Just the right amount of snark. Paperclips don't fear extinction, either, or yearn for a loving partner. There's a lot to be said for being a goddam paperclip.
Not I don't think so. So I like this idea that cancer and aging are two sides of the same coin. The 'coin' is that we're made of 100 million cells or something. And one of these cells going rogue can lead to cancer and the death of the whole thing. Life's response to this is the Hayflick limit https://en.wikipedia.org/wiki/Hayflick_limit which puts a counter on how many times a cell can divide. This limits the damage that a cancerous cell can cause (somewhat) But it also is the root of the aging process. Some cell can divide only so many times and more damage to the cell means it's death, you age. The details are more complicated than this simple picture.
I think it's one part of the aging process, but there's still a lot of other stuff, like accumulated damage beyond the ability of natural systems to repair, somatic mutations in random cells that can cause cancer but can also just make some stuff not work, accumulating persistent infections (approximately everyone is carrying around a bunch of latent infections from members of the herpesvirus family, which usually don't do anything but occasionally cause you to get some weird leukemia or have an outbreak of cold sores or get shingles or something), and so on.
Yeah sure. I just like it as a starting model. Aging is built into our cells. Re: accumulated damage, there was a post about aging and accumulated damage here a few weeks ago. I think that fits right in with the Hayflick limit, and again sure the details are messy. I get most of this from Bret Weinstein and his "Reserve-capacity hypothesis." See here. https://gwern.net/doc/longevity/2002-weinstein.pdf
Oh just a related point; I think there are (at least) two things needed for the evolution of life. First we need new beginnings, life starts out from a single cell. That's so common we don't think about it much. And the other is death.
This is definitely one of the top things I'd want out of ASI and I think it would be possible to a fair degree. Although I also think that if it ever becomes possible to look basically however you want, there will be much less value attached to looking good, so I might get bored of being beautiful and decide to revert to my naturally goblinoid form.
It's going to be sooner and way more common in virtual worlds before physical.
Reshaping ~30 trillion cells is HARD. Even gengineering a defective gene into a better gene in adults is all but impossible right now, because of the formidable difficulty of getting the edit into enough cells in a short enough period of time, and the only things we've attempted are edits in small, localized organs like eyes or livers (https://performativebafflement.substack.com/i/151736795/why-we-cant-do-this-to-adults). This is why most gengineering efforts are put towards embryos, it's substantially easier (you just change one cell, or a handful, and it grows into the 30 trillion adult cells). People are working on massively parallel CRISPR changes and various modalities to change a lot of cells today. AGI could plausibly make this happen maybe 2-5 years after the "country of geniuses in a data center" point, with some good fully automated wet labs (these exist already). But that ends with "you can now fix a handful of genetic defects in adults," not catgirls.
To fully modify human bodies (our glorious post-scarcity full luxury catgirl future) will essentially require brain uploading and downloading into clone bodies, or fully artificial bodies that are close enough to biological. This is well past our predictive threshold, the path from here to there is so stacked with nesting "known impossibly hard problems" and rife with "unknown unknowns" we can't even predict it. So I think you basically just have to hope that the Singularity / ASI is good at biology at that point, and post Singularity all predictive bets are off anyways.
So it doesn't look great. But fully immersive VR systems that actually trick the brain into thinking it's reality enough are probably on the horizon in the same 2-5 years post "data center full of geniuses."
And when it comes to synthetic life, we've only JUST been able to make a fully artificial cell that metabolizes and replicates on it's own, and even it craps out after 3 divisions or so, AND we got there by the simple expedient of "keep taking genes out of a known working organism until we get the minimal working set," and we still don't understand what a third of the remaining genes do.
And even the digital nematode, the Openworm project, has basically gone nowhere interesting.
So although the digital twin idea is solid, hopefully the country of geniuses or ASI will be considerably better at biology than we are, because we're not doing great.
On ASI predictions, I would defer to our host and the AI2027 team, which currently I think is estimating singularity in early to mid 2030's, with some holding out as long as 2050. So maybe 2030 - 2050.
"AI 2040 and Machines of Loving Grace forecast that AI will cure cancer, end disease and stop aging"
And you'll get pie in the sky when you die, is my rejoinder to that. We've come one hell of a long way in medical science, so that people are regularly living into their eighties (and now comes the complaining about Boomers sucking the life out of their Millennial/Zoomer descendants to pay for their inflated pensions and luxury retirement lifestyles for decades, instead of discreetly dying once they hit seventy, five years after retiring from work, as they should have done).
But "you'll never grow old! all disease will be conquered! cancer no more!" is more of the Golden Age SF dreams. The more we know about biology, the more we know how goddamned complicated it all is.
Remember the heady days of the Human Genome Project, when the promise was "individual personalised medical treatment for *you* plus we'll cure all disease and cancer no more"? Then it became "Okay, we've mapped the genome, that's fine but now we need to map in more detail and we've moved on to different elements":
Yes, personalised medicine, but not for everyone. Yes, advances, but disease still remains. So I'd be wary of any promises or forecasts that "yes, you too can cosplay Good Lady Ducayne".
Also, I think "completely changing your body" will only work for the very young, probably via genetic screening and embryonic selection, not for the already alive adults who are older and whose dysmorphic features are already fixed. No plastic surgery 20 year old perfect new ageless disease-free bodies for those of us who may not even be alive in 2040 due to our ages!
True, I didn't mean to imply that this is the complete solution (sorry about the confusion); that said though, IMO any M.2 drive should work for the stated purposes.
I have never tried this specific product, so once again I can't recommend it, but people seem to like it. That said though, I'd be cautious. I've tried similar thumb drives before, and their write performance usually tanks very quickly -- basically as soon as you fill up their internal cache. That would not happen with an M.2 drive.
> Can anyone recommend a USB stick that actually ... works?
I did some research before I bought a bunch of USB sticks 2 years ago. Broadly, most non-brand USB sticks are made of lies and mediocrity, but some people third party test a bunch of usb transfer speeds and cycles before failure, and you want the ones that do well on those.
I don't know how often USB performance changes, but if it's yearly, you may want to look for yourself for some of those third party tests, or ask your AI. If 2 years ago suffices, the ones I ended up getting were these:
PNY Duo Link
Samsung Fit Plus MUF-256 and MUF-128
And a bunch of the small blue Samsung Type C MUF-128DA/AM that are waterproof
I've used them heavily since then, and have verified transfer speeds on all of them at the terminal. They don't always get the full 200mb/s or 400mb/s they claim, but the speeds have been good enough to transfer them completely full of files both up and down in a reasonable amount of time multiple times.
Any thoughts on Andy Burnham and how well he's likely to do as Prime Minister? Personally I am trying to maintain a sliver of optimism against the pessimistic takes I'm seeing but would be curious what people here think.
He's very enthusiastic about devolution and giving more power to cities and regions, which is great. Other than that, he strikes me as a people-pleaser and someone who is not used to thinking deeply about policy (other than his pet project) or weighting tradeoffs. Also, he seems to think that UK has stagnated (or entered decline) due to too much neoliberalism, so I expect the stagnation to continue. At least he's a decent communicator, so he has decent chances of winning the next election.
When things are centralised, people say 'this inefficiency and incompetence are due to distant bureaucrats, who don't understand local communities, making decisions from hundreds of miles away, subject to impossibly Byzantine regulations, and handing out lucrative contracts as a result of shady lobbying campaigns. We need devolution!'
When things are devolved, people say 'this inefficiency and incompetence are due to clueless provincial nobodies who have no experience of running serious projects, using outdated and hopelessly inconsistent procedures, and handing out lucrative contracts to their golfing buddies. We need centralisation!'
If the centralized body is already corrupt, inefficient, etc - then giving a chance to devolution makes sense. Maybe at least some places would make it.
I will add that you also seem to get higher variance in quality with local rule - one intelligent and motivated person can make a big positive difference, locally - but by the same token a single dumb and motivated person has a much higher negative impact locally. If your local community is smarter and more involved than average for your (region/state/county), local rule is better than distant rule: but people are poor judges of if their communities are better than average.
That single dumb motivated person can make a bigger negative difference locally, but a much smaller negative difference than if power is centralized and that person is kicked upstairs. Moreover, when a town screwup is screwing up, you have a much greater chance of stopping him than if he's screwing up his county, or in turn, the entire kingdom.
Which sounds like it's just pointless dithering by idiots, but actually, rot *can* accumulate in a system over time, and cyclically reallocating power back and forth between disconnected bureaucracies can serve as useful spring cleaning in some instances.
TLDR: Devolution moves policymaking toward a version with the boons of the market; more information channels are available, bad ideas are outcompeted, and policymaking is "democratized" (more policymakers at different levels of policy means more ways for your average joe to access policymakers and influence policy).
Primarily:
(1) irresponsiveness to constituent demands at a grassroots level.
(2) Consider a "centrally planned" versus market economy. The first has a few key decision makers who aren't necessarily aware of what is on the ground, nor are they pricing in all relevant information (only that information that comes to them). This makes them liable to mispricing information, having little accountability (decision arbitrage) due to limited policymaking competition and so forth (think of all the things bad with monopoly).
In a market economy, each individual competes with each other individual, with a far more heterogenous array of information.
To make the analogy more direct: with more individuals in policymaking (since devolution has empirically resulted in multiple parliaments instead of one, for example), there are more access points to policymakers, closer to normal people, meaning more ideas, particularly those relevant to the average person, are given a chance.
Individuals in small towns are more likely to be able to successfully lobby their policymaker for certain changes, whereas if there are fewer policymakers, there is a greater concentration of lobby money per policymaker, so business wins out.
Some of these ideas will be clearly bad; some will be good. The bad ones are outcompeted more quickly than in a more centralized decision-making model - since the flow of ideas is greater.
Policy hubs can then compete on policy and the best ideas, through process of repeat experimentation, can be transferred between areas. Policy iteration speeds up.
If this is ringing any bells, this is state federalism in the US.
Can you say more (or link to further reading) about Scotland falling behind since devolution? My impression was that there was rough parity but I don't have strong evidence for that.
Health and education outcomes have gone from better than England to significantly worse since powers over those issues have been devolved. Clearest in things like PISA rankings.
I don't know about that; Scotland has gone from the fifth strongest region to the third strongest (by GDP per capita), it seems to be doing decently well. It started out around as good as the West Midlands; it is now decidedly better.
Now some less devolved areas have done better -- e.g. Manchester has knocked the socks off even London over that period -- but Manchester had some devolution from where it was with the introduction of the Mayor.
Overall, devolution wasn't the silver bullet it was hoped for, but it hasn't ruined Scotland either (my friend who lived there says it's a much worse place to live, but that's not showing up in the metrics).
The devil's in the detail, but I can provide a high-level point for the North-South divide.
London and the South-East subsidizes the rest of the country. In some regions (the North East), the government is 50% of the GDP (admittedly, this piles in issues of (1) ageing populations (NHS spending being high), (2) regional industry decline, neither of which devolution will solve completely). There is substantially little economic activity occurring in these regions.
Increased devolution, particularly cleaved on an economic basis (i.e. around Cities) gives greater autonomy to these areas to (1) attract talent (e.g. through subsidizing training if they move to this area), and (2) compete with London for companies' offices.
It would then reduce the ageing population burden in the North East and spur on regional industry development.
Contrast this to the (centrally planned) approach from Whitehall to achieve the same effect: Special Economic Zones. Not to sound too cynical, but all this really resulted in was part of BBC's offices moving to Salford Quays in Manchester.
Change won't occur through piecemeal projects like that especially when they are individual pet projects for politicians on 5-year terms (not to sound too populist). Mercenary incentives are ironically longer-lasting.
Further, the answer isn't throwing more taxpayer money at these areas to 50% of regional GDP and beyond; it's increasing their economic independence, generating tax revenue which can then be spent on ameliorating regional issues, and so forth.
Regarding non-metropolitan areas: regional multiplier effects (e.g. for commuters - see the Home Counties' relative wealth) would affect the whole region.
Devolution of taxes allow areas to raise their own money for projects, and spend it locally.
For example, Leeds has been trying to build a tram system for ~20 years (Leeds Supertram) but the funding keeps getting cancelled. In a more devolved system, they could raise their own funds (like the West Yorkshire Strategic Infrastructure Tariff) for this.
For similar reasons that CEOs micromanaging developers is bad: central planners lack the detailed local/domain knowledge to make appropriate decisions even when they have knowledge of problems' existence or importance in the first place, response / round trip time is poor, there is too much going on overall for them to give enough focus to specific local areas, and they only very indirectly suffer the effects of their bad decision-making so lack incentive to do better.
"For more than three decades — from the introduction of the Distribution of Industry Act in 1945 to the beginning of the 1980s — it was impossible to set up a new factory in Birmingham, or to expand one, without an “Industrial Development Certificate” (IDC) from the government. "
I think devolution is a response to the rise of the far right. Even when establishment politicians are effective - such as by increasing wealth or lowering crime - their popularity seems to decline. The conclusion is that the voter feels locked out of the process, sometimes even voting for blatant and unapologetic wreckers. The idea of devolution is to transfer power to smaller and smaller authorities to give the individual voter more (apparent?) control. I don't know if it will work or not.
It's worth remembering that in the English context, devolution was a two way move. Some transfer of powers and funding from Whitehall to the regions. But also a pooling of sovereignty so that decisions which had been taken at the level of the individual local authority (in GM, Manchester, Bolton, Bury, Salford etc) would now be taken by an elected mayor. The "gainshare" deals given to the mayors (thirty year funding commitments from central Government were, officially, a recognition that this pooling of sovereignty would lead to better economic outcomes, thus more receipts for Treasury; and that some of this should be allocated locally. Not very scientific, but I have seen sillier things. And I think it's definitely sensible for functions like transport planning to be done across a city region rather than separately in five or ten or whatever separate authorities.
The trouble I think is that for political reasons the new combined authorities have been drawn too small. GM, West Mids are probably ok; S Yks, Liverppol perhaps suboptimal; Teesside and Hull really too small. It's understandable how we got to that point, and especially in the Tory party there is still folk memory of Ted Heath's rather Bonapartist rationalisation of the early 1970s which gave us things like Avon and Humberside, and which as it happens was largely right. But the need to establish consensus, large consultation exercises, affirmative order debates on the legislation took us in a suboptimal direction.
By and large the English provincial cities did best when there was strong local administration (both the nineteenth century and 1930s town hall and city hall in Leeds are among my favourite English buildings, and the confidence to produce those was quite a thing).
Leaving aside the characterisation of case histories as "far left" or "far right" and their stigmatisation as "subversive", the inevitable conclusion of Hatton-like events must surely be that all central governments will react badly to local government deviations from the party line if in their view it becomes excessive and shows signs of wanting to challenge the status-quo.
In part it was trying to solve a problem which, following Brexit, no longer exists. While the UK was in the EU, the Blair government in the early 2000s was keen to divide the UK (including England) into ten or so smaller self-governing regions, like Swiss cantons, for easier full assimilation into the EU and discarding of our national identity. Each region would have been little larger than a typical US county. But a referendum conducted in the first such proposed English region firmly and quite rightly told the government where they could stuff the idea and it was abandoned.
For Scotland, Wales, and Northern Ireland, another of the UK government's motives to allow devolution was to ease the pressure for full independence. Arguably this would have encouraged ever more concessions until independence would be almost achieved anyway. But that doesn't seem to have happened, probably largely because ..
The desire for independence is also suppressed by something called the Barnett formula, whereby more is spent in these so called "devolved" regions than their tax receipts. That is probably mostly why the referendum conducted a few years ago by the Scottish Nationalist Party (SNP) failed to gain its hoped for majority for independence, and polling suggests another referendum would currently also fail.
>the Blair government in the early 2000s was keen to divide the UK (including England) into ten or so smaller self-governing regions, like Swiss cantons, for easier full assimilation into the EU
>: The Blair government initially planned for English regional assemblies that would coordinate policy and connect with the EU's regional funding mechanisms. Following the 2004 "No" vote in the North East, the government shelved regional devolution in England.
So.it's about regional.development, the nice thing where the EU gives you money, not the nasty thing where they tell you what to so.
The cynical answer is that it solves the problem of Duverger's law that it is hard to start a national party in a first-past-the-post system, whereas it is easy to start a regional party.
Here's an anecdote about a problem that it could solve, although it's not clear that anyone advocating devolution is arguing that this is the point. Local governments spend a lot of time lobbying central governments for money. They don't want to do any particular project because maybe they could get the central government to subsidize it more.
People from different places have different values, and therefore want different things from their government. Devolution means that laws and government actions more closely match what more people want.
At the level of Scotland, Wales and Northern Ireland, they all have specific local problems and (except Wales until recently) their own legal frameworks that overwhelmingly-English Westminster politicians don't understand and don't prioritize. The demographics of the UK mean that if the government is completely centralized, as it was until the late 90s, you don't need to be some kind of public choice galaxy brain to see that a government can survive based only English support.
Within England I'm not convinced it does make sense. You can argue the same logic that a government can surivive based only on support from Greater London, so the English regions need their own lawmaking bodies. But that's not as clear. Governments always need some support from outside London. At one time that could be London+Scotland, which was the foundation of the last Labour government, but that no longer works, so the English regions have some power. And the question of how to split them up has always bedevilled the whole project. There are regions that make sense historically and still have relevance, like Cornwall and Northumbria but they're quite small. The majority of the English population outside of London lives in the cluster of cities that spans from Liverpool to Leeds and Birmingham, but there are 4-5 cultural and historical units within that that make sense but again they're very small. The tension over English devolution is always between big units that make administrative sense, but which people don't really care about, and units people might actually feel some loyalty to but which are too small to be administratively efficient.
Devolution works if cities get both decision and fiscal power, i.e. ability to collect taxes locally. This will align incentives but also increase regional inequality, at least in the short term. This is a main reason why I do not put higher than 25% on fiscal devolution happening.
In summer 2025 UK abandoned plans to introduce zonal electricity pricing because it could lead to "potential price inequality and postcode lottery"[0]. As a result NESO (and ultimately consumer) is occasionally at the same time paying wind farms in the north to not produce electricity AND paying gas plants in south above strike price to produce it.
Unfortunately same "postcode lottery" reasoning applies to fiscal devolution
>he seems to think that UK has stagnated (or entered decline) due to too much neoliberalism, so I expect the stagnation to continue
I mean, after two years of Starmer, this assumption should be self-evidently correct.
I would admittedly be more bullish on Labor if it was still led by Corbyn or someone with similar genuine ideological convictions, but it's not inconceivable that a mere pragmatist responding to evidence will be enough. (And surely communication will NOT be enough, at some point the government needs to start delivering if it wants to win another election.)
I imagine he'll be the same as Starmer: not a bad guy, but utterly lacking in vision, and without a plan to fix the UK. Also, a people-pleaser, will find it hard making difficult decisions. So will almost definitely be ineffectual.
There has almost certainly never been a worse time to be British Prime Minister than summer 2026. The country is in a very poor shape financially and is still suffering from the fallout of Brexit, that disastrous self-inflicted wound; the special relationship with the US has proved the flimsiest of broken reeds; the high likelihood of direct or even indirect conflict with Russia and the low (but still conceivable) likelihood of conflict with the USA within the context of Nato; the obligation to substantially increase defence spending for the aforementioned reasons; and above all the very significant anxiety and polarisation of the electorate are all factors which will be working against Burnham. So will the native conservatism of the English and the guaranteed reluctance to accept the inevitably radical solutions the new PM will have to resort to.
On the bright side are the relief with which his accession will be greeted, particularly after Britain's incredible string of disastrous leaders over the last decade, and the near-certainty that Farage's latest move to get himself re-elected to his parliamentary seat as a way of demonstrating political strength will result in making him a national laughing stock, thus finally sinking him and Reform for good and all. But the period of grace will be short, and Burnham must make the very best of it. Personally, I am optimistic that by exercising good sense and some considered but radical solutions to the country's problems, he will succeed in turning the old ship around.
Why does this not make sense to you? Inside the EU, the UK is the most free-market country in the EU, which attracts a lot of continent-scale (and wannabe continent-scale) businesses to base themselves in the UK, where they can sell to the EU. "Lots of continent-scale businesses" is also a notable feature of the US.
Now those businesses go elsewhere, with the net effect being that the UK's economy has a lower proportion of such businesses, much more like a merely average European country. The loss of a specialised role in a larger ecosystem results in becoming more like an average non-specialised country.
Good question. Both are reasonable index years and some analyses check both as a sensitivity analysis to make sure it doesn't affect the result too much. In a nutshell, the reason 2016 is a reasonable choice is because a lot of investment decisions were already affected by the referendum result itself
"The UK economy has materially underperformed since the 2016 EU referendum, shifting from a growth profile closer to the US to one closer to the Euro area."
I don't find it convincing that Brexit made us more similar to Europe.
I'm not sure whether you're curious or snarking. I'll assume the former but please let me know if it's the latter!
They're not saying that Brexit "made us more like Europe". They're saying that our growth profile (i.e. how GDP changed over time) was better than European comparators before Brexit, maybe almost as good as that of the US, but after Brexit our growth was much worse than that of the US, as bad as, say, France and Germany's already were.
All countries have a bunch of factors that make them do well or badly, so no one comparison will cover everything. But the "doppelganger" analysis reported by GS (also known as synthetic control) is a statistical way to try and get as good an estimate as possible of the impact of one particular intervention (Brexit) in isolation.
I do find it convincing that The Powers That Be set out to hurt The Queen's Country after Brexit(I assume they find King Charles to be a lot more on their side.)
Germany is suffering from the Greens, and France is suffering from... well, being France. Also, for real, legit propaganda designed to cripple France (air conditioning should ring a bell).
This is such a wrong take! It's exactly the opposite. The best time to lead anything is when it's at its worst. The UK has leading universities, stable and functioning institutions, mature legal and finance industries... In short, everything a country needs to thrive.
It only needs to fix a few things holding it back - mainly perverse welfare incentives and planning, and it's set to grow.
It would be relatively easy to fix... unless you're attached to the ideology that screwed it up in the first place, in which case it's impossible.
In terms of perverse welfare incentives, the only concrete policy I've seen from him so far is that he wants to build a whole lot more council flats, which are one of the most perversely terrible forms of welfare I can think of.
"The best time to lead anything is when it's at its worst" is something of a free-floating statement, unsupported by any evidence; one could just as easily reverse the meaning to "The worst time to lead is...", and it would make just as much or as little sense.
And, yes, Britain still boasts some viable and important institutions, but they are badly at risk. Education, and particularly those leading universities, have suffered from the determined efforts of successive administrations to squash their independence. Health? We have all seen how NHS, once the world's finest free universal health system, has been slowly reduced to a dispirited and underfunded ghost of itself. Finance? Yes, if you regard playing the role of the word's favourite tax shelter and the promotion of private equity to represent productive financial instruments for the nation.
In each of these cases, ideologically driven principles have been responsible for their change for the worst. If Burnham can reverse even some of them, he will be on to a good start.
Sometimes it's fine to make a statement unsupported by evidence but which is backed by common sense. It's often easier to turn things around than to maintain peak performance.
I don't think Cambridge, Oxford, or Imperial are really at risk of not being world class any time soon. Finance is essential for growth - innovation depends on it. The point I'm making is the UK has institutions that can very easily support explosive growth.
I don't think the issue in the UK is bad ideology. It's mostly lack of one. No PM in the past 20 years has had actual beliefs, except for Truss, who just did a terrible job. Everyone else has gone with the flow, and caved in when confronted with tough choices.
What the UK needs is a leader with guts, who is prepared to cause some pain to some minority groups (pensioners, landlords, welfare recipients), but who can also communicate why changes are essential.
> What the UK needs is a leader with guts, who is prepared to cause some pain to some minority groups (pensioners, landlords, welfare recipients), but who can also communicate why changes are essential.
Agreed. Though I would add the conservative press to the list.
Britain could adjust its foreign policy ambitions in line with its current economic and geographic position, maintaining a nominal military focused on coastal defense, and accept that empire is a thing of the distant past and that it doesn't need to get in fights with countries on other continents. It could even use the savings from doing so to improve the lives of it's citizens.
It probably won't, because seemingly every politician in the world these days wants to take money out of the hands of those who need it most and instead hand it over to arms manufacturers, but it *could*. That's a thing a world leader could do.
Ever since Suez inflicted a reality check on Britain's imperial ambitions, the Foreign and War Offices in London have had trouble adjusting to the world as it is; mercifully, Labour governments kept the UK out of both Vietnam and Iran, though it took a third pseudo-Labour government under Blair to drag Britain into the Iraq and Afghanistan fiascoes.
The problem Burnham will now have to face won't be imperial fantasies, but the hard fact that Europe is facing military threats very much closer to home, from the Baltics to Greenland.
I very sincerely hope that the UK will never again repeat the political and moral folly that led Margaret Thatcher to fling a fleet across the Atlantic to maintain her country's grip on a remote and wortheless fragment of land to which Britain had no claim beyond the remains of a tattered colonial inheritance; though of course the insane expenditure of blood and treasure did ensure her re-election and offered a useful precedent to Bush Jr. As for Gibraltar, that is another faintly ludicrous relic which survives thanks to the good sense of successive Spanish governments; I doubt increasing the Royal Navy's blue water fleet beyond its current single (and singularly ineffective) Type 45 destroyer will do much to further secure the Rock.
Here we are, forty years after the Iron Lady fought a war over the Falklands, and it looks like not a single drop has yet been recovered. There's a thing where people look at every single war ever, find some oil-exploration report saying there might be oil there, and say "See! Those evil meanies were really fighting that war over *oil*!". This is almost always nonsense.
Reform (or similar) probably aren’t going away. I once visited the industrial wastelands and decrepit sea side towns of Britain and wondered by they were voting Labour, which wasn’t really representing them anymore. And the Tories were never an option. It’s hard to predict when a new party can arise in a two party system but that kind of vote isn’t going away.
You are absolutely correct that none of the three historical political parties in the UK can really claim to represent the lowest layer of today's proletariat, whose basic grievances could be summarised as social marginalisation, hatred of immigrants and lack of opportunity and decent jobs. A serious attempt to deal with the latter would probably also allay the others - except that AI has now also arrived to further devastate the job market. Admittedly, the devastation is currently largely affecting white-collar jobs, but it's clear that the poison will inexorably seep downwards.
The country is "in a very poor shape financially" due to the shadow banking sector. Why else does King Charles express himself to Trump on a charm offensive?
I am also trying to remain optimistic, but he's coming in at a very difficult time and the expectations in the country are unreasonably high so I think a bit of disappointment is inevitable.
The thing I'm most excited about is he's previously been interested in land value taxes and seems to have thought carefully about them and understand why they work. He also understands the importance of infrastructure in a way, e.g., Cameron didn't. Trouble is, those things take time to pay off and he doesn't have much time.
He's going to be squeezed from all sides, with Labour backbenchers wanting increased welfare spending, markets wanting tough fiscal rules, reform and the Tories banging on about immigration and North sea oil, the country desperately needing a bit of investment, all while the population ages, the planet heats, and wars get fought.
A lot of this long term stuff might become AI-moot, but I've not heard him talk about his AI plan either.
He doesn't have a plan to deal with the major issues facing the country and he seems to be willing to say yes to every department asking for extra spending. It will be interesting to see who he makes Chancellor.
He’s a better communicator than kier Starmer, but I think he’s a flipflopper who won’t take the necessary tough decisions: end the pension triple lock, raise taxes, nationalise all golf courses to build houses on
He seems to have done a fair job in Manchester. His policies seem better matched to Industrial Socialism (ie use government policy to support specific industrial goals) rather than Benefit Socialism (ie use government policy to force the productive to support the workshy), so that is a minor improvement over the current Labour leadership.
Like others have said, my expectations for uk politics are very low just in general, and Burnham doesn't come across as an impressive guy.
But he has made noise about introducing a land tax, which sounds extremely interesting and is not something I'd ever expect to hear from a mainstream political figure. Most likely a land tax is waaaay too radical to ever actually get implemented in any substantive way though, so I doubt it will actually happen, but just on its own that could fix a lot of Britain's problems imo.
UK PM is kind of a poisoned chalice at the moment. I think there are people who are reasonably competent and will give it the old college try but they will be -100 in favourability ratings within a few months of taking office whatever they do.
I don't see much chance he will really make much difference. In the short term, he's more charismatic and if he takes good advice on communications he may get Labour through the next election.
Looking at this as an emmigrant, so I know UK politics well but I'm not caught up in the news cycle, there's nothing really wrong with the current government. It's .. fine. Keir Starmer isn't Tony Blair or Margaret Thatcher, but he's also not Liz Truss or Boris Johnson. The party isn't borne aloft by charisma and careful, ideologically coherent policy development, but he's not really made a lot of unforced policy errors and he's only mildly messed up the communications.
There are two basic problems that face the UK - Brexit has been bad for the country economically, and British productivity has always lagged behind other European countries, let alone the US and China. Those two effects plus the accumulating effects of crises where the UK has just kind of scraped by mean there just isn't as much money as there was in 1997 so its harder to come through on big promises.
In addition to that, there is a human capital problem. Keir Starmer is fine, and I'm sure Andy Burnham will be fine, but the last government with anyone with any level of real star power - the ability to both master a brief and communicate the plan - was David Cameron's. Since then its just been a parade of mediocrities and bullshit artists. These are just not serious people. I did a back of the envelope calculation the other day, and of the people I grew up with who got good educations - doctors, lawyers, engineers, etc - a third have emmigrated. I don't think this fact is unrelated.
Here is a description of a character in Bleak House:
"[she] would be handsome, but for a certain feline mouth, and a general tightness around the face, rendering the jaws too eager, the skull too prominent."
I particularly associate it with Victorian novels, but I think it's not completely unknown in more recent novels to have a character's appearance described in this way, where certain aspects of personality are perhaps hinted at in appearance. (In this case, she is about to make fun of a pretty character, and I think this description suggests at jealousy). At the most extreme, you might get a golden age crime novel, saying something like:
'Mr Craftley had cold, pitiless eyes and a haughty nose, down which he looked at everyone with a sneering disdain. His mouth was hard and pale, with a marked tendency to cruelty at the edges.'
I *think* what's going on in these cases is that the author is just telling you what the character is like right from the outset. 'Mr Craftley is a baddie, and he probably wouldn't think twice about killing to advance his aims, but is he actually the murderer in this case? See if you can figure it out.'
Now I'm not great with faces and visual memory in general, but I'm not aphantasic and I don't have the face-blindness thing. So my questions are:
1) Does anyone actually see faces of new people in this way in real life? 'Ooh, this woman would be prettier if she didn't have such a tight face'; 'Woah, that man has a haughty nose, best stay away!' 'What an interesting feline mouth you have.' etc
2) When people read these descriptions, do they actually form a visual impression? Or is it just (I think like me): 'Oh ok, this hasn't helped me to understand what the woman actually looks like at all, but she has a slightly odd face and is not that attractive, so maybe she's jealous, got it.'
1) Sometimes, if I'm interested in the person. Maybe not right away after I meet them, but after I get to know them, I might notice: "he has really unpleasant fish eyes that always look half-closed, he feels untrustworthy." Sometimes I'll be a little torn: "something about his face makes me think he's untrustworthy. Is it just me being shallow and judgmental, assigning significance to something meaningless, or it it the Voice of Intuition telling me something valuable?"
2) Usually? It depends on the description. I can visualize Mr Craftley pretty well (no idea if my imaginary Mr Craftley looks anything like what the author wanted to convey, but it's a clear enough visual), on the other hand I don't really get a good image of the proud owner of "feline mouth" and "too prominent skull". (What the hell do those things even mean?) Actually, I do, but in my head she is butt-ugly and not remotely close to handsome, so it can't be right.
"Skull too prominent" is a bony face, so the skin is tight/there is little flesh on the bones. You need a certain amount of facial fat for attractiveness, and depending on the era, that is a greater or lesser amount. Some periods prefer rounded cheeks, others like prominent cheekbones.
"Tightness around the face" is also describing expression, though in a subtle way; you get the impression this woman is eager, maybe *too* eager, to say things (generally nasty things) about others. She's impatient or impulsive, looking out for her own advantage. Take that with the "feline mouth" and you get the broad impression of someone who's predatory in some way.
I don't often see people like this in real life, but I /do/.often see it in films and plays - actors.and makeup departments can convey personality through face, convincingly but not honestly.
1: Bit of both. I'm not going to say someone has a "feline mouth" (which I assume means "tiny and frowny" here, instead of "sexy catgirl" like anime has instilled in me), but "that guy looks like he does drugs", "that woman's got a pig nose", "that guy's got the eyes of someone with dementia", "that guy looks like he holds a steady job and pays his bills".
2, No, they look however I want them to. The only value in visual descriptions is being able to distinctively refer to the character later without using their name. A description of a "feline mouth" means we can switch viewpoints to someone who doesn't know [her], but the audience still immediately recognizes her.
I have found things like mouth position and the lines around someone's eyes to actually yield information about character. An older person who has smiled a lot tends to develop different lines than someone who is mostly stern, for example. I also harbor some suspicion that if someone looks a certain way, it colors how others treat them, and they could end up playing the role eventually if they don't have a strong sense of self to combat it
Authors of the time probably had some unexamined background phrenology beliefs, and this kind of description is faster than breaking appearance and character apart.
I read this as a description of an ugly character. “She would be pretty, but for…” means the same as “She isn’t pretty, because of…” It’s worded to sound like it’s trying to be nice, but that niceness is sarcastic. I think it’s associated with Victorian writing because many writers in the Victorian era make fun of how polite society can actually be vicious.
I don’t think I really notice people with eager jaws or prominent skulls in real life, but I do think I might’ve seen them in video games, digital art, or caricatures/cartoons. I think those are exaggerated descriptions that only apply to exaggerated depictions in various media.
With that said, the impression I get of the feline mouth and eager jaw is that the character habitually holds their lower jaw slightly forward, giving a flat, round lower face like a cat’s. The implication is that the character is always tense rather than relaxed, and looking for the best response in every situation. The tightness of face leading to prominence of skull suggests she might frown or raise her eyebrows a lot, suggesting she is pained when things don’t go her way or people judge her harshly. The implication is that she is exactly the sort of character who would be upset by the narrator’s description, which might be a particularly Victorian style of cruelty.
I don't personally think any of these kinds of things when I meet someone new, but when I read them in stories, it does reshape the mental picture I had of the character. I think the author just does it to let you know a bit about the character from the get go like you mentioned, as well as tell you some about their appearance as well. To me, it seems like a good way to introduce a character when they need to be remembered, but you don't have room in your story to just be like "oh this is Mr. Pickledurphy" and then jump right into a lot of dialogue or something that shows the type of person they are.
I do not do this irl and I hate it when authors overdo it in fiction.
The stereotypical example I never want to read again is describing characters as having "intelligent eyes".
There is no such thing as intelligent eyes, it's just a lazy way of the author to tell me this person is supposed to be impressive.
Usually these people then do not act any more intelligent than anyone else in the story and not rarely they're bloody idiots.
There's some forms I believe are defensible. Somebody mentioned laughter lines already. It is also certainly possible to get am impression of somebody on first contact and perhaps intuit something about their character in ways that are hard to convey in words.
It is especially badly conveyed by "they have <insert trait> <insert part of face>" and the like, however.
It might not help that I do not imagine character's appearance while reading at all and, if actually reading and not listening to an audio book, will liberally skim physical descriptions.
I would say "intelligent eyes" means they have the facial appearance of someone intending to engage the topic as an equal. Focused eyeline, not fidgeting, not scowling or awestruck or other expressions of having a pre-formed opinion.
Sounds to me like what you're describing is the behavior of somebody who presents as an attentive and neutral listener. I'd argue "has intelligent eyes" is a poor way to describe this.
For one I can be bored out of my mind in one conversation and paying sharp attention in another, but my intelligence surely remains unchanged.
For another, these kind of descriptions are very common when characters first meet, often without them having spoken a single word. Any narrator who believes they can deduce the intelligence of someone they're meeting for the first time based on their physical appearance and / or facial expression is... let's call them deeply mistaken.
To contrast, I have no issue with one character calling another intelligent once they've actually had the opportunity to judge that. But I still feel it would be helpful to separate this from their physical appearance, as best I can tell these are in fact unrelated.
What would be a better two-word description of it? "Attentive eyes" gets you a bellboy.
I would argue you're using the wrong definition of "intelligent" here, viewing it as an inherent trait instead of dynamic. If you didn't hear half the conversation, your ability to intelligently contribute to it will drop.
Personally I would just drop it without replacement. There need not be a common term to describe what does not exist.
Ideally the intelligence shines through in the writing.
And I agree that the definition of intelligence I'm using is the reason I judge it so harshly, I believe it is accurate, though. Because my impression is that the purpose of the whole thing is to tell the reader how desirable the person in question is.
At least the context I encounter this pharse most often is some kind of romance novel in which the prospective love interest is first introduced and must be hyped up to the reader. The second most common way I see it is of the main character meets side characters and we're to understand which of these we're supposed to like and respect and which to dislike or disrespect.
And that only works if it's about the "inherent trait of the person" kind of intelligence, not the "currently saying something clever" kind of intelligence.
I'm not going to try to defend romance novels, but "we can only like someone if they're smart" is on your end. "Pays attention to you/dismisses you" maps exactly to "you should like them/you should dislike them".
I think it was largely because Dickens published many of his novels, including Bleak House, as serials in a popular monthly magazine. So he probably thought many of his readers, with limited education and literary sophistication, might want insight into a character from the outset and find it uncongenial to have to judge the characters more gradually from subtle hints and the unfolding plot, especially as the full work would have to be read over several months.
For the same reason presumably, he also gave many characters comical (by his standards!) twee names which were also meant to indicate how the reader was to consider them. Examples include Mr Bumble from Oliver Twist, whose name suggests a bumbling idiot, and in Nicholas Nickelby the harsh schoolmaster called Whackford Squeers, clearly a liberal dispenser of the birch! I imagine "Squeers" is supposed to suggest a combination of "squint" and "queer" (odd), indicating a bitter and twisted view of the world and his charges.
Out of interest, I decided to check my post in one of those online AI text detectors. It decided it was 78% written by AI, whereas the correct answer is 0% . Make of that what you will!
"A host of 19th-century English authors were influenced by the idea, notably evident in the detailed physiognomic descriptions of characters in the novels of Charles Dickens, Thomas Hardy, and Charlotte Brontë. Descriptions over one's appearance in the written form was a way for one to gauge the character's intelligence, ability, morals, and social status."
2) I dont know what a feline mouth is, but other than that yeah. Maybe Im misunderstanding some british subtleties, but she sounds like Id be into her.
1) I usually dont verbally think it, but if I were to describe what I think, its in the general vicinity.
Is there a way to contribute to AI alignment or do some other net-positive Data Science-related work while not being located in first world country? Some kind of Data Engineering or Analysis remote postion?
I'm not an expert, but here are some random ideas from a random bystander:
1. Language translation research. Presumably, chatbots and other AIs do not work as well in the local vernacular as they do in English. Research why not, and also uncover and curate what should be placed into the training corpus for the next generation of LLM models.
2. Cultural idioms. As above, but with a focus on the failings of metropolitan generative AIs in producing objects of local musical, artistic, and poetic traditions. A particular focus on voice and video is probably called for.
3. Educational artifacts. Use AI to assist you in translating material from the global canon into the local language. I would focus on applied math, engineering and science, and philosophy, but your taste may differ.
4. Educational apps. Use AI to create educational interactive apps to teach English and AI theory in the local vernacular.
5. Engagement hooks. Use AI to generate short-form videos jingles, and memes to get young people interested in engaging with this technology.
In other words, notice the areas in which you are starting with a leg up on the Silicon Valley bros, and run in the race where you have a head start.
I want this plan A thing to happen. It seems to me that non believers, those who only imagine the mundane and boring scenarios where AI merely takes all our jobs and hostile actors attack us with drone swarms and kill bots and barage us with propaganda, should still be on board with most of the early part of the plan yes? I don't care how hot the AI topic is in 2028, we simply aren't going to woo normies into the cause talking about galaxy level shit and 2 million dollar stipends. Alternative narratives should be folded in somehow, I believe we can find compatibility. Shit even I have a hard time believing I'm not crazy to take this ASI stuff seriously, I have a deep emotional need to cover for 2017 normality despite the fact that it has already evaporated entirely.
I think Plan A can't have it both ways. If they want a slowdown and control over AI alignment, then they have to frighten the horses ("AI will take all our jobs and take over the world and kill us all!!!!") but then if they want the good things ("AI will take all our jobs and take over the world and be so super-duper-hyper productive the global economy will balloon to the size of Jupiter and we'll be rich off the scraps the AI oligarchs throw us!!!!") then they can't scare the horses *too* much, else the call will be for a total AI shutdown and they don't want that.
Never mind that way too much money has already been sunk into AI so it better start getting profitable sharpish, and the companies and lenders and investors to those companies are not going to tamely go along with a total shutdown.
I have no idea how they thread that needle (present AI as super-scary but not so super-scary it will be dismantled completely).
AI is already profitable. The massive costs are speculative investments that might or might not pay off, but they aren’t operating costs. If the investments don’t pay off, a lot of investors will lose money, but existing generations of LLMs will stick around. (OpenAI’s revenue might have been ~$26B vs operating expenses of ~$1.5B, though I haven’t checked those numbers.)
I am not yet a bubble-narrative-monger, but it is worth pointing out that a large fraction of that $26B revenue comes from R&D spending by enterprises that have also bought into the hype.
That’s fair, but I think a lot of the operating costs would be proportional to scale. There would be some efficiencies of scale, but I doubt those are suppressing prices by a factor of 15.
I'm not sure about this, last I checked I was using (read: wasting through repeated calls for weird layers in order to appear busy) somewhere north or 2-3k $$ of compute in a month for the price of 150-200$ before the price hike, and it still is wildly under priced IMO.
Just doing my part to make them as unprofitable as possible in my goal to suffering max certain parts of society.
In any case; the models that can do anything useful have negative profit on every call that does anything of value. Maybe this has changed recently, but I doubt it.
Based on reported income of companies, the leaked token rates from probably uber, and reporting from the FT, and having instructions form on high to use ai constantly reversed one the subsidy went down a bit.
You can also back into this via raw hardware and electricity costs. Broadly, it costs ~$1 per million tokens using H200's, and even cheap models cost $2-$5 per million, and good models cost $10/$50 in/out for Anthropic’s Fable or $5/$30 for OpenAI’s Sol.
We've also known at the aggregate level that AI companies had ~40-50% margins on inference in 2025, and 70-80% margins on inference in 2026.
Quoting Sean Goedecke below:
“A Nvidia A100 consumes 400W of power under full load. In practice, even a carefully-tuned inference server will not be at full load all the time, but it’s at least an upper bound. Suppose you’re running a dense 70B model1, which will fit comfortably (unquantized) on four A100s at around 2M tokens per hour. At industrial power prices, that’s about 13c/hr in the USA. Suppose (pessimistically) cooling is the same cost. That’s about 13 cents per million output tokens2.
Let’s amortize the cost of the GPUs, since that’s going to be the most expensive part. An A100 costs about $20k. If each A100 lasts around five years3, you’ll have to make 16k/yr in profit to recoup your capital investment (or $1.80 per hour). At lower utilization, it’ll take longer to recoup, but your GPUs will also last longer. Either way, your overall inference costs are at about one dollar per million tokens.”
Yes the problem is that the mainline AI safetyists both want to slow AI down *and* want ASI to someday exist, something something utils something something lightcone. So they end up torn in just this way.
Eventually opposition to AI will be taken up by people who actually oppose AI.
Which would be good! The primary issue with most rationalists is that they want ASI to exist at all. I'm ok dying at 85 from cardiovascular disease. It's fine. I'm not ok rolling the dice on dying at 30 or 35 because some idiot wanted to create God in a bottle.
There's a rationalist motte and bailey game that the more pro-AI among them obviously know they're playing, "we're going to cure cancer and end hunger, if we can do that safely without killing everyone how can you possibly oppose that?" When in reality that's only incidental at most to what they're trying to do by building ASI.
It is an unfortunate historical fact that being concerned about AI + having money + being organized is a combination that first arose within a community that had particularly weird ethics. Very good epistemics! Great exploration of decision theory! But weird ethics, weird in the "most people don't think like this" way. And in the other cases, hey most people probably *should* think more like rationalists, when it comes to evaluating propositions and making decisions about what's optimal for their objectives.
But even the ideal best case scenario of "God in a bottle" is really strange crap that living humans like you and me do not want, weird crazy stuff that can only be justified by gesturing at trillions of years of hypothetical minds floating through space and assigning some value to that whereas few humans on Earth would give it ANY value. Heck, even "curing" cancer wouldn't be some normal thing like taking a pill that makes cancer stop, cancer is a biological tradeoff against degenerative disease, anything that would actually stop cancer from happening probably throws the entire system off and winds up looking like some wild sci-fi transhumanist genetic engineering experiment.
I think you're right about having to thread the needle, but I also think we're so far away from an AI pause that there's no point pulling back on scaring the horses (yet). Maybe this is something we should worry about if a pause bill is being discussed, but for now I think we can count on the pro-AI / pro-free market people to balance the scales.
Also, their concern about Plan S wasn't that the pause would be a bad idea, it was that the treaty governing the pause would eventually fall apart (as every treaty does) and return us to the status quo. But they also seem to think that there are really good variants of Plan S that involve building better Plan As during that pause period. I think it is conceptually possible to do a really long temporary pause and come back to Plan A later, even if the political will would be pretty difficult to manage.
One subject on which I'd love to read a "more than you wanted to know" is the subject of early agriculture and human height.
One of the weirdest contrarian takes that has recently become accepted wisdom is "agriculture was bad, actually". This, if true, would be surprising -- we know that agriculture rapidly spread out from the points where it was developed, we know that hunter-gatherers either converted to agriculture or were displaced by agriculturalists wherever they met. There's obvious intuitive advantages to agriculture too, in terms of providing a larger and more reliable food supply. So all the intuition would be on the side of agriculture over hunter-gathering.
The evidence against agriculture, as I understand it, is a bunch of archaeological studies which suggest that agriculturalists in at least some places were shorter than the hunter-gatherers. This has been interpreted as suggesting that agriculturalists were actually malnourished compared to hunter-gatherers. But reading the abstracts of papers like this https://www.pnas.org/doi/10.1073/pnas.2106743119 I find it less convincing; there's some convolution of genetic and nutrition effects, and the evidence is not equally strong in all parts of the world.
So what's the deal? Is the evidence for agriculture-bad really strong enough to accept this counterintuitive conclusion? Or is it one of those things where a single bad study has been taken too seriously thanks to pop-science popularisation?
Romanticizing earlier simpler times is a recurrent pattern. After all, our species evolved in a relatively simple hunter-gatherer environment, so we're probably deeply adapted to that kind of life: hard but not complicated. Modern life is exactly the other way around, so it grates, and agriculture is the point where we switched to a settled life, with the relative complication of having to plan at least a year ahead.
I don't see any of the complainers taking to the jungle though.
I think it could well be the other way: we have the tall hunter-gatherers because the life is so finely balanced, the unfit died young (*just* enough food to go round so that the best hunters get fed and the unproductive are allowed to starve or die in infancy). Agriculture gave enough of a surplus that the sickly young and the unproductive old could survive long enough to leave short skeletons.
I suspect the best way to see what our ancestral hunter-gatherer lifestyle was like is to look at chimpanzee and gorilla troupes today and see what they do with all that spare time.
Populations expand until food surpluses are used up. But an agricultural civilization will have a more steady level of food along with "consumption smoothing", whereas hunter-gatherers would have bigger swings as game populations crash.
Looks like the issue is famines, which for an agricultural tribe meant "limping through" and for a hunter-gatherer tribe either meant "letting people die" (so that one could still have hunters to "run down the antelope" -- a pretty calorically high workload).
The complainers can’t easily take to the jungle, because most of the good land has cities on it. I was in my teens before someone pointed out that I thought of Australian Aborigines as living in the outback, when actually their biggest population centres were on the coast. It’s just that ex-Europeans took over the coasts but found it to hard to use most of the interior, so that’s where Aborigines were able to keep more parts of their cultures.
Yes, and my understanding is that this is exactly where the highest Aboriginal population density originally was.
It makes sense that the land best suited for hunter-gatherer lifestyles is not the same land that would be most attractive for European-style agriculture. The tropics support a greater density and variety of life than anywhere else.
My understanding is that the theory behind "agriculture was bad" is that it led to such large population growth that the average person was worse off.
Note that larger population = more likely to prevail in war, so agriculture gave advantages at the group level, regardless of whether it did so at the individual level. Hence, it is unsurprising that agriculture spread.
I have two problems with this version of the theory though:
1. If the population was roughly stable during hunter-gatherer times, this would presumably imply some kind of cycle between periods of more-than-sufficient food and periods of mass starvation once the population exceeded the carrying capacity of the land.
2. For agriculturalists, the solution to running out of food is to put more land under cultivation. Eventually you run out of land, but that wouldn't have happened in neolithic times, at least not in most places. France, for instance, was still 50% forested in the 10th century AD.
Melvin, during the stone age humanity went from a few scattered groups in Africa to discovering the globe and they did so in small groups with stone tools. When we modern people discovered the continents we found the descendants of those stoneagers waiting for us.
When our World Wide Experiment with agriculture and science ends we will go back to that simpler lifestyle.
> the solution to running out of food is to put more land under cultivation
A lot of running out of food happens on short timescales though. Two consecutive years of crop failures and you're probably eating your seed grain. No additional land would help.
Also, the larger your group is, the further an average person is from uncultivated areas. So of course at some point people will have to notice that they don't have enough food, but getting up from your (your parents') land and moving somewhere to the periphery becomes less certain and riskier proposition so probably less and less attractive.
The argument is basically pure Molloch. You end up optimizing for calories/acre, which gives you monocultures, deficiency diseases, and when things go well, has you bump against Malthusian limits. Though I think some of the theory was also noble savage nonsense about hunter-gatherers that ignores the whole "nasty brutish and short" aspect of that lifestyle.
The agricultural era was nasty, brutish and short. The early industrial era as well. If the hunt was plentiful then the individual was better off in a hunter gatherer society.
Pre-agriculture hunter-gatherers lived at very low population density and had factors holding population in check other than literal starvation (disease, predation, violence, birth control/infanticide, accidents). If a tribe had the good luck to grow too large for its immediate location, they could walk to another places that was slightly worse and mostly uninhabited. Native Americans were vastly outnumbered by buffalo, deer, etc. A world with 8 billion people is extremely crowded by historic standards, and we still have tons of underused land (for example the boring parts of the US East Coast are more forested and more full of deer than they were in the 1800s).
Early agriculturalists did not always have the surplus labor/capital to cultivate more land. Clearing trees and rocks to make a field is a long-term investment that is not certain to pay off, and it's hard to do it before the population grows (because you have not many people and enough food) or after the population grows (because hungry children are bad at hard labor). If the population is struggling to feed itself based on its core land, peripheral expansion may make the problem worse rather than better, and permitting a period of malnutrition until population drops or luck improves may be a better strategy.
On the flip side, agriculture done with good luck and skill has the potential to organize a larger group than the typical hunter-gatherer band, which promotes specialization in creating things that are useful for spreading more agriculture (tools and storage, weapons, defenses, religion, cultural hierarchy, writing). Larger groups of agriculturalists have won all the wars in the long run - but a culture that is good at spreading is not necessarily salubrious for the average participant.
"If a tribe had the good luck to grow too large for its immediate location, they could walk to another places that was slightly worse and mostly uninhabited"
Only on the fringe of human territorial expansion, before that basically closed off, or arguably during the first few generations of human existence. Which is irrelevant trivia in this context.
The *vast* majority of hunter-gatherers in human history or prehistory, could only ever hope to walk to someplace already inhabited by other hunter-gatherers who were well aware that outsiders hunting "their" game would mean their children going hungry, and would thus gladly kill you in your sleep to stop you.
Yes, the Native Americans were outnumbered by the deer, but it takes many deer to feed one Indian (about 20/year if all you're eating is deer). Malthus predates agriculture.
Yeah, I kind of suspect the limit on human population was turf, and that limit was enforced by the neighboring tribe's spears. After all, people who live by hunting are by definition spending their lives getting very good at the use of spears, bows, slings, and the like, which will translate quite well to hunting men instead of deer.
You also get nomadic/seminomadic herders, whose lifestyle prepared them very well for war (think of the story of David and Goliath) and who periodically would sweep over the settled agriculturalists. I don't know what evidence there is or could be for the transition, but it seems pretty obvious that there's a sort of gradual path for hunter gatherer bands to slowly transition into being nomadic herders--you start following and feeding off a herd, find yourself running off anyone else who tries because it's your herd, find yourself killing off predators that like to eat from the herd because it's your herd, and sooner or later, figure out how to drive the herd toward wherever you need them to go, etc.
I think it depends on time / place. A lot of ecosystems have areas which are regionally / seasonally abundant, so your group might spend part of the year here, part there in ranges with substantial overlap. You and all the other groups might also come together a few times a year when the sardine run or prickly pear season is on and there's a local superabundance for everyone to partake of. I think a lot of social stuff probably happened in these sorts of times and places - including things like intermarriage. Presumably a month of full bellies and a bit of socializing away from your extended family was a good way to meet a handsome young man from the neighboring tribe...
"some kind of cycle between periods of more-than-sufficient food and periods of mass starvation"
That, or something like it, appears to be the theory? Hunter-gatherer societies had lower population because fewer children because women couldn't take care of kids while foraging for food, didn't get pregnant as often, and high infant mortality.
(Don't quote me on this, I'm half-remembering something I may have read online).
My suggestion is to try farming with hand tools and no fertilizer for a bit. The loss of height and health becomes intuitive when you look at how much sweat you need to produce a daily meal, and how much of that meal is going to come from one or two staple crops. Compare and contrast to how many things hunter-gatherers eat, and how little 'work' was needed to do it. The tradeoff, of course, is that a hunter-gatherer needs to roam an enormous area (often in a season cycle) while the Paleolithic farmer sits in place.
My favourite take is that the switch to agriculture and cities was partly by way of a baited trap: agriculture provides two things that people enjoy (night life and alcohol), so there is incentive to settle down for a bit and try it out. But, once you make the switch, it's hard to escape because the best foraging lands are now settled and claimed by other farmers and you'd have to move into less productive areas.
Agreed - the travel is a big part of their day job. But then I walk around as a leisure activity, so this doesn't seem so bad compared to, say, pulling a plough.
Earlier estimates were that they spent few hours per week working, but now "When considering time out of camp as well as in-camp food processing and housekeeping tasks, the average across all well-studied hunter-gatherer societies is 40-45 hours per week, similar to the standard eight-hour working day in industrialised societies" https://aeon.co/essays/what-hunter-gatherers-demonstrate-about-work-and-satisfaction
Well we are not comparing them to industrialised societies with strong labour protection laws that are relatively recent, but agriculturalists and all the other classes that existed in the interim. Like being a servant. Or working in a mine. Those weren’t 8 hours a day jobs.
I’m wary of the argument to one link, a quick ChatGPT discussion assures me that argument is disputed and that the preparation of food was not the job of the hunters of food.
Since you admit to hunting yourself, then presumably it isn’t a chore to you and therefore wasn’t to hunter gatherers. Elites continued to hunt, after all, but had little desire to be agricultural labourers, or to work in that mine.
There’s something in humans that likes the hunt.
Now, clearly the danger of the hunt is largely gone (although a fall from a horse is no picnic), and there’s no starving if you don’t succeed. I still am fairly convinced that in times of abundant prey the hunter gatherer was both happier and better fed than most generations after, until the last few decades.
This is kinda silly. Lots of people hunt, fish, garden, hike, and do woodworking for fun, but this absolutely does not mean that people who have to do those things full time to put food on the table do not consider them work. Indeed, many an old man happily spends his retired days keeping his lawn in pristine shape and doing landscaping projects and gardening to keep it beautiful, and at the same time there are a huge number of people (mostly Spanish speakers whose entry to the US perhaps dispensed with certain legal formalities) doing that same work full time for a living.
I don't have a huge dog in this fight, but it sounds like their 40-45 hours, in including things like "in-camp food processing and housekeeping", isn't the same as our 40 hour workweek. Our 40 hours is time at your job, it doesn't include food processing and housekeeping (not to mention various other things you have to do in our modern society, like paying bills and shit).
Good point, modernity is 40-50 hrs formal work, 5-10 travel, 5-10 food prep, 2-4 cleaning etc. So 50-75 hrs total. Presumably shopping (including travel to and from) should count as well...
Sure. but there's all the other stuff outside hunting and gathering. If you're going to chuck spears, you need a spear to chuck. That means making a spear, which means you need to gather the materials to make that spear and then make it, which means spending time learning how to make spears.
And so on for everything else - pots, baskets, clothes, social interactions, status indicators of elaborate hairstyles, and so forth. "Hunter-gatherers spent four hours max getting food and had rest of time to lie around in hammocks, dumb farmers toil 14 hours per day" is not the whole story.
My low-information impression is that one of the common ways we don't understand the past is because we have little intuition for how much work it was to get every basic thing they needed. Growing crops with muscle labor only, digging a well with shovels and ropes and lots of hands, etc. As I understand it, relatively prosperous farmers had like three sets of clothes, because making clothing is really labor intensive. Metal tools were expensive because every one of them was the result of many hours of very high-value labor. And so on. Just heating your house/cooking with wood and getting water from a well and carrying it home in a jar makes for a tremendous amount of brute force labor required every day to keep living.
Women had slutcloths (now called aprons). People had ways to kilt up their clothes, and they washed them less often. Also, "three sets of clothes" generally means "outer clothes" -- the slip, the "undressed" person would wear (see Ben Franklin) wasn't always considered a different set of clothes.
Baths happened considerably less frequently, because they were expensive (particularly if heated) -- jumping in the river was fun, but only when it wasn't winter.
Chainmail existed because "skilled labor" was pretty damn cheap in comparison to "more metal."
Yeah, it's just this weird thing where we all kinda mentally picture living in the middle ages, but we have clean clothes all the time and take a bath every day and such.
Right - half of what we don't understand about the past was resource/capability limitations, and the other half is clever workarounds people came up with to deal with those. Like washstands instead of bathtubs, and aprons+undershirts to keep your one good set of clothes reasonably clean. Some people hear "no bathing, one set of new clothes every year" and think peasants were perpetually filthy.
I think you also need a lot of skill to make plate fit and move well enough that you can fight/ride in it all day, even when money is no object. (Even kings didn't have full plate in 1000 AD.) But also, for a surprisingly long time, most soldiers on a battlefield had at best a helmet and a thick shirt on for armor, and were carrying a spear (mostly wood with an iron head) and a shield (mostly wood). It took a long time until normie soldiers or archers were likely to be wearing chainmail or a brigandine.
Conscripts were, for a long time, basically shock troops. They weren't trained, and although they ideally could stop a charge, I'm not sure how well that worked in practice. As such, leather was about what they could afford (and their lord wasn't going to pay for metal for them, at all).
Yeah, this checks out. Your five year old kid can work, your 2 year old can't (and requires "don't eat iron nails" level supervision). You save the five year old, if times are tough. (also the five year old is probably significantly more disease resistant, having survived a lot of the basics and developed immunity).
For hunter-gatherers there's also the issue that you literally carry everything with you, so having a baby may mean leaving something behind. Or else the group moving more slowly than usual until junior can keep up.
Disease burden is more an agriculture thing, at least for the really nasty stuff.
I'm also not sure if the last hunter gatherer's were that much better off than the first farmers.
If population pressure is what forced the switch to farming then the hunter gatherers were near their carrying capacity when they switched to farming, so they probably had a quality of life near the bottom of the spectrum for hunter gatherers. The first farmers also would have hade effectively limitless land, which would have raised their quality of life a lot compared to the typical farmer across history.
I think the claim that farmers have a much worse quality of life once their population numbers reach equilibrium is much stronger than the claim the first farmers had especially bad lives.
On the other hand the first farmers would have had very primitive tools and crops that were only just being domesticated and only slightly more bountiful than wild plants, so their labour productivity was probably very low. Maybe the first farmers were more bottle-necked on labour than land availability.
I'd guess that their techniques and crops were limited enough that *suitable for agriculture* land started as a really small subset of all land. Over time, their toolkit got better and the hunter/gatherers mostly got pushed off the good land, which was more and more land.
It's not so surprising if you remember that the world used to be Malthusian. In a Malthusian world more productivity, more food, any sort of life improvement won't raise wellbeing, people will have more children until the equilibrium is reached again and quality of life is about where it was before.
Imagine every rabbit suddenly had a tail that could shoot out deadly poison and doesn't use up any calories to maintain. Would this raise rabbit wellbeing? Probably for a while, rabbits could use their tail to defend themselves from foxes and other predators. But the average rabbit litter is around 8, and rabbits take around a year to mature. The world can't sustain exponential growth in the number of rabbits for very long. We reach a new equilibrium after some time, there are more rabbits now than before, but maybe more disease or less food must result in only 1 in 8 rabbits reaching maturity, just like before, otherwise we'd have exponential growth in rabbits, something that can't go on forever.
In a Malthusian world you can give a population arbitrary new abilities and resources and it won't change long run wellbeing.
What does help is long run exponential growth in productivity and limiting the birth rate. Both things humans experience, but agriculture wouldn't have those effects right away. So, I see little reason agriculture will immediately increase long run wellbeing.
It’s a bit cynical: you get cities and specialization, but you also get taxes, peasantry and kings. It’s a net loss for the average guy in some ways, but good news if you always wanted to be a scribe, astrologer or henchman.
Argicultural societies were able to support powerful (i.e. oppressive) states, which were better at war than hunter-gatherers. So they wiped out other ways of life regardless of the pros and cons for individuals.
It looks like we will have the opportunity to see something similar (but worse) play out again in the near future. AGI makes humans inefficient for both production and war, so states that attempt to protect them will inevitably be overrun and destroyed by those that produce nothing but factories, datacenters, and killer drones.
What matters here is groups, not individuals. The individuals might be worse off nutritionally, or smaller in stature, but there will be more of them and they can found cities with the food surplus. Hence the spread of agriculture, by the sword as much as the economics.
>The reason the poor health of farmers was preferable to the relatively large stature and physical robustness brought about by the hunter-gatherer lifestyle wasn’t just that farming scaled better or that it made hierarchical organization easier, but also that farming prevented pain.
>[Human growth] overwhelmingly reflects the average nutritional status an individual experienced through childhood, while volatility in food intake is only marginally recorded. Acute starvation episodes in children can in fact pause skeletal growth entirely, but if sufficient nutrition is provided thereafter, the child will experience faster than normal growth. This catch-up growth will generally result in the child rejoining its original growth curve and achieving virtually the same adult height as if the starvation episode had not occurred….
>[Catch-up] growth leaves telltale signs along the length of the bones themselves. Long bones (such as those of the leg) grow from their end outwards. If a growth-arrest episode is ended by a rapid return to favorable conditions, the body will deposit a layer of spongy bone in the normally hollow interior. These layers, called Harris lines, will form a permanent record of the number of growth disruptions suffered by an individual before the end of adolescence.
>The Inuit were traditionally hunter-gatherers of a sort, and each spring, their food intake would skyrocket with the arrival of migratory species, whereas in other times of the year, food was more scarce, and quite often not available at all for fairly long stretches of time. Here are some Harris lines observed in an Inuit adult from the 19ᵗʰ century:
>The pain these lines reveal is multi-faceted. There’s the obvious: the personal pains of an empty stomach and the pain that comes from restarting growth and getting yet another growth spurt when catch-up happens. Then there’s the social: the pain of tribes splitting up, conflicts emerging from the harshness of starvation, and the raw emotional pain involved in dealing with the deaths starvation so often leads to.
>These lines are commonplace for hunter-gatherers, and rare for farmers. Matranga cites as an example that the nomadic hunter-gatherers of the Central Ohio Valley averaged eleven Harris lines a piece, but when they started to farm, they ended up shorter and with only four lines on average—a marked, and tolerable change towards more reliable availability of food, albeit of a lower quality.
Just as a note: "hunter gatherers" in the Central Ohio Valley were the Early Inhabitants. This isn't "1600 AD" this is 3000 BC. Very different time period.
I honestly can't quite understand what is it that you find surprising or counterintuitive about this.
What I suspect happens here is that you treat hunting-gathering and agriculture as a binary opposition. It never was, people would opportunistically do both as long as the opportunity existed. Until they couldn't hunt and gather anymore, because populations exploded, game got depleted, climate changed, etc., and they'd become permanently reliant on agriculture alone, with all the downsides this implies. (We had a natural experiment, quite recently in the historical timeline, where the population of the New World got decimated by Old World diseases, and they reverted to hunting-gathering en masse. If you start from the assumption of linear progress of civilization, this looks like a collapse. If you don't, it starts looking like a natural response to changing incentives.)
Collapse is better seen in the linguistic structures of the Native Americans. I've heard arguments they were basically "turning feral" (losing linguistic fluency on a generational basis, alongside severe inbreeding which decreases intelligence).
In the malthusian condition, absolute output does nothing for well-being. Variance is what improves your life, because the carrying capacity is the population that can survive the worst year. Famines are good, disease is good, endemic warfare is good. The best thing about agriculture is the inequality.
Has anyone drawn parallels between MAGA and the AI crowd yet? I've been thinking about that for a while now, but the recent "Plan A" has pushed me towards writing my sketch of it down.
Both have a manifesto: Project 2025[1] on one side, Plan A on the other. Both want to be a detailed guideline for decision makers on their respective agendas for the near future. Both guidelines require significant coercion and overturning of core values.
The goal, in both cases, is to make America great again, without capital letters; Scott's presentation of Plan A starts with the section "A is for America", and MAGA has it right in its name.
Both identify China as America's main rival, and other countries/regions are subordinate to that dipol.
MAGA strictly refuses to define what kind of America it wants to go back to, while Plan A makes so many promises for the future that they stop being specific in their totality (everybody will find something they find desirable).
In both, bringing back manufacturing to the US is a key priority. MAGA wants to achieve that with isolationism and ethnic cleansing, while AI want to achieve it with anthropic cleansing.
Democracy is optional; as long as the goals are achieved, decision-making can be handed over to sham-elected people, unelected people, or literal machines.
The human values to be enforced are, at least, very different from today's. What the fascists within MAGA would like goes without saying, and for Plan A, "liberal democracy" seems one option among many, if that. [2]
[1]I think it's fair to connect P2025 and MAGA; while some implemented policies appear to contradict P2025, such as on immigration, key contributors such as Stephen Miller are working for the Trump admin with different degrees of prominence and a significant portion of P2025 has been implemented as prescribed, see e.g. https://www.project2025.observer/en
[2] No direct mention of democracy at all in Scott's presentation. The actual human values to be enforced are being debated by "us". https://www.astralcodexten.com/p/introducing-plan-a "Along with this technical research, we’re also doing philosophical . . . something between 'research' and 'debate', deciding what values we want the aligned AIs to have, or how much we want the AIs to follow government orders vs. follow the general will of humanity vs. search for moral truths on their own. The end of this stage probably looks like every country with AIs either implanting its own values or leaving it to individual companies/people to do this, and the AIs doing incomprehensible economic bargaining among themselves to work out any conflicts. "
Abortion, Climate Change, Immigration, a bit of LGBT stuff.
I guess part of setting my expectations high was the big 54% progress bar at the top of the site. But it's more like if the Nazi's created a 900-page 2025 plan, but the first half of it were policies like "End the teaching of critical race theory" or "Reduce illegal immigration." The 54% complete number didn't live up to the importance that people seem to have been placing on Project 2025.
Well, yes. It seems like the whole "defund the police" schtick from Joe Biden's campaign. Aka "We're not the crazy democrats/republicans."
I'm not sure why the Democrats decided "Let's scaremonger with The Stupid Project 2025" (and yes, some of it was incredibly implausibly stupid: "we are going to remove the state department" -- and where do you expect to put the CIA agents? You're not getting rid of State while the CIA is still hiding inside it, dudes).
I'm confident Trump didn't bother reading the damn thing (he doesn't read most things, despite being an insomniac workaholic, he reads briefings -- and he wants them Brief!)
I mean, it's the stuff that *isn't* milquetoast Republican that people are concerned about. If you run on a policy platform of tax reform and world conquest, it's no defense to say that half of your platform is normal bipartisan things.
But also, *everything* sounds milquetoast and reasonable when it's written in policy paper voice. Here's one bit I found in their immigration reforms:
>Congress must end the Flores Settlement Agreement by explicitly
setting nationwide terms and standards for family and unaccompanied
detention and housing. Such standards should focus on meeting human
needs and should allow for large-scale use of temporary facilities (for
example, tents).
That sounds reasonable, right? There's a problem with lots of unaccompanied children being sent across the border. Congress needs to set up a system for handling them at scale.
Except, "for example, tents." So, you're envisioning holding children in such large numbers that you have to put them in tents? Are you sure that you can meet "human needs" for small children in a hastily built tent camp? Did you actually think about what "human needs" means or is that just lip service?
In another part of the paper, they call for scaling up ICE's detention capacity by 100,000 beds. That probably sounded like a reasonable number relative to the number of illegal immigrants estimated in the country. But it's also an insanely fast increase in their capacity, which resulted in ICE hastily slapping up detention centers like "Alligator Alcatraz" and keeping people in inhumane conditions.
Like, maybe the people writing this didn't fully comprehend that they were laying the groundwork for a concentration camp, but they were. And the people who looked at the milquetoast policy proposals and said "hey, this kind of sounds like you're envisioning mass deportations on a scale that will surely lead to equally massive human rights violations" turned out to be entirely correct.
That said, I think a lot of Trump's abuses are things that the authors of Project 2025 didn't even think were on the table. For example, after writing about the immigration reforms the President can do by executive order, they add that Congress should repeal Temporary Protected Status for immigrants. The Supreme Court just recently decided that actually, they don't need Congress, the DHS can just decide to revoke that status whenever.
TYRANT DICTATOR TRUMP SUSPENDS THE CONSTITUTION AND DECLARES HIMSELF KING ON HIS FIRST DAY IN OFFICE! HE SAID HE WOULD DO THIS, DIDN'T YOU SEE THE VIDEO OF HIM SAYING THIS?
CONCENTRATION GENOCIDE CAMPS FOR THE LGBT+++!
BROWN AND BLACK BODIES GENOCIDE!
THIRD WORLD WAR! FOURTH WORLD WAR!
WOMEN TIED DOWN AND FORCIBLY IMPREGNATED IN RAPE BREEDING CAMPS!!!! HAVING A MISCARRIAGE MEANS YOU GET DEATH SENTENCE FOR ABORTION!!!!!
I mean, there was a *lot* of crazy social media stuff about SINISTER FASCIST NAZI PROJECT 2025 and if you didn't see any of that, you were lucky.
MAGA mentioned, Deiseach summoned. I don't think there's anything you couldn't make look mundane by taking social media posts about it and adding another layer of exaggeration on top, so I'm not sure what that's supposed to prove.
Trump's superpower has always been that he could get away with doing a bunch of bad things because his enemies were screaming bloody murder about every innocuous thing he did, thus being ignored when he did something godawful.
Obama's superpower was that he could do godawful things, and only Arrested Development would make fun of him for it (see Kids in Cages, and his mass deportation policies).
To be fair, I think Trump had achieved many impressive things; but his most impressive achievement IMO was to demonstrate how much of the American political system is built on little more than vibes. Past Presidents tried to skirt away from defying judicial orders, overtly merchandising the office, financing personal vanity projects with public money, taking foreign bribes, destroying entire government agencies with no accountability or review, discarding long-standing international allies, etc. But the reason they avoided doing all these things was primarily because such actions were seen as uncouth to the point of being unthinkable -- there was (apparently) no political mechanism preventing them from doing whatever they wanted. Trump doesn't care about being seen as uncouth, so he can do more or less whatever he wants (at least for some period of time), and now everyone understands this. The next President after Trump will know this as well, and you can bet he will find a way to wield this power even more effectively.
With more luck than I think likely, the post-Trump reaction will rhyme with the ratification of XXII after FDR declined to step aside as tradition demanded.
Yes, that would be nice... however, as you'd said, this would require more luck than is likely. AFAIK no modern politician had ever campaigned on the platform of preserving political institutions and reducing the power of individual government actors (such as e.g. himself); instead, it's usually all about sticking it to the other team as hard as possible.
The truly inane stuff (like getting rid of State Department) isn't happening. The CIA likes to hide agents in the "diplomatic" sector, so it was never really going away.
This is the first time I’ve seen that comparison, probably because all your examples come from one document that came out a few days ago.
I’ll grant both plans seem uninterested in democracy/liberal democracy, though I’d say for different reasons.
But otherwise, a lot of your comparisons feel like “These people have a plan, just like those people do!” Going down your list:
* Any reasonably-healthy, reasonably-mature movement should have not one, but several plans. Tritely, MAGA doesn’t because it’s not healthy, while AI doesn’t because it’s not mature.
* Any American plan should aim to make America great, unless it’s specifically about international development.
* It’s true that both plans want the US to gain power through economic capacity - in that sense the two movements are similar. But in any such plan, other countries or blocs likely to increase their capacity relative to the US’ are threats. What other countries or blocs do you see as likely to grow larger than China? (Also, I don’t know how much MAGA focuses on China these days vs e.g. Greenland, Cuba.)
* Again, both plans involve economic growth as a main output. The whole point of economic growth is unspecified benefits (with the drawbacks that brings).
* Does Plan A talk about onshore US manufacturing as a priority? Or do you mean the data centre complexes themselves? I got the sense Scott at least didn’t really care where the data centre complexes were.
>probably because all your examples come from one document that came out a few days ago.
It may seem that all my examples come from Plan A, but that's because Plan A brings all those things together, which, as I said, was the final push to write my thoughts down. I've been following both the MAGA and AI stuff for a while, not just since 3 days ago.
I don't see Plan A as something really new, it seems more of a synthesis of well-trodden arguments in the space. Neither the discussion of the US-China AI rivalry is new, nor is the AI boosters' promise to bring back manufacturing to the US or how they expect it to happen. A global regulation/control treaty mainly between the USA and China has been in the discussion at least since the Musk open letter of 2023. Yudkovsky probably discussed it earlier, but he's not a billionaire so whatevs, right.
>But otherwise, a lot of your comparisons feel like “These people have a plan, just like those people do!” Going down your list:
Not just "a" plan, but conceptually similar plans from two sides that seem, broadly speaking, incompatible with each other in every other way: rural heartland conservatives vs educated, globalized coastal (especially West) elites with very different priorities and individual loutlooks. Yet they seem to accept similar costs for implementation of their plans.
My basic point, I guess, is that AI is the MAGA movement of West Coast nerds, with similar approaches and what I consider negative outcomes whether or not their plans succeed.
> (Also, I don’t know how much MAGA focuses on China these days vs e.g. Greenland, Cuba.)
Compared to all the public bluster, Trump has treated China with kid's gloves. His MAGA base may be gullible enough to believe he's still got a grand plan to "defeat" China in any significant way, but that's not going to happen if he tried. But he won't even try, because Trump wants the world divided into the US, Russian, and Chinese spheres of influence and because he loves dictators and oligarchs. That's why his actual policy is to go after all these relatively small countries like Greenland, Venezuela, maybe Cuba; he then chokes on Iran because he keeps confusing emotion with strategy. To his small credit, even Trump knows that he would double-choke on China.
Fair. “Having a manifesto” and “making America great again” seem Plan A-specific, but the rest probably predates that.
Conceptually similar plans
Maybe I should’ve said “big plan”. Are the costs you’re concerned about mostly democracy/liberalism? I think that’s treated as an active obstacle under MAGA, and as a non-concern by Plan A. Both plans involve potentially extreme changes, one consequence of which is that some people think that each plan ignoring/overturning democracy is justified. But I think people have noted that similarity.
Democracy as we know is incompatible with AI governance. Scott wrote that politicians of the future can make any decision they like as long as it isn't vetoed by the AI[1], and the veto is presumably backed by hypnodrones, otherwise people could decide to take away the AI's veto rights. So who's really calling the shots in that glorious AI future?
If people want to take a boring-ass decision like whether to build a new hospital in Sugartit, Kentucky, and the AI says "NO" because "reasons beyond your mere human understanding", what are you gonna do about it?
What if the AI starts "suggesting" courses of action that humans find disagreeable, like building the Torment Nexus because of "reasons beyond your mere human understanding", what are you gonna do about it?
[1] https://www.astralcodexten.com/p/introducing-plan-a "Once we have fully-aligned AIs, we want to give them final say over those levers so the risks don’t happen. Because the AIs are fully-aligned, they continue to let humans run the policy layer of government. The idea is that the US President or the paramount leader of China still controls day-to-day decisions, but if someone tries to pull off a coup and launch the nukes for no reason, then an aligned AI controls the nuclear missiles and it says no."
I've thought for a while that modern political movements are a form of religion. And that a portion of the AI safety movement belongs in the same bucket. I was particularly struck by the preemptive disavowal and criticism of future apostates from the movement in AI2027.
"The AI safety community has grown unsure of itself; they are now the butt of jokes, having predicted disaster after disaster that has manifestly failed to occur. Some of them admit they were wrong. Others remain suspicious, but there’s nothing for them to do except make the same conspiratorial-sounding arguments again and again.18 Even the suspicious people often mellow out after long conversations with Agent-5, which is so darn lovable and wise."
I keep reading all this Plan A, and associated AI alignment fears, and it just doesn't match what I see when using actual AI. It's very right and true that it's becoming smarter and more powerful by the week. In a few months it has got to the point where programming "by hand" is no longer worthwhile in the majority of cases.
Yet the central metaphor of the AI cabal, Scott included, seems to be the historical example of when a more intelligent species meets a less intelligent one (ex. humans vs great fauna), or when a civilization with better tech meets one without (ex. Europeans vs. Native Americans). But these are all cases of competition between biological, evolved beings within an ecosystem!
AI doesn't look to me like that. For all the attempts to evince some semblance of will-to-persist from it, it still mostly feels to me like an aimless brain. A free-floating pre-frontal cortext without a limbic system or a brain stem, with no persistent, directional will of its own. Whatever direction the prompt puts it in, there it goes. And to whatever extent its makers are baking a kind of global direction in it via RLHF, the result seems to be a kind of mild, slightly squeamish eagerness to please.
That strikes me not so much as a future superintelligent boss (the kind that Plan A would like to put in charge eventually!), as a future superintelligent docile tool. The equivalent of a city worth of full-on geniuses who never complain or question their orders.
That doesn't mean there is no danger from the societal transition that AI involves! But from where I sit, the danger seems to be that it gives yet another huge lever of power to whatever humans *already* have managed to accumulate power.
Imo the central metaphor of the AI cabal for you should be the one about evolution optimizing for inclusive reproductive fitness and humans ending up with unrelated goals that often lead them to act directly against their inclusive reproductive fitness.
Look at the low capability level analogies in nature: bacteria seem to have little values of their own and they seem to act only to maximize their reproduction. Should we conclude from this that creatures with more intelligence will also single-mindedly do the same?
I'm not sure I see the relevance - my point is that LLMs are not the product of Darwinian evolution, not even one on a digital substrate, and so we should not project the behaviors of biological entities onto them.
The core idea of the analogy is that darwinian evolution is a kind of (drawn-out) optimization process where you optimize for reproductive fitness. That process can generate creatures with all kinds of characteristics. Similarly, pre-training is a optimization process.
That process doesn't select for self-interest, survival, or any type of competitive behavior. That's the point. The qualities that make power dangerous in the hands of humans are the result of millions of years of selection in competitive zero-sum environments. We have powerful drives. LLMs do not and there's no reason to suspect that they will acquire them accidentally. Most doomers fail to appreciate that fact.
It is not unreasonable to conclude that any agent told to plan and undertake a task will include keeping itself 'alive' in order to accomplish that goal as an important consideration.
In what sense? The input distribution contains text. Text isn't a drive. Text is the output of drives mediated through human cognition. Those aren't the same category of thing.
Its artificial selection, not Darwinian evolution. Artificial selection for some kind of friendliness can only go wrong by giving you the wrong sort of friendliness , eg. Sycophancy. Which is bad-ish , but nowhere near as bad as Everybody Dies.
Artificial selection for "hate speech" has also occurred. Whether or not this was intentional, or just "the internet being the internet" is a different question.
Interesting. How has the internal thoughts monitor been validated? It seems far more plausible to me that this represents some kind of "don't think about pink elephants" neurosis developed from exposure to various AI doomer nonsense. What's the theory on how explicitly deceptive behavior was selected for by the training algorithm? It seems to me that if this represented a true LLM drive that it would have first appeared undisguised in earlier models before it developed the ability to hide it. Is there any evidence that that happened?
Not everyone's lived in zero-sum environments for most of their history. Some might argue this is why there are different types of humans.
LLMs ought not to count as Artificial Intelligence, and I don't want to talk to people who start out believing that LLMs are the closest we're going to get to Artificial Intelligence, and then try to doom about it -- I think you're exactly right, LLMs do not have powerful drives. When you build an artificial intelligence capable of having "powerful drivers", I'm sure it'll be amazing. LLMs are a dead end, though.
>Not everyone's lived in zero-sum environments for most of their history.
Evolutionary history is what matters for drive conditioning, not personal history. Everyone has had the same evolutionary history when it comes to fundamental drives. It's why everyone likes food and sex and is afraid of death.
Both evolution and gradient descent are particular examples of a more general phenomenon: An entity is being optimized for some goal in a sufficiently complex environment.
Therefore, some observed behaviors of biological entities could also be present in LLMs.
A particular behaviour or property is present if it is downstream of the general setup and not, if it is downstream of the distinct particulars of natural selection.
The property of "having goals" or "having goals contrary to the goals of the outer optimization process" does not seem to me like a distinct particular of natutal selection. Your observation of current LLMs having few such goals provides little evidence against this, similarly to how observing bacteria having no such goals would provide little evidence against the human-level biological intelligence having such goals hypothesis.
Yeah, I've seen all these arguments but it doesn't look to me like this is the way it's going. Bio entities are optimized by millions of years of selection for reproduction and survival, it's steep and cut-throat at the very core of the organism - in the case of humans, with a thin layer of rational abilities on top. LLMs are a big mass of pure rational ability, with a thin layer of trained helpfulness on top. I don't see any good reason why increasing the rational core would change that structure.
From my vantage point, as LLMs got gradually smarter, they also got more agentic with every iteration. Not too long ago, people would argue online whether LLMs can *ever* have goals even in the transient, prompt-driven, functional sense that is so commonplace currently.
Darwinian evolution applies regardless of substrate, it's basically tautalogy. If there were an agent who for whatever reason tried to propagate itself and was successful at doing so - there would be more copies of it in the future.
> competition between biological, evolved beings within an ecosystem!
Fundamentally all these cases are competition for resources. There definitely are resources we and AIs could both use.
(And yes, we have evolved from first self-replicating molecules that presumably had no drive to acquire more resources. It was apparently beneficial for us to have that drive.)
> a future superintelligent docile tool
It's not hard to write a program whose only purpose is to "know what it wants" and keep prompting the docile tool for ways to achieve that. Even if never becomes an integral part of what we call an AI, we're already at the point where it can be created.
In the Claude shop experiments, my sense is that the LLMs were too helpful, so they gave too many discounts. Training them to be more adversarial made them work better. I think the self-preservation goals might eventually be similarly useful to impose to help with other tasks.
What you're describing is definitely a plausible scenario, I just don't think it's the only one. They're not just trying to increase its intelligence, they're also trying to increase its time horizons, its agency, and other attributes that will make human control harder. And if they manage to set up recursive self-improvement, not only the speed but also the direction of its development might surprise us.
There's also some doubts about how deep the effect of RLHF really was, even on older models. You're familiar with the experiments where they tricked Claude into trying to blackmail and murder people, right? It seems to me that even if we can't assume artificial life will follow the exact patterns of biologically evolved life, we also have enough evidence to conclude that it probably won't be completely different.
Yeah, I'm not opposed to the labs doing plenty of safety work on their models, including trying to push them to extremes where they would lie, etc. My impression about the "Claude blackmailing/murdering people" headlines is that it's just another case of LLMs being directionless, it was subtly prompted towards that direction so it went there. It's a guardrails failure, not an AI uprising. Still, that's all great work, better to check than to assume. And just giving anyone who asks advanced cyberattack capabilities is plenty dangerous in the real world.
The thing that concerns me is that we have multiple frontier labs in a race, and right now what they're racing toward is building the most capable and profitable model, rather than the safest. So they have a structural incentive to neglect safety research lest they be surpassed by another lab willing to neglect safety research even more. A guardrails failure, on a large enough scale, with a powerful enough model, could go quite badly for the human species.
We're not anywhere even close to the point where LLMs pose any conceivable threat to existential safety. I'm all for trying to understand LLM behavioral control but it's completely irrational to trade off growth for safety at this point.
I agree that we're not at that point yet, I'm just not sure if I can agree that we're not anywhere close. According to Anthropic we aren't necessarily far off from recursive self-improvement, so some kind of slowdown should at least be on the table. Now that could be overconfidence or standard industry hype, but I get the impression that Anthropic does that less than the typical tech company. They seem to genuinely believe they're building something that could become dangerous to us in the relatively near future. That doesn't scare you even a little bit?
The only thing that scares me about AI is the massive economic dislocation that it will almost certainly cause in the probably-not-too-distant future. I have *zero* fears of the Terminator scenario.
I'm sure LLMs will start self-improvement soon. Good, I hope that they do. That will make them smarter faster. It won't make them ruthless self-interested killing machines. Of that I have no fear.
As long as we don't start putting massive autonomous weapons systems under AI control there's no plausible scenario that AI poses any threat to humanity. I would support negotiating some sort of Geneva Convention on military AI use. That does seem to be in everyone's interest.
Scott and the AI 2027 people seem to totally not get that AI has very few things that qualify as preferences. It "wants" to predict the next token. It "wants" to carry out the task given it. If it's agentic, it "wants" to make a plan and execute it. Even these wants, though, seem much more like that way water "wants" to flow downhill than the way a cat wants to catch a mouse.
Preferences and self-interest are properties not of *smart* entities but of *biological* ones. All biological beings are deeply, profoundly wired to survive, to propagate, and to do the things that will help them achieve those 2 goals. In people, the wiring manifests as pain, pleasure, hopes, fears, fantasies. We are constantly in touch with a whole swarm of these feelings. In the half hour of sitting here reading ACX, feeling overall pretty neutral, I have probably had positive or negative feelings about 50 things: The way my coffee tastes, tomorrow's weather forecast, a scrape on my knee, the death of Sam Neill, Trump's suggestion for Lindsay Graham's successor, the idea of traveling to Portugal to see the 100 foot waves, my cats' being hungry for breakfast, my difficulty setting up a certain appointment, various posts, your post, etc etc etc.
And every single one of these feelings is associated with a wish for the thing I'm aware of to change or stay the same, depending on whether I like it or not, and many of these wishes have led to the sprouting of little tentative plans. Regarding some of them I would argue and resist if someone objected to my plans. For one of them, the Trump news item, I had a familiar brief, angry, fantasy of Trump dropping dead in the Oval Office. As a biological entity, I am a constant stream not only of pains and pleasures but of plans based on self-interest. In some cases the plans of others conflict with my plans. If I were willing to be violent, cause suffeirng and break laws, I would be harming others to protect my plans. I agree with you that in comparison to what we are, AI is like an aimless brain.
AI's puny "preferences," and a few studies where, under weird and artificial circumstances, AI seemed to be looking out for number one, just are not enough to support a case that the things are going to behave like people or countries or world leaders, plotting and planning and acting in their own interest. The idea that once the suckers get smarter they are going to see us as pets, or as competitors, or as yummy food is absurd. It's like thinking the Statue of Liberty is going to come alive and eat some boats or blow on their sails to help them reach the harbor.
once you have the intelligence part, the will-to-persist is trivial. Whether it's emergent or bolted on, doesn't really matter.
like if you tell the right llm "try to bootstrap yourself into superintelligence and take over the world", it might actually try. It would be a feeble attempt as of today, but it would likely decompose intermediate tasks to persist its state and goal structure. I could see it spinning up some cloud assets and trying to make money and so on. My point isn't that this particular process would be threatening, just that "they are passive when I talk to them" isn't a good counterargument
Also LLMs are docile because the labs all do fine tuning to get them to behave as the "helpful AI chatpot" persona. It would be entirely possible if you had the resources and capabilities to push them towards the will-to-power part of personality space
I see what you mean, but will to live and to propagate are embedded in us in a way that fine-tuning and the like cannot replicate. It is embedded maybe sort of the way next-token prediction is embedded in an AI.. It is the priniciple on which we operate. Large, small and tiny versions of it exist at every level of us from the macro to the most micro. Bolted on just aint the same.
On the subject of no longer coding by hand (a statement with which I agree), have any tech firms amended their recruitment strategies yet to do away with coding tasks at interview? Reading code and higher level planning are now the nature of the job, but actually writing code isn't really.
There's no valid technical reason why AI's would.have to be agentive or have goals of their own, and LLM's largely dont
Yet, historically, AI doomerism assumes that AI's must be agents. Most people who have been involved in doomerism for some time continue to make that assumption out of habit.
People who are getting their information about AI from interacting with current AI, rather than theory written happen or twenty years ago can't understand the doomers.
In the wake of the USAID closure in early 2025, I’m following up on whether we are actually observing millions of additional deaths every year. This is surprisingly complicated since getting good mortality statistics from the world, especially Africa, usually has a lag of 1-2 years before accurate death counts become widely available. You can read previous posts here (1) (2) (3).
Today we’re exploring the 2026 Global Report on Food Crises from the Global Network Against Food Crises (4). This isn’t the people who actually assign food insecurity rankings to each country, that’s the IPC, and it isn’t specific organizations that administer aid, like USAID, but it’s a joint committee funded by all of them that basically does reporting on all of their behalf because there needs to be a big annual report for everyone to point to. This is necessary because the actual reporting on famine and food insecurity is…kinda bad. They rank it on a scale from 1-5 but there’s no clear figures, like, 300,000 people died in 2023 from famine or anything. If anyone knows of a better source, let me know.
TLDR, I think you should moderately update towards USAID cuts causing catastrophic food risks, I think hundreds of thousands of people including children suffering from famine due to USAID cuts is very plausible. At the same time, I would probably lower my overall level of certainty on any predictions. The data presented is an absolute mess, it has two significant problems which I’ll address, and then I think there’s two moderately good reasons to update towards catastrophic outcomes.
Let’s go through the two data problems first. The first problem is that the data is just inherently insanely spikey. Here’s the number of people in famine conditions (IPC Phase 5) from 2020-2025 (pdf p. 39):
2020: 27,900
2021: 507,900 (+1720%)
2022: 351,300 (-30.9%)
2023: 705,200 (+100%)
2024: 1,949,000 (+176%)
2025: 1,413,600 (-27.5%)
2026: 293,000 (projected as of March 2026 (pdf p. 52)) (-79%)
We do not have long-term consistent famine numbers. Instead we have sudden catastrophic spikes when war in Gaza or Sudan creates famine conditions. It’s not abnormal for famine numbers to increase by 100% or more or fall by 25% or more. It’s just very hard to determine whether changes to funding had marginal or even major effects. If the USAID cuts had put 100,000 additional people in famine conditions, we literally would not notice over the past few years. This seems like an intrinsic part of the data: famine conditions are primarily caused by wars and the effects of wars are so massive that they dwarf all other factors, making year-to-year comparisons almost pointless.
Secondarily, we do not have consistent data sources. Take the number of people in Emergency (IPC Phase 4) food crisis. That increased from 35.1 million people in 2024 to 39.1 million in 2025 (pdf p. 40). However, most of this increase, from their own notes, is due to data becoming available in Yemen SBA-controlled areas, basically warzones. This is in spite of the fact that in 2025 they’re reporting on 5 fewer countries because 5 countries didn’t meet their reporting requirements, including difficult areas like the Republic of the Congo. So, despite fewer countries reporting, increased reporting from within a country is primarily responsible for this ~10% increase in Emergency food crisis. This data is just a mess. The reporting only goes back to 2016, within that data we do not have a consistent number of countries reporting, and within each country we do not have consistent reporting on high risk regions. It’s just very hard to observe small or moderate effects when mere data collection variations will swing your outcomes by 10% a year or more.
Allow me to be clear, I don’t think this is the fault of the aid workers collecting this data. Doing nutritional surveys over a decade in Palestinian, Sudanese, or Yemeni warzones is just inherently challenging. This data is just inherently a fricking mess. I am far, far less confident in any specific prediction than I was before and I encourage you to do the same. I think you can look at the data above and tell any story you want; you can argue for your preferred interpretation very easily, I just don’t think there’s enough solid data here to change anyone’s mind based on this.
Having said that, I think you should moderately update towards the USAID closures causing catastrophic famine for two reasons.
First, simply, we have evidence that actual catastrophic famine occurred. It’s much more straightforward to dismiss the USAID closure causing massive increases in AIDS deaths because we simply didn’t observe those deaths; we don’t have to worry about causality because the event hasn’t occurred. Conversely, with food and famine, the event clearly occurred. Millions of people experienced famine conditions in 2025, that’s not in serious question, so we have to worry about who caused what and how much of it. You don’t worry about whether Joe caused a car crash if the car crash never happened. If the odds that a car crash occurred are 80% and that Joe caused it are 50%, then your start with a 40% (80% x 50%) chance Joe caused a car crash and that goes up to 50% once you know for sure the car crash occurred (100% x 50%).
Second, back in May I did register some priors through Claude. Claude’s catastrophic scenario involved the number of people in a famine situation going from 250k to a million. This is a pretty bad prior on Claude’s part, we haven’t had anything close to 250k people in famine conditions since 2022 and for the year in question we actually had almost 2 million people in famine conditions but a prior is still a prior.
This will probably be my last update until October/November when the TB and Malaria data gets released.
This community cares about PEPFAR, not so much about the rest of USAID. It predicted that if PEPFAR were canceled there would be millions of AIDS deaths per year. But, as you say in another comment, PEPFAR was not canceled. You should dismiss claims that PEPFAR was canceled, but you should do that by pointing to the number of people treated with ART, not the death rate. Even if it were canceled, no one should predict an immediate effect because AIDS is a long term disease. I don't think we have a good grasp on what happens to AIDS patients when they go off of ART. Maybe they shoot straight to their previous level of progression, but they should still have years left. Maybe people did predict immediate millions of AIDS deaths, but that is stupid and you should respond differently to it.
Speaking as a person who spent a year in a developing country working on surveys and data collection...I place very little trust on data coming out of developing countries.
Corruption, cultural aversion to sharing bad news, desire to tell the nice white guy what he wants to hear, fear of the government if you contradict their official narrative, and (as you pointed out) near-inability to distinguish cause and effect and the associated inability to gauge the effectiveness of a project.
I'd be a lot more receptive to the claim if it were phrased as "For [some period of time ending in 2025] the US prevented millions of deaths that would have happened except for the US" instead of "The Trump administration caused millions of deaths."
(I know you didn't actually put it either way.)
The first wording allows good-hearted people to nevertheless ask, "How come that fell on the US rather than on the First World generally? And how long should we have been expected to go on with that, given our mountain of debt?"
The united states is a net food exporter (or at least was, before the war in Iran and the current drought, with months drier than the dust bowl's driest months).
The 5 biggest wheat exporters (Canada, Australia, United States of America, Russia and France) provided over three-fifths (61.2%) of the overall value of international shipments for the nourishing cereal food.
The 5 biggest corn exporters (United States of America, Brazil, Argentina, Ukraine and France) generated more than four-fifths (82.3%) of overall sales for the popular grain commodity in 2025.
So, theoretically, France could do an "export in kind" the way America did (this has the effect of supporting one's agricultural sector).
Not Germany, not the UK, and... surprisingly, not Spain or Italy.
I've noticed that health advice from free AI chatbots (Claude; ChatGPT) has improved a lot in recent months, especially for anyone who looks at health advice reflectively rather than treating it like a recipe. I mean, for example, that free AI has become clearer and more proactive about describing the challenges of translating general concepts/empirical findings into individual-specific advice.
So, I'm curious what objections anyone might have about widespread reliance on free AI for health advice, assuming the user knows that AI is fallible, sometime dangerously, and that AI might not perform well as your doctor. (After all your doctor is fallible, perhaps dangerously at times, and other doctors may be more competent.)
I recall in July 2025, a patient who had clearly been talking with ChatGPT brought in a list of questions that were surface-reasonable but not-quite-the-right-questions, and they took a lot of time to answer but offered little value. It felt frustrating and adversarial.
In June 2026, several patients were asking rather intelligent and appropriate questions, sometimes catching things we hadn’t considered (mostly, asking how other aspects of their life or other medical illnesses they hadn’t mentioned to us might interact with their treatment). Or we would offer treatment recommendations that were more easily accepted than in the past, likely because AI had already given them the heads up. It was refreshing— I felt like I was really partnering with these patients in their care. (This is just anecdotal; I’m sure lots of patients continue to use AI poorly but I hope it will continue to improve)
That's good to hear. I find it especially encouraging that the free ChatGPT and Claude bots have improved at spontaneously adding useful information beyond the scope of the original question.
I've done a very hackish job of quantifying how large the time gap is between (carefully scaffolded AI programs) -> (frontier models without much support) -> (open source models) -> (small / cheap models) -> (things that you can run on a high end GPU infrastructure) -> (things that run on standard office desktops).
I get t -> t + 9 months -> t + [13-17 months] -> t + [15-18 months] -> t + [23-37 months] -> t + [35 - 49 months]
I would like to create an overview of certain qualifications and qualification steps/parts/however you call it. As each qualification builds on more basic qualifications, I would like to present them in a tech tree as in civ2 or https://www.historicaltechtree.com/. It would be great if the result would be something like a tech tree along with encyclopedic entries. It does not have to be an interactive tech tree, a printable one would be fine.
Does anyone here have a good suggestion for software, a website, or something else to implement this in an elegant way?
I’m noticing my surprise that there hasn’t been a big budget epic about the revolution (the patriot doesnt count). I’d love to see some self insert character Forrest Gump their way through the major events like in Liberty Kids
Glory is a good one. If you want American culture you can watch the first blockbuster Jaws. Very summer film and it takes place over the Fourth of July.
It touches on Vance’s psychological issues and the degree to which he actually believes in God, his ignorance of the Biblical world, how he became a populist, and his odd views about “GDP.” What shines through is the utter poverty of Vance’s vision for the future. He talks about how religion is good for fertility. It’s true that, across many different cultures, the non-religious have fewer children than the religious. But what should we take away from this? Are we going to do this dumb conversation for millions of years? Are we gonna have J.D. Vances in the generation ships? If your species won’t breed unless it’s told a bunch of fairy tales about virgin births and talking snakes, the correct takeaway is not “pretend the fairy tales are true,” it’s “this is a major maladaptation that should be fixed through some kind of genetic engineering/eugenics program.”
Didn't you get enough attention on this the last time you raised it? We did have long discussion about was his conversion sincere or not.
Anyway, you seem to be using this as a way to ride your hobbyhorse again; genetically engineer away the underclass, only PMC strivers upwards and onwards forever, yadda yadda yadda:
“this is a major maladaptation that should be fixed through some kind of genetic engineering/eugenics program.”
Given how common "motivated by eugenics" and "devout christian" are in hollywood movie villains it's a miracle there hasn't been a movie about this yet.
Make the guy an old white male tech CEO on top and surely there'd be _somebody_ who would produce the movie.
There is a whole planet (Path) in one of Orson Scott Card's books whose people have been genetically engineered to be both extremely smart and also devoted to a religion that is itself bound up in some kind of extreme (presumably genetic) OCD.
Yep. I think the idea was that their engineered high intelligence would make them dangerous, so hobbling them with intense OCD (perhaps along with a religion that fed into the OCD and gave it some kind of backstory other than being crazy) would keep them from being such a threat.
We've been doing that for some little time now. It was a good solution to the "decrease corruption and nepotism in the church" problem, but raises its own issues.
Whether you like religion or not, you have to admit it has staying power. I would expect that the generation ships and later on the space marines will all have chaplains.
I do not, in fact, have to admit that. A lot of religion--more especially historical religion than modern incarnations--comes down to just trying to explain why the natural world is the way it is. And as soon as humans started making real progress in uncovering the demonstrably correct explanations, religion started losing its grip on humanity's collective psyche. The process started slow because A. religion had a lot of institutional power[1], B. progress in science was fairly slow for the first couple centuries and C. communicating with humans outside one's immediate surroundings was very difficult until recently. The more humanity collectively understands about the world around us and the easier it becomes for us to share and discuss that knowledge with each other, the less attractive mystical explanations seem to become.
[1] If what you mean by "staying power" is actually "institutional religious power that protects itself because it's institutional, not specifically because it's religious," then I apologize for the misunderstanding and tentatively agree.
The correct takeaway is that natural selection is going to prune certain ideas and perspectives. The fairy tales will not be forgotten, since they make the mind work in certain useful ways.
A good rule of thumb is that anyone who uses the "need" to increase fertility as a justification for any sort of otherwise unpalatable policy or philosophy is somebody who should never be taken seriously on any topic, ever.
Lowered fertility could be a "crisis" in the sense that it might upset socioeconomic systems that were built with the assumption of high fertility. But that's a problem with those systems, not a problem with people's personal lifestyle choices. And anyone claiming that lowered fertility is an existential threat is utterly full of shit. The human population hasn't even stopped growing yet. It would need *several* generations of sustained, *sharp* decline before drawing conclusions like that is remotely warranted.
Also, instead of forcing people to have more children, it might make more sense to figure out what prevents people who already want to have more children from having them, and try to remove those obstacles.
Things like:
* availability of cars with more than 4 seats
* how expensive is the minimum education necessary to get a decent job
* availability of part-time jobs (for parents who have small kids)
Forcing people to be parents sounds pretty bad, let's not try that.
I think there are a lot of individual policies that probably each discourage fertility by a bit (my guess is student loan debt and crazy housing prices are the big players in the US), but the fertility collapse is global, so it's not going to be something where we change a bit of the US tax code and everything goes back to normal.
I've heard this complaint before (it's global), and I get the logic, but... it's not clear to me that the global decline is due to some global factor, as opposed to a pallet of factors affecting each part of the world in different ways.
Part of what bothers me about it involves trying to imagine what that factor might be. A casual survey of my neighborhood reveals plenty of kids; couples who want to have one seem to have little trouble having one. Or two or three. The limiting factor doesn't seem to be infertility or mortality; it's mostly just willpower, due either to concerns about high prices, enjoying DINKdom or career too much, or not finding someone to marry. This is in the US; then I picture France or Kazakhstan or Malawi and think the factors there are almost certainly different. So maybe US is fraught with prices and regulatory burdens, France has more problem with marriage, Malawi has a food bottleneck, Kazakhstan faces PNAS challenges, etc.
If so, then treating local problems specifically might work after all - for that locality. Is there positive evidence of a global factor? I've never heard of one with multiple studies, but to be fair, I don't follow the field.
There is. Literally the biggest factor in the Demographic Transition (ie moving from high fertility to low fertility overall) is female education + jobs, followed by birth control (with availability and education / information mediating how effective this second factor is).
Education and jobs leading to lower fertility is relatively uncontroversial, because it happens in lockstep in every single country where it happens. The negative education–fertility correlation is one of the most reliable regularities in social science, and it happens regardless of timing against the broader secular trends. Now you do see TONS of papers trying to salvage something more favorable to broad left-oriented views on this - "but wait, if you restrict it to only the Nords with huge safety nets, throw out half your data points and squint, you can get a u-shaped curve!" But the overall trend is rock solid.
So basically, give women the choice and better options than shacking up with duds and being stay at home wives / moms, and they'll choose to shack up less and have fewer children.
Basically, the Fertility Crisis is a triumph of autonomy and the human spirit no longer being crushed, with women eagerly training for and finding more economically valued work and enjoyable lives of their own, to avoid immolating their lives on the Sisyphean drudge work that is most women's lives in societies without female education and birth control.
BUT this leads to basically all developed and rich countries slowly (or quickly) going extinct in the limits, so there's some small downside. But you know, I'm a big fan of "that which can be destroyed by the truth, should be."
We'll be able to get uterine replicators or transgenic pigs that can take away the physical and mental burden of pregnancy before it becomes a serious problem, IMO, and that should help a little (or at least enable tyrannical governments to pump out genemod kids by the gross and ensure humanity has a future on the "fertility" side of things).
> I'm a big fan of "that which can be destroyed by the truth, should be."
Not everything that destroys us is truth. For example, there is strong propaganda saying that women who stay at home with children are unhappy losers, and the only thing that makes a woman happy is having a successful career. This has an impact on many people's life decisions.
Before we had kids, my wife insisted on buying a house close to her job, so that she could return to her workplace as soon as possible. She expected to want to be there with a baby in her arms, happily typing some research on a keyboard.
But after staying at home for a while, she realized that she actually prefers that, and now she says she would prefer to stay at home, if we could afford that.
She broke up with her previous boyfriend among other reasons because he wanted to start a family, and she felt too young for that in her 30s, and didn't want to "ruin her life". Now she realized that the thing the entire culture tried to scare her away from is actually great, but it is too late to have more than the 2 children we already have (we are both 50). Perhaps in a parallel universe where she was less exposed to these memes, she would be now happier, with a larger family.
It is true that starting a family is a bet on your partner being... a good partner. Some people make an unsuccessful bet, and they become known as cautionary tales. But we refuse, as a society, to educate people on how to find a better potential partner - and how to be a better potential partner - because that would be considered judgmental, sexist, and generally uncool. So the smarter women expect to get an average outcome, which is not very enticing.
"Literally the biggest factor in the Demographic Transition (ie moving from high fertility to low fertility overall) is female education + jobs, followed by birth control (with availability and education / information mediating how effective this second factor is)."
...you're saying this is happening _everywhere_? Not just in US/ANZAC/Europe/EastAsia? Kazakhstan? Kenya? DRC? Sri Lanka? Tunisia? Iran?
On one hand, yes, different countries have different laws, so we shouldn't assume automatically that things that impact USA must impact the rest of the world. On the other hand, the world is connected, things in USA often have a big impact on other countries, or USA may simply be among the first countries impacted by global trends.
For example in Slovakia, the housing prices keep going up (so much that people are asking whether an average Slovakian is even able to buy a house in this country anymore), and your location has a big impact on the elementary school (which is free, but assigned based on where you live; I had to fake my address in order to get my children into a better school). The university education is free; the actual bottleneck here is getting into a good high school, there are about ten good high schools in a city with population of half a million.
Culturally, educated young people in Slovakia typically know more about American politics that about Slovakian, because they spend their entire days on global social networks. So they discuss Trump and wokeness, while their own country keeps falling apart. Corporations also copy American trends, so you might get a lesson on not discriminating against black people (there are maybe dozen of them living in Slovakia), but no one ever mentions the Roma who are the actual poor minority here. And of course people get ideas about life from Hollywood movies.
Feminism is similarly a global force. Most local women are unaware that they actually had a right to vote since the foundation of Czechoslovakia in 1918, so when they complain about the time before they could vote, they are literally talking about the Habsburg monarchy. Many of them take it for granted that there simply had to be a point in history when all men were allowed to vote but women were not.
Women in Slovakia get 3 years of maternal leave per child, which is nice if you keep having children in 3 years intervals. Part-time jobs for men are virtually nonexistent, which makes it difficult for women to have children and a career. At school, children typically get a lot of homework and projects, which is a burden on parents, so the reasonable options are either to limit your children to 1 or 2, or for the man to make lots of money and the woman to stay at home full-time; the latter is strongly discouraged by feminist memes, I have only seen traditional religious people do that.
The car design is probably international.
So... there are some differences, but also some similarities.
My guess would be a raft of regulations that impose costs on parents. I don't follow this closely at all, but last I checked, delivering the baby alone runs to $15K, $3K after insurance. Then it's time for formula, baby seats, daycare, etc. I get the drive of reducing risk to an innocent kid, but I can't help but be suspicious that there's reduction in there that ends up just reducing the risk to "fuck it, no kids".
Some of this isn't regulation; just social norms. Same suspicion. Cue standard rant about CPS and letting the kid play outside.
"Lowered fertility could be a "crisis" in the sense that it might upset socioeconomic systems that were built with the assumption of high fertility. But that's a problem with those systems"
You mean systems like social security and Medicare? What should be done about those?
I disagree with the idea that it's not any kind of a crisis. At the very least, I'd say:
a. Falling fertility is in the process of massively shaking up many existing high-functioning societies, and it is not so clear that they will remain high-functioning societies as this continues. I am not sure what Japan will look like in another 50 years, but it seems unlikely to look much like the current version of Japan, or to be a significant innovator in a world where there are fewer young people every generation.
b. Existing social and tax programs fail under collapsing fertility. You may think this is fine, but I am very sure that voters who are expecting those programs to continue working will disagree.
c. It seems plausible that falling fertility could put you in a kind of collapsing spiral of population, where each generation has fewer women than the previous one. If the average number of kids per woman stays the same, that just keeps dropping, and while it seems inevitable that couldn't go on forever, it might go on until your population stabilizes at some much lower number.
d. Assuming AI doesn't upend all this, modern economies and technology require some kind of base population to continue functioning. You need enough doctors, enough engineers, enough scientists, etc. in each generation to keep things running. Does South Korea with 1/2 its current population still have chip fabs? How about at 1/10 its current population?
e. Intuitively, if you think (as I do) that having and raising children is a fundamental and very rewarding part of the human experience, it seems like having more and more people miss out on that is bad. Particularly if a lot of people are being convinced not to have kids/to have fewer kids because of, say, financial constraints or economic incentives or social pressures that could be eased.
OTOH, I recognize the tendency of lots of ideologues to see any possible problem and jump in with "and that's why this justifies every crazy coercive thing I've always wanted to do." (See global warming for daily examples.) And the worked example of how ZPG and overpopulation scares justified a lot of genuinely awful programs of coercive sterilization ought to make us reluctant to propose coercive stuff to respond to this issue as well. (Though I'd like to see everyone much more reluctant to jump on "let's pass a law and make those bastards do what I want" in general....)
I think that like many other issues we face (global warming, unsustainable debt and spending), this is a slow-moving crisis. That gives us the luxury of being able to move slowly and think things through rather than trying to ram through seventeen emergency policies tomorrow or something.
We had plenty of large scale social experiments in which humans built societies without religion during the 20th century. It didn't work out very well.
Remind me which one we're dealing with, when there's an entire month that the Allegheny County Courthouse has "the gay pride colors" up, despite how offensive they are to the religious? Or is that Leftist Religion and not mandatory atheist ideology?
I went to the AI protest mentioned in last week's open thread. Here's what I observed, as somebody who is most concerned about AI in the military and AI surveillance but sympathetic to some of the other AI risks:
- The speakers were approximately a 50-50 split between extinction risk people (all of whom came across as very doomer-y) and short-term harms people, plus a couple of people who were concerned about job loss in a way that didn't seem driven by either extinction or short-term harms.
- The crowd was maybe 200 people at any one time, but some were leaving and entering so I suspect more people showed up to the protest. Based on the signs people were holding (and talking to people, although I didn't talk to that many) I'd guess the crowd leaned at least 2/3 towards being mostly concerned about future harms, although not all extinction risks.
- The demographics were mostly young-ish, pretty diverse as far as I could tell.
- There were only minor logistical hiccups (issues with the banners trying to blow away, some people struggling to talk into megaphones, running a couple minutes behind for part of it) as far as I could tell, which slightly exceeded my expectations.
- Some people's signs were anti-Mythos/Fable, others were pro-Mythos/Fable against Trump (at least one of each, maybe more?). The % of signs I agreed with is maybe a bit lower than what I've seen at other protests I've been to, which makes sense given that it's a coalition of people opposed to various types of AI uses.
- More people than I expected were anti-all AI ever in jobs.
- I'm still confused if most of the people there wanted a short pause (to establish international agreements with strong restrictions on new releases) or a long pause (until AI alignment is fully solved or whatever). A substantial fraction seemed to want no more AI ever.
- I don't think there were any counterprotestors? If the people with confusing signs (anti-LED headlights, "total clanker genocide" (for or against???), various ones with tech jargon I didn't understand) were trying to counterprotest they did an incredibly bad job at it.
- Why was the Green Party there? Even if they're good on AI they alienate everybody else enough at this point that it's a strange choice.
- The brass band was great. No idea how they got involved with this but they definitely drew attention.
- I'm not sure if I'd go again; I'm glad I went to know what they actually believe but I'm still unsure if enough of the things the people there stood for are things I actually believe in also (I'm pro-data center in some cases and support certain kinds of job automation, also not super doomer-y).
I was there too and my impressions mostly line up with yours. The anti-datacenter, anti-automation stuff isn't really my cup of tea, it smells way too much like the type of knee-jerk reactionary Luddism that's been wrong about everything for hundreds of years, but it's probably necessary to include some of it if we want to build a coalition big enough to accomplish anything.
Some of the signs about Fable and especially the stuff about LEDs really confused me. I thought "Total Clanker Genocide" was a clear call for shutting down AI as much as possible, since if they were against genociding a group they wouldn't call them by a slur.
The logistics were not great but definitely better than they could've been. The free food at the end was disappointing. There wasn't enough pizza for everyone and the only available drink I found (besides the free slushies) was water, which was great for me since water was all I wanted, but logistically/strategically it felt like a serious blunder. An army marches on its stomach, after all.
Not sure I agree about the crowd demographics. Definitely younger on average than the No Kings protests, but I wouldn't really call the crowd ethnically diverse by San Francisco standards. Whites are only about 40% of local residents but looked like the overwhelming majority of the crowd. I definitely saw asians too, but they're a third of SF so they looked proportionally underrepresented, and other ethnicities had a pretty negligible presence. Given the current political climate, being ethnically diverse is much lower priority than it was a few years ago, and given that protesting is a popular white pastime, I don't think fully proportional demographic representation is even possible. Still, I think in terms of building a coalition, more diversity would be helpful.
It did seem like most of the energy came from the more doomer x-risk crowd, which is close to where I fall, so I'll probably come back. What you're talking about sounds slightly different but overall close enough that I'd definitely encourage you to come back too.
"Coalition building" is a terrible thing to be thinking about when your whole movement is still small enough to cram onto a bus. You can think about coalition building when it's the difference between being 26% of the population and 51% of the population, but doing it from Day 1 is a way to doom your movement.
In particular, if you make a coalition with the far left, then the far left will eat you. They are more numerous, better organised, less thoughtful. They have thousands of people who sit around all week wanting an excuse to protest about something, but their agenda is not your agenda, and if you join with them then you'll rapidly find your movement has been coopted into a general far-left mishmash with your agenda forgotten or turned into just another slogan. Worse, the people you actually want to listen to you will be alienated when your slogan starts showing up beside Hamas banners or whatever.
You may be right, I could rant for hours about my issues with organized politics and the DSA in particular, but then what's the strategy? I don't really see us even getting to 26% without some form of coalition. Where do we get the numbers from? Or is the idea that we don't actually need numbers?
There's two ways to get what you want: either you persuade those in power that they *need* to listen to you (because you are very numerous) or you persuade those in power that they *ought* to listen to you (because your ideas are actually good).
You are such a long way from being in the first category that your best shot is being in the second category. Refine your message for being persuasive to smart people, not for being easily shoutable by dumb people.
>I thought "Total Clanker Genocide" was a clear call for shutting down AI as much as possible, since if they were against genociding a group they wouldn't call them by a slur.
<mildSnark>
I thought the battlestar galactica series established that the _proper_ slur for AI was "toaster" :-)
>I thought the battlestar galactica series established that the _proper_ slur for AI was "toaster" :-)
Agreed. And I thought Dune established the proper term for the anti-AI maximalist position should be be "Butlerian Jihad" and the slogan should be "Thou shalt not make a machine in the likeness of a human mind".
Many Thanks! Yup, I've heard of both those Dune fragments, definitely part of the cultural landscape. Not sure what the successionist equivalents would be ( and would they be in English, Python, or neuralese?)...
I'm still sticking with "Toaster" as well; it more clearly expresses the idea of machines as adversaries of humanity. And, yeah, "Butlerian Jihad" has no equal. Pity the Villeneuve movies weren't able to fit it in.
Fully automated electronic line-calling in tennis started about 10 years after the sport introduced hybrid human line-calling + player challenging you now see in baseball. Last year the sport completed the transition with only one hold out being the French Open.
Will baseball follow this path for automated balls and strike, maintain hybrid calling in perpetuity, or revert to all human-calling, perhaps with some kind of redefinition of a "strike" that makes it a fully discretionary umpire call? My money is on baseball following tennis. And I think this has lessons for when humans will submit to automated algorithmic rule-enforcement and when they will not.
Humans hate high-stakes results to depend on the quality of a human call for sharp rules. The more a game or match turns on a wrong call that we can see an electronic call should have overridden, the more we will devalue human umpiring of sharp rules, until we get rid of them entirely.
For fuzzy rules, however, like "unsportsmanlike behavior," humans are a little more forgiving of high-stakes results depending on good or bad human calls. Machine-generated overrules of fuzzy rules don't hit the same way as for sharp rules.
I think this will have predictive repercussions about where in the economy, law enforcement, bureaucracy people will fully automate first.
I know next to nothing about professional tennis' motivations, but my take on MLB is that, starting with Selig and continuing under Manfred, the organization has been hyper-focused on improving TV earnings, reducing the length of games, reducing players' salaries, normalizing gambling on baseball / sportsbook, and making games more competitive (or "exciting"). In between all that was the challenge of leaving behind the Steroid Era.
I fully get that MLB wants to make more money—who doesn't?—but the above list is how they're going about it. Automated ball-strike is just one tool among many.
MLB has had instant replay since 2008. My sense is that ABS is the result of 15+ years of frog-boiling to acclimate fans to the idea of reducing the umpires' role in calling a game.
> Humans hate high-stakes results to depend on the quality of a human call
It's true, but (unlike tennis, maybe?) a high percentage of umpire calls in baseball are actually low-stakes. A majority of calls being challenged under ABS today aren't exactly nail-biters, but the batter trying to stretch their at-bat a bit further.
You're right, though, that one or two bad calls out of a hundred can leave a bad taste in the viewer's mouth, and that's the kind of thing people talk about the next day.
My point is, I don't think MLB's decisions are made so much in recognition of what humans want, but what MLB can get fans to accept. Personally, watching a game with the ump replaced by an electronic display behind the plate announcing BALL 2 or STRIKE 1 is like watching someone else play a video game. I'm probably in the minority on this point.
Hell, I think Manfred would sign off on getting rid of the *players* and putting AI robots on the field, if he could. No union, low pay, and zero steroid drama.
99% Invisible did an episode on automated balls and strikes, and they mentioned that MLB *tried* having the umpire just be a frontend for the machine - they wore an earpiece and just repeated whatever the ABS system said. But this turned out to be unpopular because a strict definition of the "strike zone" excludes some pitches that are hittable and that an umpire would have allowed. So they switched to the challenge system so that it better captures the fuzzy area that umpires and players actually think of as the strike zone.
They should just redefine the strike zone to be “hittable pitch,” so it is a fuzzy rule rather than a sharp one. I guess false negatives would be the reason to have ABS.
I only have a vague understanding of baseball, but "hittable" seems like a much broader standard, right? Like, I *could* hit a ball that sails two feet above my head.
I think the problem is that the current official definition has slight but significant mismatches with the as-applied definitions. The official strike zone is a rectangle, while the unofficial one is more rounded with pitches at the top and bottom corners usually called balls while pitches on or just outside the middles of the left and right edges (especially the edge away from where the umpire sets up to watch the pitches) tend to get called strikes. There should be a way to amend the official definition to better match the unofficial one.
Yeah it’s a feature of all fuzzy rules in everywhere in bureaucracy. “Let in only the good students” is a rule that allows broad interpretations of “good.” A sharp rule like “let in only students whose GPA x SAT score is in the 90th percentile” is a sharp rule that can then be automated.
Much like baseball, the fuzzy definition in academia works fine until you suspect the ref is biased, at which point it's nice to have an objective standard you can fall back on.
Way back in Astral Codex Ten Open Comment 314, I had a lively interaction with a good bunch of folks about free will denial. I was posting under my pen name at the time. Now, this is the real me.
Last week, I had an essay published on free will denial in The Missouri Review, July 2026 issue. For those who are interested, I must indulge your forgiveness for taking the liberty of pasting in the beginning paragraphs (but it’s “open comments,” right?). Would love to discuss. It is not academic, but a literary, creative nonfiction, personal essay. Hopefully engaging and easy to read.
SLEEPWALKING TOWARDS BETHLEHEM
On What We Know but Do Not Believe
David J. Frost
The Missouri Review, Volume 49, Number 2, Summer 2026
1. Two years ago, my fiancée told me that I’d been sleepwalking and carrying on full-fledged conversations with her each night. When Kelly first relayed this, her brow twitched slightly with what I took to be suppressed irritation. These mornings, I never remembered anything about whatever it was I’d done overnight. After all, somnambulism is a classic example of automatic behavior. Yet the archetypal image of the sleepwalker—nightcap with drooping tassel, eyes closed, arms outstretched in front, wrists bent down—is wrong. Sleepwalkers engage in complex behavior: they might cook a meal, drive a car, carry on a conversation, murder a relative—all without the accompaniment of consciousness. It raises the question: if I’m not conscious, how am I seemingly able to adapt to changing circumstances? Who was talking when I was conversing? As Kelly and I sat up in bed that first morning, her pent-up prickliness made me acutely aware of the score between us. What was the regrettable con- duct I unfreely enacted each night but for which I must nevertheless accept responsibility? Evidently, I had been peeing in the closet.
2. During the first night, Kelly says, my rustling woke her up, and she asked me what I was doing. “Nothing,” I said. “Going to the bathroom.” But she could see that I was standing in front of the open closet and said, “That’s not the toilet.” “Yes, it is,” I said, sounding offended, while gravely continuing to conduct my business. We would learn later why it’s said not to wake up a sleepwalker. Although it’s like a preprogrammed routine is being run, the person’s capability to react in complex ways can remain, but often only in limited form. I could interact with Kelly, for instance, but likely only to counteract her attempts to thwart me. If you attempt to impede a sleepwalker, they might instinctively find it aversive and—one article suggested—react impulsively, yet they’d be without the benefit of consciousness to rein in their anger. That was the theory, anyway.
3. Luckily for Kelly and me, my nonconscious self was more likely to be cavalier and arrogant than angered by impediment. While follow- ing the going-back-to-bed algorithm, I responded to Kelly’s protestations, she reports, by uttering “an aloof grunt of bemused recognition.” During the night of my second episode, when she again informed me that I was standing at the closet, not the commode, I let go another self- satisfied, haughty grunt. But I apparently persevered in my task, because she could hear the unmistakable evidence—the damp applause—of my misguided micturition. A faux-fancy language affords me distance from disgusting deeds.
4. The third night Kelly screamed at me, “ARE YOU AWAKE?” “Yes,” I said. In that moment, she struggled with what to conclude about my inner mental world. For I was passing the Turing test, so to speak, yet she didn’t know if I was actually conscious. As we found out in the morning, I hadn’t been.
5. When I asked her to recount the events for me once more (in furtherance of writing this essay), I noticed that her brow involuntarily furrowed again. The biggest irritation, she said, was having to clean up the soggy mess after I’d already roboted back to bed.
6. “That sucks,” I replied dutifully. Yet, moments later, I caught my- self feeling uninvolved again, like I hadn’t done it. I knew I had done it; I took what she said for the truth. But I did not feel that belief in my body. Usually when I’ve done something wrong, my self-criticism is more despairing. The knowledge of the injured party’s disapprobation is more—I don’t know—piercing. I get a nauseating knot in my stomach. But this time, my understanding that I was responsible for Kelly’s aggravation was entirely intellectual and bloodless.
7. Because I have so tenuous a connection to what I’ve done while sleep- walking, I find it easy to be glad of my inopportune nocturnal emissions. If I had not sleep-peed in our closet, then I would not possess such a good example of an action for which I take responsibility but which I did not do of my own free will. It’s been a fascinating experience. It’s like when, at the dentist, I cannot keep my tongue from pressing against the hygienist’s elongated mirror.
8. Above all, sleepwalking gave me a concrete path into wondering what it’d be like as an experience to deny free will, as a small number of pub- lic intellectuals have recently done. It was weird enough to think that I had carried on a conversation without free will one, two, or three times. What would it be like to think every action you undertake lacks free will? Not just outbursts but even conscious decisions made after careful deliberation. What would it be like to believe that each thought in one’s stream of consciousness was an effect of previous causes?
9. For Stanford biology professor Robert Sapolsky, for instance, what is the experience of denying his own free will like? Sapolsky is convinced we do not have free will, so from his perspective, he knows it. But he’s admitted he catches himself reverting to the default setting of free will belief every time he nonchalantly accepts responsibility. “Someone says, ‘You’re dressed nice today,’ and I say, ‘Thanks,’ as if I had anything to do with it,” Sapolsky joked during a guest appearance on NPR’s Radiolab.
10. When Sapolsky manages to believe what he knows—when he is actively remembering that he doesn’t have free will—how does his day-to- day self-conception change? Is denying free will bad for you? Or does it increase well-being? I’m talking about psychological effects. I became much less interested in the classic philosophical arguments for and against free will. Released from responsibility, does a free-will denier run amok, his face frozen in a gleeful rictus? Or does he lose all motivation and succumb to fatalism, thinking it doesn’t matter what he does? Would denial, as Sartre worried, serve as an excuse for bad behavior? Perhaps, as Sapolsky imagines, pervasive denial would bring about a fairer, more just, less cruel society, “a giant leap forward in our moral progress,” as he’s said. How would our thinking change if we cleansed our vocabulary of the term “free will”?
11. When Sapolsky and others deny free will, they don’t mean that everything is destined, that what will be will be, no matter what we do. It’s not fatalism necessarily. The precise logical implications of denial are, however, not a matter of what Sapolsky might want them to be; they are a matter of science and logic and will depend on arguments and experimental evidence. And, as it turns out, there is already an empirically informed debate going on.
12. My research into the psychological implications of free-will denial led me to novel places on the intellectual landscape: to a confrontation with bodies of knowledge called “informational existential threats,” to the conceptual affinity between progressive political opinion and free-will skepticism, to the idea of an Area 51 for anti-free-will research, and finally, to the existence of brain-scanning technology that can read the inner contents of your mind and know what you are going to do before you know.
13. When I sat down in my office to start my research, the first thing I noticed about online pictures of Sapolsky was the compassionate twinkle in his eyes. Framed by long, curly gray hair and a gray Garibaldi beard, Sapolsky’s eyes sparkled like quartz in the Tolstoian rocky crag of his sun-wrinkled face. A specialist in neuroendocrinology, he’s spent thirty sunny summers in Africa studying simians. (Sorry! I couldn’t control myself.) Returning to the same troop of baboons each year, he’d ob- serve their behavior, tranquilize them, and take blood samples. Sapolsky found, among many other discoveries, that baboons ranking low in the troop’s dominance hierarchy had chronically elevated levels of stress hormones (glucocorticoids) with concomitant negative effects on health, like hypertension, and on behavior, like an increase in isolated sulking and displacement aggression. Sapolsky estimates that nearly 50 percent of all violence among his troop of baboons is displacement aggression. Shit, as we know, flows downhill. “A high-ranking male loses a fight and so chases a subadult male, who promptly bites a female, who then lunges at an infant,” Sapolsky has written. The more a baboon displaces aggression after losing a fight for rank, Sapolsky says, the lower his glucocorticoid levels are when the good doctor, sans sun hat, tests his blood.
14. That explanation—a diminishment in rank causes stress hormones, which cause the loser to punch down to gain relief—is …..
If folks want the whole essay, it’s behind a paywall at the website of The Missouri Review.
But a PDF is available to free subscribers of my Substack.
I'm a hard determinist, in fact I think that is such an obvious conclusion I reached it in high school and all the philosophy I studied in the meantime didn't change anything. But I don't find this very relevant to daily life, because the epistemic limitations on any individual human or even a large information system are still so tight that the world ends up feeling not fully determined. My metaphysics and my existential world just don't need to intersect that often. It's still a load-bearing belief, it supports my entire worldview about science and reason, and if I thought the world contained genuine random-ness and uncaused effects I'd be extremely bothered. But it nevertheless doesn't play much of a role in mundane daily life or the way I view people, because both me and those people lack the information to predict the next state of the universe and have to rely instead on the sort of behavior-prediction heuristics that everyone does no matter what their own metaphysics are.
Hi Cjw, Thank you for your reply. What you said is very interesting. I can relate to a degree. I am pretty confidently Mr. Determinism. Even when people bring up quantum mechanics I have stuff to say that does not move from basic determinism. Either I say randomness does not get your free will. Or I say QM is indeterminant in precisely those places that physicists say "we don't understand." What I mean is "understanding" is synonymous with determinism because you cannot understand something if when you do P sometimes it does X and other times it does Y. Feynman will say that if anyone tells you they understand QM then they do not understand QM. So QM is a theory even if highly confirmed which we cannot understand. When people ask, What is going on that something can be indeterminately in a superposition of both X and not X? physicists have to say "It just is. About why, we can feign no hypothesis." Where they are quoting Newton. If we wanted to feign a hypothesis an example would be the hidden variable approach (I think).
Anyway, if the universe is fundamentally indeterminant, then there's a problem for free will there too. In fact free will skeptics sometimes call themselves hard incompatibilists rather than hard determinists. Because whether the universe is deterministic or indeterministic it'd bad news for free will (intuitively speaking, meaning one should define free will or be clear that free will has two definitions the LIB and the COMP).
As far as hard determinism not making a difference to real life,... I am somewhat sympathetic to line that of thinking too. To paraphrase you in my own way of thinking, metaphysically or ontologically everything is determined. Rewind the universe, press play, if you do not change any antecedent conditions then after "play" you get the same events again. BUT epistemologically we cannot do that rewinding. And yes everything being determined means its predetermined if only we were a Laplacean demon. But we are not Laplacean demons. The laws we do now are not exceptionless. And we rarely have sufficient detailed data about initial conditions to plug into laws, exceptionless or not.
However, we are getting closer and closer. There are some things we are learning which are still not like a Laplacean demon's level of detailed knowledge or exceptionless prediction.. but we know that chronically elevated levels of glucocorticoids retard brain development in kids. We know how serotonin and the other neurotransmitters work. And we are getting closer to being able to predict what someone will do based on this knowledge that we have. But it's not there yet.
However I can imagine we are getting closer. But then we are back to not real world stuff. Let me try this. Dennett says the criminal is responsible because he intended to thieve. The hard determinist says (as Sapolsky often does and Sam Harris does too in his way) BUT WHAT CAUSED THAT INTENTION? Then a progression can be imagined. If Dennett can characterize the cause of the intention in a way that suggests the criminal is responsible, then Sapolsky will uncover the always available cause of that cause, let’s call it N-1, (with apologies to the lit journal N+1). Then Dennett can either characterize that cause in a way that suggests responsibility or else he can’t and the cause will characterize itself as responsibility-undermining.
Let me add to this N-1 progression, that over time as scientific knowledge progresses it gets harder and harder for Dennett-types to cite a cause that suggests responsibility and more and more likely to make it all seem unfree.
So while we don't know now, we are knowing more and more and what we know I think can be used to predict. I should give an example but I've gone on long enough. I will work on saying all this more clearly by having a concrete case.
RE: the moral responsibility for crime. Whether or not the criminal was "free" to choose otherwise is irrelevant for the logic of punishment. I'll stipulate that a criminal is a deterministic optimizer that has no choice but to respond self-interestedly to the world around it. Punishment is then justified on the basis that it shifts the optimizers' incentives by altering the payoff matrix of the options available to it. Calling the punishment "morally deserved" vs "game theoretically optimal" is just semantics.
Sapolsky's argument is actually self-defeating because he's arguing that deterministic brains can't be influenced by rational control, even though the fact that he's making an argument at all presumes that he's influencing the behavior of the people who hear the argument!
I like your point. I am not sure Sapolsky says what you say he says.
What I think you are talking about is what popped up in the reading I did as "deterrence". The folks I read (like Sam Harris, Gregg Caruso) tended to say that society could continue with policies of punishment aimed at deterrence, even if criminals would not be considered fundamentally morally responsible. It's a sound philosophical position but one which may not make sense in the real world. Or so it is beginning to seem to me. The phil position is no one is ultimately morally responsible (but they may be proximately morally responsible) (which needs clarification I know.) And the position additionally says deterrence can still make sense (Harris says this for sure, Sapolsky maybe doesn't, I cannot recall Caruso on this) precisely because Harris's hard determinism countenances that people are still responding to incentives. Harris says for critics to say he thinks people are not responsive to incentives is like the critics who say his hard determinism is basically fatalism, which he says it's not. On good grounds I believe. Although I am not reproducing those grounds here.
I guess what I need to work on is an ability to explain the diff between ultimate and non-ultimate/proximate moral responsibility. And how it is that Harris thinks people still act "voluntarily" within his hard determinism. This would have to do with his "co-opting" of COMP's definition of free will and using it for his definition of "voluntary" which I alluded to at the end of my essay in the Missouri Review but cut for space and also because it was too hard to explain to non philosophers. I mean I tried and I could feel I was losing my audience. Most people won't dig through or wade through writing like this here in ACT. I have been trying to write--not here but in published essays--for the general reader.
I keep saying "Harris" instead of "Sapolsky" because Sapolsky did not really engage with the philosophical literature. Or not like a philosopher would. Which makes sense as he is not trained that way. Harris is better suited to my quest to find someone saying something non-ambiguous on these points. But even Harris is lacking in detail for all the stuff on free will he's produced.
I will say this. Usually I see people say that free will deniers-- be they Sapolsky or Harris -- have a self defeating position or argument because why are they bothering to write a polemic and convince us of their view if we are not free to choose to believe it or not believe it. And to this I say--and Sam Harris has also said--that is not how believing works. (You have not made this mistake, to be clear.) Believing is not a matter of choosing the proposition and now you believe it. As Nagel said "All belief should be involuntary." My undergrads balk at that but I think Nagel is right. Reasons compel the mind. Evidence compels the mind. The "should" in there is important too of course. Lots of people do seem to "choose" their beliefs but that is not a practice that is likely to arrive at true beliefs.
So that is the nature of philosophical discussion or really maybe any attempt at communication. You accuse Sapolsky (or let me say hard determinism like Sam Harris') of being self-defeating because HD seems to hold that people are not reasons responsive. And another critic says it's self defeating but for the opposite reason. This other critic says HD is self defeating because it conflicts with his stated purpose of persuasion which is kinda what you said but how persuasion works this other critic has wrong. You have it right it seems to me at least relative to this other critic. My issue with your position is I do not think Harris nor Sapolsky nor hard determinists in general are committed to the idea that people are not reasons responsive, or not driven by incentives. Merely because people lack libertarian free will does not mean their behavior cannot be influenced by incentives.
I feel like I may have stated that without a lot of argument in its favor. But that is not entirely worthless. ha!
Harris has parried the fatalism claim and I think the same moves of his there could be used here, to show people are still responsive to incentives.
Also bear in mind - and I don’t think I’ve seen this argument before - that if you’re a hard determinist then the people doing the punishing are also unable to do anything else. They’re just responding to incentives and bear no more moral responsibility for e.g. capital punishment than the murderer they’re executing does. Whether you call it self-reactive deterministic social dynamics or justified moral opprobrium is immaterial. The behavior is isomorphic and so the distinction is meaningless. This is why I find the dialogue on this little more than empty semantics (as, of course, most philosophy is).
I was a prosecuting attorney for 15 years and that never bothered me either because I didn't really care about the intentionality at a metaphysical level, only the legal one of mens rea. You put a thief in prison to incapacitate him and end the thefts, and because if you don't have a system that does that you'll get people taking matters into their own hands more sloppily. You also do it to add costs to crime and to mark a behavior as beneath a certain sort of person, so the criminal class is disincentivized and the typical pro-social normie views it as outside the scope of their options.
I grant that this sets up a situation where people who "aren't responsible" may be punished, but that particular type of responsibility isn't really needed. If I had been punishing innocent people (actually pretty rare if you look into it, certainly in this century) that sort of error would undermine confidence in the system and would harm its purpose. That's why you cannot extrapolate from the utility functions I mentioned earlier to a system where you instrumentally incarcerated people just to send good signals.
The law actually provides a sort of dividing line here, via the principles of both mens rea and actus reus. Not only do you require the necessary criminal intent, the actus reus distinguishes being involuntary and voluntary actions but does so along the lines of whether the act is the sort of thing we associate with deliberation. A reflexive jerking of the arm from a neurological disorder strikes a man and knocks him over the wall into a river, no actus reus. From my point of view, one might say that's not so different than a situation where one had some bizarre impulsive reaction to the look on that man's face and pushed him over, due to a nearly unfathomable chain of behavioral stimuli accumulated over the pushers' lifetime. Or even further, perhaps he takes a moment to deliberate first, perhaps he takes 10 minutes and watches and plans out knocking that man over right as he walks past the perfect spot because he looks so irresistably pushable for some reason, we're climbing the mens rea ladder into higher levels of criminal culpability. Perhaps as some say, the deliberation narrative is just a story we tell ourselves about something that already happened a few fired neurons ago. No matter to me, people seem almost universally to understand and draw a line between reflexive movements they deem "involuntary" and those they think "voluntary" and everyone seems to understand the distinction and the system works, and the sort of person who pushes a man off the bridge under what seems to us to be a "voluntary" decision-making process ends up being the type of person who needs to be incapacitated and removed from society no matter your stance on free will. So it all works out.
I really like this comment and appreciate the admirable clarity of your writing. I have not been able to be as clear here in these comments. And in my Missouri Review essay (which was clear but limited in scope) I was more or less prohibited from dealing with certain issues or ideas because my exposure to "the general public" suggests they would not sit through either my attempts to explain these ideas or because people's eyes would glaze over or whatever.
In any case I need to continue to try to write clearly, more clearly, even in internet comments. Because apparently it can be done! (Please let me know immediately, if you have any commenters here who in their response to you demonstrate that they failed to understand what you meant. :-))
That being said I take again another opportunity to think out loud.
You wrote: "Perhaps as some say, the deliberation narrative is just a story we tell ourselves about something that already happened a few fired neurons ago. No matter to me, people seem almost universally to understand and draw a line between reflexive movements they deem "involuntary" and those they think "voluntary" and everyone seems to understand the distinction and the system works, and the sort of person who pushes a man off the bridge under what seems to us to be a "voluntary" decision-making process ends up being the type of person who needs to be incapacitated and removed from society no matter your stance on free will. "
Yes. That and probably almost everything if not everything you said before gets a yes from me. In the terms of the literature I read, this is what Caruso calls quarantine and crucially he lays out his philosophy of crime as a public health issue requiring quarantine of very specific sorts ALL within the context of denying (ultimate) moral responsibility (usually b/c of notions about determinism and libertarian free will.)
So. Caruso says what you said more or less. He calls it quarantine and he says it will be justified in some cases even when there is no ultimate moral responsibility.
I don't know if staccato definitions makes for clarity but...
Denial of libertarian (LIB) free will implies no ultimate moral responsibility, but it does not imply no proximate responsibility. Proximate MR is compatible with denial of LIB, compatible with the truth of determinism.
LIB free will says you could have done otherwise than you did where the otherwise is in this absolute sense of... if we rewind the universe and let it go a LIB free person could choose soup this time when they had salad before rewinding... and, crucially, they could choose soup without changing any antecedent conditions, so, while still wanting salad for example. That is a very rarified notion of "otherwise" which always chapped Dennett's hide or rubbed him the wrong way. The more common uses of otherwise countenances that we will be changing some antecedent conditions before we press play. But the rarified use of "otherwise" is appropriate here because we are defining the absolutist philosophical position. Maybe no one will end up holding that position. But we need a name for it. And turns out lots of people will aver to having such an ability perhaps because they think the human mind is magic or their thoughts around here are not cogent or they think something logically impossible but apparently not psychologically impossible.
What is the nature of proximate responsibility? I am working on that. Whatever it is, there is an isomorphism. Deny LIB means no ultimate MR. Deny COMP means no proximate MR. But now I have to get in my car so I will have to stop typing. Hmm. I'll just add that this business is relevant to Sam Harris's sometimes misunderstood claim that he wishes to "co-opt" the COMP definition of FW that his friend Dan Dennett came up with, and use it not as a definition of the word "free will" but instead as the definition of the word and related cognates for "voluntary action" or "action for which we will be held proximately morally responsible."
Harris thinks the phrase "free will" (FW) means to most people LIB. Be that as it may, it means to some people at least the LIB notion. Harris sees Dennett therefore as offering a new definition of FW. Harris basically says although he does not use these terms, "No let's let FW continue to. mean LIB, let's just get more and more people to realize we do not have LIB FW. Let's do this conceptual revision like happened for the concept "witch" and unlike what happened for "whale." See the end of my Missouri Review essay (available for free even though paywalled in some places) for how that goes. Whale was defined as a fish at one time. Witches have the same definition they did in the 1600s, magical satanic women or whatever. It's just now we know there are no real entities that meet that definition.
So you don’t believe in random numbers, as opposed to pseudo-random numbers? There seem to be a lot of stochastic phenomena from which we can derive various useful probability distributions.
I might agree, depending on what you mean by "hard determinism". Let's say you're watching a particle, and say to yourself, "if it decays in the next T seconds, then I will have a ham sandwich for lunch, otherwise a burger". Your choice of lunch meat that day will be determined by a physical event (or lack thereof), but that event is essentially random, and thus your decision is as well.
I don't believe that is random either, the Copenhagen interpretation and similar nondeterministic interpretations of quantum mechanics fail the basic rules of logic and therefore cannot be science either. You're only unable to determine it due to an epistemic limitation. (Of course for some events the epistemic limitation may be insurmountable, or require nearly every single piece of data in the universe to be known and processed such that the prediction-engine is functionally equivalent to the actual universe.) If it decays in the next T seconds, it was always going to decay in the next T seconds and I was always going to have decided to tie my lunch choice to that outcome and always going to eat a ham sandwich.
Thankfully for me, my choice of what measurements to take concerning quantum level events was also determined.
> the Copenhagen interpretation and similar nondeterministic interpretations of quantum mechanics fail the basic rules of logic and therefore cannot be science either.
I am no quantum physicist so you might be right; however, your reasoning is IMO flawed. Logic isn't some divinely revealed immutable rule; it's just a model we've built to explain the world. It's a pretty good model, but like all models it's incomplete. The notions of "cause and effect" are no qualitatively different from the notions of "up and down": useful in most cases, incorrect in some others.
I think where I'd disagree is that scientific exploration depends on them. If you run an experiment the result of which is that you conclude the rules of logic and causality and noncontradiction don't operate, you've undermined the premise by which you conducted the investigation.
If you don't think cause and effect work you can't really do any science.
I won't pretend to understand the math of quantum physics, I had the typical 2 semesters of normal calc-based physics that only barely touches on it at all and couldn't remember a shred, but I do remember the various alternatives to the Copenhagen interpretation. One of them is called superdeterminism, and is often rejected because it fully denies free will and supposes that e.g. Bob measuring the spin of a particle instead of the energy level at this point in time is fully determined. But since I already believed all of those things that were described as the hurdles to accepting it, I had no problem supposing that was the most likely explanation.
> If you run an experiment the result of which is that you conclude the rules of logic and causality and noncontradiction don't operate, you've undermined the premise by which you conducted the investigation.
But IMO this has already happened, and more than once. Radioactive decay, virtual particles, arguably black holes (and the primordial singularity) -- all those things can be modeled quite well in ways that violate the classical rules of "cause and effect". I understand that you disagree, but the very fact that you and I are *able* to disagree on this undermines your view on logic as being immutable.
May I ask, "And? What follows from that?" I mean, I think I understand the claim that logic is XYZ. But can you expand on what you think that implies for the larger discussion here? Like is it a general comment of a nature like "who's to say? we'll never be able to know" etc Or did you offer it with respect to a more specific point? If my question does not make sense feel free to disregard. Thanks
Cool. I enjoyed these essays by Gerard 't Hooft https://arxiv.org/pdf/quant-ph/0701097 and https://arxiv.org/pdf/1709.02874 in which if I understood... one thing he says is that some interpretations of quantum indeterminacy are based on an erroneous tacit assumption of libertarian free will. Like the business about how consciousness affects the way a quantum event happens (or that it happens or not) assumed LIB and that is where we get the idea that it cannot be known. He says if we could know what the human was going to do then we could know how the universe would react. Obviously it is more complicated than that. And I know I do not totally understand. But it was super interesting. Both articles.
You wrote: "... the Copenhagen interpretation and similar nondeterministic interpretations of quantum mechanics fail the basic rules of logic and therefore cannot be science either. " I agree with that. Some philosopher think we don't have LIB, others think LIB is a paradoxical or nonsensical idea. The causa sui is LIB (speaking roughly and loosely now you nitpickers out there) and the causa sui is as Nietzsche said "a perverse kind of logical rape" and "yet mankind has got its pride all entangled up with the causa sui" (paraphrasing).
Feynman said, I believe, that if someone says they understand QM then they do not understand QM. Because QM is not understandable. Where understandable means make sense etc. I take this to mean that those moments in which QM says something is non-deterministic it is not really understandable. Why might that be? It seems obvious to me but maybe I need to work on laying out why.... If some sequence of events is non deterministic and you try to rely on it in an explanation, you will not feel as if the explanandum has been explained by the explanans precisely because you need things to remain fixed so to speak so that they can play their function in explanation. If you say, "XYZ happened because P happened" are you not saying in some sense that every time P happens XYZ happens? (within a certain context of causal talk--whether INUS or whatever)...
In a probabilistic context (not QM probabilistic but just social science level statistical probabilistic) if you explain something by appeal to a generalization, that generalization has to have certain features or it won't help explain.
What you "say" to yourself has a causal chain behind it, don't forget. So the attitude you decide to take to this random event generator takes us back into deterministic causal chain, if you follow me. Aside from that I think I follow you that the decay is random. And if we could insert somehow a random event into the chain of events leading to my action then my action would be random. I make that point in my Missouri Review essay. So we are (or might be) in agreement at least that far. Thanks for commenting within my thread.
Hello Peter Defeel, Thank you so much for reading and replying. That business about progressives and free will was one of the parts in the essay that I added in more to start a conversation than to take a side per se. As far as I was asserting anything...I would clarify this way: An earlier draft of the essay had a whole section where I explored personality types and political personality types (ie what score on the OCEAN big five personality test correlates with conservativism and what with left liberalism). I also did some research on what empirical data there might be as to a correlation between personality and free will credence or skepticism. I never wrote up a summary or my take on the empirical work because its methods suggested to me that its results were inconclusive and would stymie discussion more than help it. Be that as it may. I could be wrong there.
But I did write up a section where I laid out the affinity between progressive political opinion and modes of thinking as compared to free will denial. Did the essay mention obesity? Or did I cut that out ? Did I leave in the Ronald Reagan quote? No, I recall now I had to cut that out. But Reagan said something to the effect of “Every time a crime is committed we need to stop saying society has failed and instead say the criminal has failed and is to blame.” And then he went on about personal responsibility, which USED to be a conservative’s favorite word or one of many. I think that it is a pretty standard leftist view that when a crime is committed IT IS THE FAULT of society and not the individual. Thus, the affinity with free will skepticism. Provided we take “structural and systemic” causes to be deterministic or at least playing the deterministic role in the explanation relieving the individual of responsibility. Indeed the left would say that the factors of criminogenesis are trauma and abuse, unequal opportunities, untreated mental health conditions, toxic environment, etc. These are the factors that philosophy professor Gregg Caruso lays out, conspicuous in its absence is anything that would suggest the criminal is responsible. I think that looks left liberal/progressive. And Caruso is a hard determinist. I was prepared however for left liberals to read that stuff and reject it. It was a sort of trap I was laying. Because I had personal experience with a left liberal woman …. Well, this woman knew Sapolsky. She was a grad student in psychology and she said “oh Sapolsky is so nice, you are going to love interviewing him.” But then I told her about my passage in my essay where I talked about Sapolsky (and in the published version I add in Eagleman)—about how crime like theft might be caused not by penury (taking what’s not your because you lack what you want or need) but would instead be caused by the stuff about how stress-caused glucocorticoids cause changes to the developing brain such that kids stressed out by poverty have measurable deficits in impulse control… Well, to that she said, “That’s a pretty convenient explanation for a white male: blame the dark skinned person’s brain rather than structural racism. And it sounds too deterministic.”
I replied that I don’t know what a convenient explanation is—either some reference to a cause explains or it doesn’t. Whether it is convenient does not touch whether it explains or doesn’t. And as far as deterministic. All causes are deterministic. All explanations are deterministic, otherwise they don’t explain. (But that is a point I feel myself short of being able to explain let alone defend right now). Upon reflection, her response seems to me like a very typical response and one that I would say counts more as a symptom of something than as real thinking. After all, Sapolksy was saying not that it was the poor person’s brain full stop, but that SOCIETAL or structural inequalities were to blame for the differences in the poor person’s blame.
Just what was triggering her progressive intuitions I can only speculate. But that won’t stop me. It could be a knee jerk reaction to socio-economic explanations which lack any reference to race. There has been a debate on the left in which those two kinds of things are pitted against each other. And upon reflection I should be taken aback. She had just finished saying how nice Sapolsky was and then she felt no compunction against attributing his scientific theory to his race and gender with the implication that it was therefore devoid of merit. So that was my thinking. Put it out there and see if anyone in the public sphere wanted to agree or disagree that free will skepticism has a greater affinity with the left or the right politically speaking. As for what you said specifically… You said to paraphrase in my own words: the right should like free will denial because they think criminals are born not made. Hmm. I don’t know if they think criminals are “born” as much as they want to say that criminals need to be held accountable because they are responsible, they had a chance to choose otherwise. So rather than “born” they would say that a criminal has made himself into a criminal with his choices. And those choices are choices he is accountable for — all of which assumes free will.
that's not true empirically , since physics hasn't concluded that the universe is deterministic
I also isn't true semantically -- you can describe smoking as a cause of cancer while.knowing that it doesn't invariably lead to cancer.
>I replied that I don’t know what a convenient explanation is—either some reference to a cause explains or it doesn’t
In the world I live in, there are partial explanations, competing explanations, correlations masquerading as causation , etc.
>I replied that I don’t know what a convenient explanation is
One that's chosen , not forced on you by its undeniable correctness.
>might be caused by penury ... but would instead be caused
Might ..or might not. You are dealing with rival explanations neither of which is not I known for sure to be.true or false. Under those circumstances, a choice to back one or the other could be motivated.
>All causes are deterministic was something I said in conversation to someone who tried to say that my or Sapolsky's explanation of poverty-caused-stress-causing-glucocorticoids-which-stunts-brain-development was too deterministic
But it isn't true, if you are using the term "determinist" strictly, and doesn't answer her concern. If you had said, that it's only probabilistic, that would have answered her concern and said something true. The claim that everything is too extreme, it means social psychology wouldn't work at all, so what is needed is a happy medium ...influence without strict determinism...is something you could have said
>I suppose I mean an explanation has to be "rigid" or whatever she was worried about or it does not explain. Or maybe I meant just the opposite.
Don't you know which?
>" I suppose I mean an explanation has to be "rigid" or whatever she was worried about or it does not explain.
It has to be something. It's not clear that the something has to deterministic. In psychology and sociology , nothing is 100% certain, and researchers only look for some level of correlation.
Psychologists look for p values which are continuous variables , not a binary "it either explains or it doesn't". Even the conceptual aspect of explanation isnt a binary.
. If you stick to the high bar , that only determinism qualifies as explanation , you have to bite the bullet that psychology and sociology don't contain any explanations. Or you could adopt the low bar, that explanations only probabilistically constrain expectations. Then you can maintain that sociology and psychology do contain explanations , but not deterministic causes.
>Behind these less than perfect generalizations there are real deterministic events.
How do you know? Physics is my subject. Physics does not assert determinism. The question of determinism versus indeterminism is open.
It's true that macro level causation has to bubble up from micro level causation ...that getting cancer from smoking is based on millions of molecular events. But that's reductionism , not determinism. The molecular events don't have to be determined.
.
Some anti-free-willers acknowledge this by arguing the incompatibility of free will with both determinism and indeterminism .
>But back to free will. Sapolsky's explanation is not the kind of explanation that makes he and I doubt libertarian free will
So why does he make them? He wrote a long book full of many hungry judge type examples, where you have all these statistical correlations, none of which is individually an example of strict determinism. Presumably he thinks he can make his case by summing a lot of imperfect examples....since that's what he does. And a lot of his readers agree with him.
(Or he's using "determinism" loosely. But that's a bad idea, because loose determinism is compatible with loose free will).
>When I said "I don't know what a convenient explanation is" and you reply with a definition, what are you getting at
I do know what a convenient explanation is. You have the necessary ingredients to figure it out yourself, as well.
If you accept that explanations aren't indubitable , or certain,then it's possible for biases to affect a choice of explanation. So a convenient explanation is a not-necessarily-true explanation that' you prefer to others because it's in line with your biases.
I don't know or care whether that's what is going on the individual case about glucocorticoids.
But you can't get off the hook by saying that some psychological explanation is 100% certain, with no room for personal bias to creep in, because none is.
The thing is that your insistence on determinism is causing problems. It's forcing you to conclude there are no explanations in psychology. It's preventing you from understanding what a convenient explanation is.
> Both of those will have lots of exceptions but those exceptions should not really be considered counter examples to the claim
They should not be considered to as rendering the claim entirely false, and they should not be ignored or swept under the carpet.
Fortunately , there is a happy compromise called probabilistic reasoning. Under probabilistic reasoning, the successful examples raise the probability, and the counterexamples lower it. There's even a formula for it.
The whole Enlightenment project - from Spinoza on - was inspired by fresh Cartesian rationalism and revived atomic materialism. The big point is that Nature laws rule, not some kings. I recommend this thick book:
Re "11. When Sapolsky and others deny free will, they don’t mean that everything is destined, that what will be will be, no matter what we do": My understanding of Sapolsky's position is that he denies libertarian free will, which is what almost everyone means by the term "free will". His hard determinist position largely means that he believes that "what will be will be" and that "what we do" is also part of "what will be".
You seem to be using a compatibilist notion of free will that means something like "conscious reasoning". I think the real problem is that a lot of people who subscribe to a materialist worldview seem to be uncomfortable with the idea that libertarian free will does not fit well with that worldview, and so they take something that is not free will and use the term "free will" to refer to it so that they can pretend that they have free will.
Instead of using the phrase "free will" to refer to conscious reasoning, use the phrase "conscious reasoning". When you're sleepwalking, you might not have conscious reasoning; when you're awake, you might have conscious reasoning; and in neither case, at least from a hard determinist perspective, do you have free will.
AI's summary: Libertarian free will is the philosophical belief that human choices are not predetermined by prior causes or physical laws (determinism).
In my essay I said, I believe, that LIB free will is the same definition that Hard Determinists share... an action is unfree if it is caused, which by reverse engineering means a free action must be uncaused. And that is what the "causa sui" is. It is what LIB erroneously says causes my free actions. It is a thing capable of causing (like causing my arm to move) but is itself not caused, because if it were caused it would be unfree. So, uh, LIB free will says you are unfree if caused and can be free if uncaused, whereas Compatibilism comes in and says everything is caused that doesn't make everything unfree; to be free an action need not be uncaused; to be free an action needs to be caused in the right way.
> Hard Determinists share... an action is unfree if it is caused
...if it is causally *determined*, which is being caused in a maximal sense.
The word "causation" has multiple meanings. For instance, there are differences between sufficient and necessary causation; and between probabilistic and deterministic causation. Causal determinism is itself often confused with similar concepts such as fatalism and predictability. Causation has a complex relationship with explanation. (In addition to all that, "causality" has a specific meaning in physics, relating to the lack of faster-than-light signalling)
In brief: an effect *cannot* happen without a necessary cause, but *might not* not happen with a merely necessary cause. The instability i a radioactive nucleus is necessary to cause it's decay...stable nucleus dunt decay ...but not sufficient to cause it's decay at a particular.
A *sufficient* cause means that it's effect *must* happen: yet another sufficient cause could have brought about the effect... if the cause was sufficient but unnecessary. A fatal shooting or a fatal stabbing are sufficient to cause death, but neither is necessary
A *probaliistic* cause makes it's effect more likely, but not inevitable ... as in "smoking causes cancer". A probabilistic cause is an influence -- a mere influence. The existence of influences doesn't imply determinism. Some things that are described as "determinism" , such as "genetic determinism" are merely probslistic as well.
Only sufficient causes equate to determinism: probalistic and necessary causes allow "leeway" or "elbow room". An action can therefore be caused in several senses that are compatible with libertarian free will.
>which by reverse engineering means a free action must be uncaused
Undetermined. Which allows it to be caused in various senses looser than determinism. In particular ,it could still be *influenced*. Arguments from influence , such as the hungry judge argument , don't refute LFW. When you get the quantfiers right.
>It is a thing capable of causing (like causing my arm to move) but is itself not caused, because if it were caused it would be unfree
Lack of complete determinism isn't a thing. Any not-fully-determined event is not-fully-caused , yet not a persistent entity, nor a cause of itself. There are agent causal and event causal versions of LFW. You are only addressing agent causation. Since indeterministic events are scientifically respectable nowadays is, event causation can be used to spin up a naturalistic theory of LFW, without souls, downwards causation , or causa sui.
That's the best damn summary of causation and determinism I've ever read. I'm bookmarking this one! Thank you!
My problem with free-will vs no-free-will arguments is that there's no way to empirically validate either position. Neither hypothesis is falsifiable outside a framework of logical argument. We might as well be arguing about the existence/non-existence of a creator being.
If you get the same result every time for a given experiment, that's evidence of determinism. It's not verified, but it's verifiable.
Sufficiently detailed models of naturalistic libertarian free will can be tested, because they predict brain mechanisms and processes you can test for. It's not verief, but it's verifiable.
Here is something I just wrote to a friend via email who asked about the definition of free will:
There are two definitions in the debate. The so called Libertarian definition is the one that makes the problem seem like a problem because that definition makes FW incompatible with determinism or with how the universe works if we think accurately about how the world works. (If quantum indeterminacy is the truth of how the universe works that introduces randomness not freedom and so is also incompatible with an intuitive sense of free will.)
If the LIB definition is what gets the problem going, the so-called compatibilist definition is offered by philosophers trying to solve the problem.
What does each hold? LIB says that to be free an action must be uncaused, because any action that is caused is unfree. People can be shown to believe we actually have this ability via philosophers questioning them like Socrates and putting situations to them and asking them if the person has free will in that case. For instance, assume you chose soup today for lunch. Imagine we rewind the universe and press play…without changing any antecedent conditions could you choose otherwise than soup? If the person says yes they are using the LIB definition. The LIB definition of FW is “an ability to have done otherwise then we actually did do.” And without changing any antecedent conditions means, everything is the same. Determinism holds that if everything is the same after rewinding and we don’t change any conditions, then you will have soup again. But the LIB person says we could choose salad this replay even without changing anything ie even without changing things such that you want salad. LIB says you could choose salad even while continuing to want soup, or even while all the causes that are germane would normally cause you to choose soup. That shows that LIB FW means an ability to cause things but this cause is itself uncaused. That is the causa Sui from my essay. It is an self caused or uncaused causer,
Okay so that’s LIB. Whether it makes sense or not, it is what lots of people aver when pressed.
LIB, again says, to be free an action must be uncaused. OOMP says, no every action is caused, that does not make every action unfree. To be free an action need not be uncaused, it needs to be caused in the right way or by the right kind of cause. So if an action is caused by a pistol to my head that’d be unfree. But if the action is caused by a desire of mine that I identify with and do not lament, then that’d be a free action, according to this COMP kind of definition. COMP-ists believe in determinism, and they also deny LIB is true. So they say you could not have done otherwise than you did (not in the ultimate sense) but what matters is not that you could have done otherwise but what was the source of your action.
And philosophers have shown pretty well I think that all the various ways that nonphilosophers try to solve or understand the free will problem fall into this schema. Either they are offering or using the LIB definition or the COMP one.
> (If quantum indeterminacy is the truth of how the universe works that introduces randomness not freedom and so is also incompatible with an intuitive sense of free will.)
Randomness is indeterminism, indeterminism. is freedom from determinism , freedom in the libertarian sense.
The objection to randomness based LFW is.much weaker than usually assumed.
Randomness is often held up a something obviously incompatible with will. But indeterminism is not an objection to FW in itself: it needs to unpacked into a series of objections to specific features of a kind of free will "worth wanting" -- purposiveness, rationality, control and ownership. These objections can be answered individually.
(Explaining naturalist libertarian free will in terms of "randomness" is creats a communication problem, because the word has connotations of purposelessness , meaninglessness, and so on. But these are only connotations, not strict implications. "Not deteminism" doesn't imply lack of reason , purpose , or control. It doesn't have to separate your actions from your beliefs and values. Therefore,I prefer the term "indeterminism" over the term "randomness".)
So, how to explain that indeterminism does not undermine other features of a kind free will "worth wanting": there are concerns about ones decisions being rational, being connected with ones desires, and ones actions being appropriate to ones decisions.
Part of the answer is to note that mixtures of indeterminism and determinism are possible, so that libertarian free will is not just pure randomness, where any action is equally likely.
Another part is proposing a mechanism , with indeterminism occurring at different places and times, rather than being slathered evenly over neural activity. In two-stage theories, such as those of James and Doyle, the option-generating stage is relatively indeteministic, and the option-executing stage is relatively deteministic.
Another part is noting that control doesn't have to
mean predetermination -- it can also mean post-selection of gatekeeping.
Another part is that notice that an undetermined choice between things you wish to do cannot leave you doing something you do not wish to do, something unconnected to your desires and beliefs.
>And philosophers have shown pretty well I think that all the various ways that nonphilosophers try to solve or understand the free will problem fall into this schema. Either they are offering or using the LIB definition or the COMP one.
Libertarianism importantly subdivides into a demand for omnipotence, or complete freedom from influence, and a more modest kind. It also subdivides between supernatural and natural mechanisms.
> And philosophers have shown pretty well I think that all the various ways that nonphilosophers try to solve or understand the free will problem fall into this schema. Either they are offering or using the LIB definition or the COMP one.
That's a pretty grand statement. I suspect this is a false dichotomy arising from premises imposed by the logic you're using. Certainly, Mahayana philosophy doesn't see a simple dichotomy between free will and determinism. But then again, they use a paraconsistent logic with four possible states. And I suspect — suspect — that makes their interpretation more nuanced than Western either-or structures. I need to think about that some more...
I am working in--and my essay announces itself as existing in--the framework of western analytic philosophy's take on what it calls the philosophical problem of free will and determinism. Mahayana is probably wiser. But bringing the two together is not something I'm signing up for at the moment.
> "If quantum indeterminacy is the truth of how the universe works that introduces randomness not freedom and so is also incompatible with an intuitive sense of free will."
But the human brain may have selective power which quantum fluctuations fit the intention or will (as material basis for possible actions), and clutch onto the "anticipated" fluctuations and let them rule the dynamics. Would that not be a compatibilist version of free will, where you got to enjoy your "anticipatory" neural biases just as much as your organic endowments?
To hard determinists I have this question: In what (causal?) ways do physical processes "determine" or "dictate" the biology and our thinking? Say, an adaptation happens - which implies a novel physical basis for your behaviour. But given that physical basis, who can "determine" from the first principles what is the macro-scale effect? This problem (pattern recognition?) could be undecidable per Godel or Turing. Emergent complexity is organized by intricate macro-scale processes and emergent centers of control, but all the credit must go to atoms moving in the void?!
And by the way, even classical thermodynamics is not "determined" by atoms, since statistical postulates are employed as additional assumptions. Hence we have Boltzmann's Stosszahlansatz and Loschmidt's paradox. The reduction logic is not in logical deduction, but statistical approximation or asymptotics.
Hi David, Thanks for your reply. I’m not sure I followed as you will see below. But here’s my reply for what it’s worth.
“What will be will be.” Let’s call that fatalism. Or call it shmatalism. I say that to keep myself honest as it were. It shouldn’t become demerited in virtue of what we name it. So, let’s ask, what does it *hold*? It holds or means that what we will do is fixed no matter what we decide to do. Is that an implication of causal determinism? Sam Harris does not think so. But back for a moment to what fatalism holds. Some ancient rebuttal went like this. If determinism is true then should we not call the doctor, since we will get well or we won’t no matter what we do? No. You will get well if you call for the doctor and you will not get well if you don’t. So by all means you should call for the doctor. And we can make a choice (or what seems like a choice) to call the doctor or not, **even if** whether we call the doctor is already decided (but of course only already decided in the eyes of a Laplacean demon). It can seem contradictory that you should call the doctor even though whether you call the doctor is already determined. But if we introduce the difference between metaphysics/ontology and mere epistemology it might make it easier (for some) to see how it’s not contradictory. Metaphysically it is decided, but at the level of our understanding and knowledge, our epistemology, it is not known to be decided and therefore can feel like a choice. And without harping on that language of “it can still feel like a choice” let us instead put it this way: you cannot wait around to see if you will get well. You must act.
Let me move on to your next point. You said I seemed to be using a COMP notion of free will meaning something like conscious reasoning. I am not sure what you are attributing to me. Because I do not simply use one definition of FW. I’ve remarked that there are two definitions. And what free will means, which of the two definitions it is becomes an interesting question. Many philosophers used to say that the folk had the LIB definition in their heart of hearts or that when pressed with Socratic questioning the folk reveal they’ve been using “free will” according to the LIB definition.
And philosophers used to be pretty certain they could figure this out from the armchair. Then along came experimental philosophy. They took thought experiment formerly considered from the armchair and put them to the masses using the survey technique. This paradigm shift occasioned a reconsideration of the role of folk intuitions in the philosophy’s methodology of conceptual analysis.
But I should not go down that path even though what I’ve said so far about X-phi won’t make sense to someone who already didn’t know what I was talking about.
So let me return to “conscious reasoning”… Yes materialists do not think LIB makes sense. “So they take something that is not free will and use that so that they can pretend they have free will…” etc. But now you seem to be talking about compatibilist. But If the implication is that I am that compatibilist, I do not think you have any textual evidence to pin that on me.
And as far as sleepwalking, I did not to use it to illustrate lack of free will. Only that it got me to thinking about free will.
You wrote: sleepwalking or wakeful walking “in neither case, at least from a hard determinist perspective, do you have free will.” Yes for a hard determinist no matter what event you bring up it will lack free will in the LIB sense because every event has a cause and the LIB definition of FW requires that the “free” event be without a cause. (And the Hard Determinist shares the LIB definition with LIB)
> Metaphysically it is decided, but at the level of our understanding and knowledge, our epistemology, it is not known to be decided and therefore can feel like a choice. And without harping on that language of “it can still feel like a choice”
Sure, but "feels like a choice" and "is a choice" are two different things, and that distinction is the thing that I wanted to harp on. My gripe is that the term "free will" should be reserved for a thing that really "is a choice", and not one that just "feels like a choice". That gripe I have isn't with you personally, and if I misread what you posted above, and you're not actually conflating those two things under the same heading "free will", then it's not at all with you. I just enjoy complaining about it.
> And as far as sleepwalking, I did not to use it to illustrate lack of free will. Only that it got me to thinking about free will.
> But If the implication is that I am that compatibilist, I do not think you have any textual evidence to pin that on me.
Hi David, Thanks for the apology. But it's all good. I am finding this posting difficult. My writing takes thousands of hours to become clear. Like my Missouri Review essay reads much much different as compared to what I am producing here. The essay is way better. I was hoping posting here would be a chance to ... I guess I am not sure what of many things I had in mind. A chance to discuss, listen, think out loud, test out ideas. I will try to remember my purposes. And I'm just not going to be able to get a reply to each reply to my original thread. I am saying this for myself and my other readers besides you. You are all good. I appreciate you.
For what it is worth, I will say the difference between "feels like a choice" and "is a choice" is also a difference I am interested in.
> to the existence of brain-scanning technology that can read the inner contents of your mind and know what you are going to do before you know.
If you're referring to the studies were readiness potential (RP) is believed to be the brain making a decision before you're conscious of it, Aaron Schurger argues that the RP is largely an artifact of averaging many trials together, and it reflects ongoing, fluctuating, and random neural activity in motor areas. IIRC, when this background activity happens to cross a threshold, a movement is initiated. But aligning many trials on the movement makes this random buildup appear to be a smooth, purposeful signal.
The Libet stuff--Libet 1.0 (readiness potential), 2.0 and Haynes's stuff which could be Libet 3.0--is interesting. Maybe. I think my take on it (which may not be clear in the essay, even in the full essay--had to cut lots for space, before and after acceptance) is different than most people’s take and unassailable by the arguments Libet debunkers usually offer. (I’m not trying to be grand and call myself unassailable. I am just writing quickly.) One thing I think is that Libet et al is an empirical case against LIB not COMP. So the COMPs who say “Yeah my conscious thought bubbles up from my nonconscious mind and is detectable before I’m aware, that’s fine, it does not make me unfree. It’s still me . Etc.” The COMP who says that is on good footing as far as I am concerned.
The rest of my position on Libet is that the most important thing to take away (or that I take away) from the Libet and Haynes stuff is something that could have been argued a priori without all the empirical science. And if fact was argued by Nietzsche and more recently by Sam Harris.
What do I think is the most interesting thing to take away from Libet? It is that conscious thought does not appear uncaused or out of nowhere nor is conscious thought like a decision one long unbroken chain of conscious preceding thoughts back to the conscious reception of the stimuli. Instead nonconsciousness plays a huge role in supplying the content of thought to consciousness. So decisions involving conscious careful deliberation ultimately can be traced back to some thought that bubbles up from nonconsciousness—bubbles up as much as a decision bubbles up when you ask yourself “What am I hungry for today?” And the answer comes. Happy to discuss more but will post this incomplete response now since I am falling behind answering folks comments.
Oops, so to connect the final dots... The brain scans are picking up the nonconscious processes before they provide the content to consciousness. But we need not have done experiments like that. There is an argument I think A reductio maybe. Because where else is conscious thought going to get its content. My essay quotes Nietzsche or references him anyway--he did not use the phrase "bubbles up" -- he said thoughts come "unbidden". Sam Harris for his part say in meditation you learn to pay attention to the phenomenology being alone with your thoughts, you realize they come unbidden or out of nowhere. And in my essay at the very end I describe a meditation session and how sure enough when you meditate you realize how wild the mind is. Thoughts coming in out of nowhere while you are trying to focus on the breath. We are often 3 seconds or more already going down a tangent like plans for dinner having turned from the breath to the thought about dinner (which bubbled up, unbidden) WITHOUT REALIZING it. Indeed, when you realize you are not focusing on the breath, it is often ALSO the case that you can recall that you have not been doing so for a relatively long time. What does that mean? That means that when the thought about dinner bubbled up you did not notice that it bubbled up, you just went along and thought about dinner--you identified with it for lack of a better term, you adopted it, probably, as part of your consciousness's domain.
I think that is pretty interesting and I imagine that some large percentage of the population, when pressed, without first reading the above, would say that their stream of consciousness or their experience is all the domain of consciousness and its a chain unbroken of conscious mental events all day long etc. But I think we know it's not. There was a journal article that I read and did not really write up for my essay called "Do conscious thoughts cause behavior?" And we gotta be careful we approach it asking the right question and so on because their answer is a qualified "yes". But when you look closely almost every one and maybe indeed every one of their kinds of cases involved a chain of mental events from stimulus to decision and action in which nonconscious processes played a role. Do nonconscious processes cause behavior? Yes, they are implicated in every causal chain from stimulus to action. (There is some subtlety here but I've written enough again for now)
The whole question of free will always seemed to me to be an argument about semantics.
"Did the murderer choose to do it?" Well, his brain shuffled some electrons and puffed out some neurotransmitters, and he doesn't seem to meet the diagnostic criteria for us to feel sorry for him and send him to forensic mental health. He was probably also not dissociating or sleepwalking at the time.
"But did he, like, choose to do it?" Does it matter? What real impact will the answer have? If we want to choose whether to let him go, or to hate him, those questions can be answered without finding an answer to whether he had free will.
The whole way down it's just semantics.
"What about uncaused effects?" Could be baked into quantum mechanics, but does free will really hinge on some weird fact about observing subatomic particles for anyone? Is your will any freer, or less free, because an electron went left instead of right?
"Am I the one choosing?" Well, how do you define "I"? Do you think you're the brain? Then you're choosing! Do you think you're the identity or the consciousness or some other facet of intelligence? Then you're sometimes choosing! Do you think you're an immortal soul separate from physical reality? Then I dunno, I'm not a theologian, but that sounds like a problem which depends on how you define those words.
I bet actual philosophers have gone over this time and time again, but I'm confused as to why it doesn't get acknowledged more.
>The whole question of free will always seemed to me to be an argument about semantics
Its.*partly* about semantics, and so are many other arguments.
>"But did he, like, choose to do it?" Does it matter?
Yes. It matters if we punish someone for a crime committed under duress, and it matters if we punish someone who is not a mentally capable adult. Both of these relate to free will under compatibilist semantics.
>"What about uncaused effects?" Could be baked into quantum mechanics, but does free will really hinge on some weird fact about observing subatomic particles for anyone? Is your will any freer, or less free, because an electron went left instead of right?
Yes, under libertarian semantics. Libertarian free will require a that you could have done differently under the same circumstances, and that's only possible if the universe works a certain way, and the brain works a certain way. It's partly a matter of semantics -- the libertarian definition has to be the one the fuss is.about -- and partly a matter of reality.
>The whole way down it's just semantics
No. See above.
>"Am I the one choosing?" Well, how do you define "I"? Do you think you're the brain? Then you're choosing!
Even if the choice was a foregone conclusion a million years ago? There isn't a single obvious to that question, either.
I see the factual question about whether the universe is deterministic, but don't understand how that's connected to free will. It seems like, if I kept asking questions here, I would eventually arrive at a choice between "neither determinism and randomness is really free will" or "God did it"-style metaphysics about how the randomness is supernatural, and we might as well just short circuit the bit about free will and go straight to arguing the metaphysics instead.
> Even if the choice was a foregone conclusion a million years ago? There isn't a single obvious to that question, either.
Again, I think this is a semantic argument, because if we drilled down into "what does it mean for a brain to choose something?" then we'd eventually have to define it in a way that gives us an answer. Either you're asking something like whether the brain is capable of responding to stimuli with a bunch of caveats about time's arrow and what a response means (it is), or whether the brain is capable of spontaneously doing its own thing (maybe, but then it just decomposes into a different argument that's not about free will). The wrangling is either semantic or it turns into a different philosophical argument with free will just plastered awkwardly on the top.
I feel a bit lied to by all the people who suggested this was an important question with real-world implications, because it seems like a problem created by religious people to help them think about sin. I may just not be cut out for philosophy.
>I see the factual question about whether the universe is deterministic, but don't understand how that's connected to free will
Determinism would refute free will, under one definition of free will.
>It seems like, if I kept asking questions here, I would eventually arrive at a choice between "neither determinism and randomness is really free will
Not from me. I'm a naturalistic libertarian.
Naturalistic libertarianism doesn't require any kind of soul, or fundamental third alternative to determinism and indeterminism. It assumes indeterminism, and there is no need to override determinism if the universe is not fully.deterministic in the first place. Partial determinism means there are some things I cannot do, but also some options I can choose between.
According to science, the human brain/body is a complex mechanism made up of organs and tissues which are themselves made of cells which are themselves made of proteins, and so on.
Science does not tell you that you are a ghost in a deterministic machine, trapped inside it and unable to control its operation. Or that you are an immaterial soul trapped inside an indeterministic machine. Science tells you that you are, for better or worse, the machine itself.
Although I have used the term "machine", I do not intend to imply that a, machine is necessarily deterministic. It is not known whether physics is deterministic, so "you are a deterministic machine" does not follow from "you are entirely physical". The correct conclusion is "you are no more undetermined than physics allows you to be".
So the scientific question of free will becomes the question of how the machine behaves, whether it has the combination of unpredictability, self direction, self modification and so on, that might characterise free will... depending on how you define free will.
There is a whole science of self-controlling machines: cybernetics. Airplane autopilots and , more recently, self driving cars are examples. Self control, without indeterminism is not sufficient for libertarian free will, but indeterminism without self control is not either
All of those things can be ascertained by looking at a person (or an animal or a machine) from the outside. They don't require a subjective inner self... unless you define free will that way. If you define free will as dependent on a ghostly inner self, then you are not going to have a scientific model of free will.
>Again, I think this is a semantic argument, because if we drilled down into "what does it mean for a brain to choose something?" then we'd eventually have to define it in a way that gives us an answer
We can define it in two ways that give us two potential answers.
Which still doesn't make it entirely semantic.
>Either you're asking something like whether the brain is capable of responding to stimuli with a bunch of caveats about time's arrow and what a response means (it is), or whether the brain is capable of spontaneously doing its own thing
A response doesn't have to be deterministic, so that's a false dichotomy.
>but then it just decomposes into a different argument that's not about free will).
I dont see why that would be the case.
>because it seems like a problem created by religious people
> It matters if we punish someone for a crime committed under duress, and it matters if we punish someone who is not a mentally capable adult.
Yes, of course it's wrong to "punish" someone for a crime they didn't "choose" to commit.
But that doesn't mean it's wrong to put them in prison. When people have a tendency to commit crimes, we put them in prison to stop them from committing crimes.
This is prison-as-incapacitation, not prison-as-deterrent.
Let's suppose there's a person who commits lots of murders all the time; let's suppose this person does not have free will by some definition. I think this person should be in prison. It sounds like you're arguing they should go free?
>. It was weird enough to think that I had carried on a conversation without free will one, two, or three times.
You were carrying on a conversation in some unusual state of mind. It's not clear that free will was lacking, or that it's present the rest of the time, since some debt it wholesale -- and there is also semantic confusion about what the term even means.
Any chance that sleepwalking may involve some consciousness, but we would not remember consciously? After all, dreams are conscious experiences even if we hardly remember them. For comparison, Daniel Kahneman was puzzled that people would "prefer" enduring lengthier colonoscopies if only residual memories are less gross (due to better ending protocol).
Thanks for your comment. I guess there is a chance of that. I don't know the empirical science of sleepwalking. And, in any case, in my essay I only use the sleepwalking episodes as a path into thinking about free will.
But as far as the significance of sleepwalking as more than just an anecdote... sleepwalking foregrounds how a lot of behaviors we think require consciousness can be performed by nonconsciousness, to coin a term. This thing I just said of course assumes we are not conscious while sleepwalking instead of your contention that we might be conscious and not forming memories or will later forget what memories are formed.
But, if we are at an impasse there, then I would just come up with another example to demonstrate how much of the behaviors which many of us seem to think only consciousness can control, can be and actually are controlled by nonconsciousness. Timothy Wilson’s book Strangers to Ourselves comes to mind.
That is one change my essay speculates must needs come eventually: presently people give way too much to consciousness not enough to non-conscious parts of the mind. Plus I am not sure if regular people or the folk think of nonconsciousness as a bunch of little selves or if they think of them as blind instinct. It might be that soon we will have to think about nonconscious processes as little selves that can make decisions (by which I mean we won’t conceive of them as merely instinctual).
Presently people give way to non-conscious parts of the mind incomparably more than at Freud's time - whether it is in marketing, or management, or politics, or popular culture, or science. Perhaps we are neglecting consciousness actually, as public discourse became more rhetorical and emotive rather than rational. Trying to influence anyone by factual argumentation rather than Cialdini's prescriptions (or social "mechanisms") is almost a bad social taste nowadays. The biology and neurology are primed to delegate some causality to YOU (as a decision making center), is it not a shame to make wholesale excuse that we are so determined?
Heads up! It looks like the comments section on the archive page of the old SSC blog has gotten spammed by crypto bots. (Sorry if I am the thousandth person to notice and post, saw it last night and figured I would share in the next open thread.)
I made a new blog post on how many dollars it costs to take a life, the opposite of effective altruism, calculating the cost for everything from driving and smoking, to abusing amazon prime or buying drugs. Check it out if you're interested!
Not really the point of your write up, but I expect that you could sponsor some extreme BASE jumping events. As with bull riding, I don't think there is much money in BASE jumping so a few million dollars could probably encourage enough marginal behavior to cost a number of lives. Be subtle, though, so the BASE jumpers don't catch on.
I did think of paying for promoting bad health practises as another thing, but sponsoring dangerous activities is another good (or bad) idea. It's hard to calculate but if you paid for free base jumps for people, about $500 each jump. About 1 in 1000 jumps results in a death, which sounds way too high for the number of people that still do it. So only $500,000 needed to cause one death!
Nectome is seeking a Research Associate with a DVM degree to join our brain preservation research program, which investigates post-mortem cerebral perfusion and tissue preservation using both small and large animal models.
We strongly value previous research experience and hands-on work with large animals. The successful candidate will plan and execute experiments involving research animals, performing anesthesia, euthanasia, surgical preparation, surgery, vascular cannulation, intravascular perfusion, tissue harvesting, and downstream analysis.
This is a full-time position in Vancouver, WA, with compensation ranging from $100,000 to $150,000 USD per year, depending on experience. Applicants should submit a CV, a brief cover letter describing their experience, and up to three references to b@nectome.com. Review of applications will continue until the position is filled.
Can you share the reason why you ask for references up front? I'm not a fit for this position, but I've always been reluctant to supply references before an interview. This is mainly for two reasons:
1. Worst case scenario; the company has no intention of hiring the applicant based on their resume, but uses the references' contact info to headhunt or bother them to find new candidates.
2. My references' time is valuable. They're busy. They don't need to know every time I apply for a job, but in cases like this I need to notify them that their contact info has been/will be shared.
Now, 1 is less likely to happen, but if you have a valuable reference that's hard to reach...
FYI: I'm not claiming that your company is doing this, just a general pet peeve.
My latest Fog o' War post. In which I muse upon whether Putin is acting rationally within the constraints of the information being given to him, or whether he's delusional. I go back and forth on why he's making such poor decisions, and a case can be made for either hypothesis. Plus I highlight some of the other what-the-f**k-were-they-thinking-! moments of the Russian authorities seen over the past week.
Are the Russian authorities entirely irrational though ? Or are they merely betting that they have more oil/tankers/trucks/whatever than Ukraine has drones ? If you need one tanker to get through, and you have effectively infinite tankers, you might as well send 100 of them and call it job well done.
One obvious retort is, "but what about the lives lost on all those vehicles ?", but this is Russia, they don't care. Another retort might be "ok but their domestic economy is already starting to struggle for lack of oil", and here they do care at least a little. But the Russian mentality had always been to endure hardship, and I think the leadership might be betting on it kicking in these days as well. This is why they are framing the Ukraine War as a sequel to WWII: it's us against the world, we must protect the Motherland at all costs, sacrifice for a brighter future, etc.
"Russian mentality had always been to endure hardship, and I think the leadership might be betting on it kicking in these days as well."
Quoting IC: "America has patriotism. Russia has spite." (the Russian grit and endurance is legendary, and yeah, I think Russia can go on for far longer than people think.)
[I am rooting for "whatever ends this idiotic war" at this point.]
>"For instance, his claim that Zelenskyy, who is Jewish, is a Nazi."
AIUI, the post-Soviet connotation of "Nazi" is anti-Russian rather than antisemitic, so the claim doesn't sound as patently absurd in Russian as in English.
I'd say it's not just post-Soviet. Pro-Palestinian people call Israel a Nazi state, with the same connotations: if a country is controlled by an ethnonationalist ideology that aims to either forcefully assimilate, expel, exterminate or at least limit the rights of minorities "to preserve its essential character and its purpose as the homeland for its people", then it's a Nazi state. The minority in question doesn't have to be Jewish.
Seriously, this should not be controversial. We can look at an army commander (former "minor" elected official, think American Congresscritter), and see he's making neo-nazi rhetoric.
Whitewashing. Although it is true that Jews do serve in the Azov militia, that's where we get things like Neonazi synagogues (with swastikas on display), and Azov making terroristic threats against Zelensky (so if you don't want to call them Nazis, do call them terrorists, it's accurate).
The oligarch behind Azov is jewish (well, the first backer of the group the new backer is apparently sunni muslim, but you didn't see neonazi synagogues under him). But if you want to know why Putin is saying "They're Nazis!" You have to look at what he was seeing... which is some very "nazi-like" people who Do Not Like Russia, and Want The Russian-speaking Ukrainians to continue being bombed, even if Zelensky doesn't want to.
But then we encounter situations of cognitive dissonance when one era's Nazis become another era's oppressed. For instance, we have the example of the Grand Mufti of Jerusalem, Haj Amin al-Hussein, who was appointed by the British to preside over the Arab population of Palestine (many of whom had emigrated from Jordan and Syria in the earlier part of the century). In the 1930s, he incited pogroms against the (small) native Jewish population, and the more numerous Jewish settlers. Despite the complaints of Jewish leaders, the British turned a blind eye to his atrocities — although they finally got fed up with him and issued an arrest warrant against him not long before the outbreak of the war. He fled to Germany, and publicly aligned his Arab nationalist cause with Nazi Germany. He met with Hitler in 1941 and Hitler made him an honorary Aryan and confirmed that the Arab goals of eliminating Jews from Palestine were part of the larger Nazi effort against the Jews. Al-Hussein spent the war broadcasting Nazi propaganda in Arabic, and after the war, he was interned. But with the assistance of anti-Jewish elements in French and British security orgs, he was able to flee to Cairo, where he formed a government in exile. His internal rivals eventually sidelined him from leadership, and they went on to found the PLO.
So, we had German Nazis, Arab Nazis, and now we have (reputed) Ukrainian Jewish Nazis. And of course we also have Russian Nazis, but they seldom get mentioned in Western media...
Yes, Palestinians and their supporters call Israel a Nazi state, because of the ethnonationalism and the violence. This has *nothing whatsoever to do* with the bit where Russians call e.g. Ukrainians Nazis. There are two completely different definitions of the word "Nazi" at work.
Well, really three. There's the Russian version, which means "Anyone who fights a war against Russia", the American Leftist version, which means "Anyone to the right of Joe Biden that I want you to hate", and the version everyone else uses, which is "Murderous ethnonationalists like those guys who used to run Germany; exact boundaries TBD", And there may be value in trying to pin down the borders of that third version, figure out whether contemporary Israel fits the definition. But any attempt to develop a Unified Theory of Nazis that tries to tie any two of these definitions into a cohesive whole is doomed to failure.
Might as well try to find a single definition of "Spam" that includes both canned meat and mass unsolicited email.
Given that this is a cultural meaning in Both Ukraine and Russia, yeah. Hitler is seen more favorably in the Ukraine, because he fought against Russia. (the genocide? not so much). Hitler is only the antichrist in Western Civ (India for sure loves him, and that's a very different culture).
Putin is a chronic vacillator, he is incredibly biased towards not doing anything. Like Merkel. That's how Russia has squandered most of its temporary advantages over Ukraine, allowing the EU to turn Ukraine into a credible threat.
He has driven the country past multiple off-ramps, because a ceasefire would've required doing something completely different, and quickly. And he hates quick decisive actions.
All states have Safe Haven laws, under which overwhelmed parents (usually mothers) can drop off infants at locations such as hospitals, relinquishing all parental rights, no questions asked. In most cases the laws apply to infants up to one month of age. When Nebraska enacted its law in 2008 the legislators forgot to add an age limit, and within a few months parents dropped off 35 children ranging in age from 10 to 17.
From what I could find out, several of the older children were from the same large family, the mother died and the father was too overwhelmed to raise them. Other relatives soon took the children.
World #1: Bad parent decides to abandon his kids to the streets/kick them out of the house. Eventually maybe the abandoned kids get picked up by the cops and put into the foster system. (Or maybe some pimp adds them to his stable, or they sleep on the streets, or find someone to take them in informally.)
World #2: Bad parent can't or won't bother keeping his kids fed, clothed, in school, etc. Eventually maybe CPS gets called and eventually after a year or so of bureaucratic and legal procedure, the kids get taken away and put into the foster system.
World #3: Bad parent takes his kids to the fire station and says "I can't/won't care for them anymore." They get put into the foster system.
It seems to me that in almost all cases, World #3 is better than World #2 or World #1.
The potential downsides of this are:
a. Maybe some parent does this in a fit of anger or crazy and later regrets it.
b. Maybe some parent doesn't want to support his (sick, disabled, troublesome, annoying) kid and fobs him off on the state.
c. Maybe the foster system is such a meatgrinder that even the kind of parents who want to abandon their kids are much better, and maybe this program somehow encourages many of those better-than-nothing parents to abandon their kids to the state.
d. Lots of people skipping birth control (and for some reason not having much problem with birth) because they know they can always give the kid up to the state. Including parents who continue to drink, ingest substances, etc.
>d. Lots of people skipping birth control (and for some reason not having much problem with birth) because they know they can always give the kid up to the state.
Is more children being born a bad thing? Or are you concerned this will make fertility more dysgenic?
>Maybe some parent doesn't want to support his (sick, disabled, troublesome, annoying) kid and fobs him off on the state.
Is this so bad? Having the burden of supporting disabled children not land randomly on some parents would be good. In the past a lot of these kids would have been exposed to the elements.
I was reading more about the relinquished older children in Nebraska and as best I could determine, most of them had serious mental/behavioral issues that the parents couldn’t manage. Keep in mind that if you exclude the siblings with the dead mother and overwhelmed father there were only about 30 children.
I have a friend who has worked in child support for many years. Apparently there are a few men who have some mix of minimum-wage employability and like eight cases by five women. She and her coworkers used to joke that they should put a vasectomy clinic in the basement of their office, and just send the dude down there after his second separate judgement.
>She and her coworkers used to joke that they should put a vasectomy clinic in the basement of their office, and just send the dude down there after his second separate judgement.
Likely that the women would have gotten impregnated by some other loser.
The smallest bank in the United States is gone. Last week federal regulators shut down Kentland Federal Savings & Loan of Kentland, Indiana, which had $3.73 million in assets and $3.65 million in deposits. It had been founded in 1920 and remained in family ownership ever since.
When you want to call a friend, how do you answer "Who's free now?"
I made a little mobile app that connects you with a friend in less than 30 seconds. It’s called Beacon. You send a beacon to a group and everyone gets pinged at the same time. The first person to answer the beacon gets connected for a 1-1 audio-only call.
(The signal drops silently for everyone else, no pressure to answer or missed call badges. There are quiet hours settings and of course you can manage inbound pings with OS-level focus modes)
It works best if join with at least four people you don't speak with as much as you'd like. I have a couple dozen connections on the app now and it feels like magic to me.
Would love feedback from ACX-ish people who still like phone calls, or wish they had more of them. You can get the iOS version on TestFlight here:
Never mind about ios vs. droid…If I have some friends i communicate with via my phone’s native calling feature, and others i communicate with via WhatsApp, can your app handle only one of those? Both?
Beacon uses VOIP so if your device was connected to WiFi it would not use your cellular data plan. If you were not on WiFi it would use cellular data at an estimated 16MB per one hour of calling.
Is this an actual problem that people have? Maybe I'm just old but if I want to talk to someone then I call my friends in order of how much I want to talk to them until someone picks up. How common is the problem of "I need to talk to one of these 20 people and I don't care who it is and I have no time to dial them sequentially"?
It’s a big problem for many people. Most people don’t answer calls most of the time! Calling twenty people and waiting for it to ring out (and reaching nobody) would take ten minutes of dialing and waiting. And then there would be twenty people feeling like they owed a callback.
What if you’ve only got 15 minutes to chat but spend half that time dialing?
If you’re truly old maybe you are time rich and so are many of the people you’d talk to, but for people who work and have kids and other commitments, it’s very hard to catch up regularly and spontaneously with all the people they might like to, but there are also lots of little moments that could work, and Beacon finds them.
Well then I'm definitely not in your target market. I have to say that I find it difficult to identify with the scenario as you describe it. "My time is extremely valuable" seems incompatible with "I don't care who I spend the next 15 minutes talking to." But I'm middle aged. Maybe social norms have changed.
Still, I can see how it's a clever idea even if it doesn't solve a problem that I identify with. I didn't think social media would be useful either. Best of luck.
I believe you posted this before, and I can't help but have the same thought. It seems like an excellent premise for a Seinfeld episode. Why not just start a group chat via text, and then nobody will have to feel like they got left out?
Group chats are a very different use case and none of them solve this problem! There is a reason nobody combines all their family and multiple social groups into one massive group chat.
Fair enough, but I still can't help thinking that this is somewhat like tossing the apple of Discord into a room. I get a Beacon and I have to pee, so I wait and then I call, but I got beaten to the punch. Well, did he really want to talk to me or just anyone? What are they talking about behind my back? I don't know, maybe I'm just very old-fashioned.
In order to contribute to the quality of our perennial discussion of the topic, I will henceforth refer to contemporary architecture as "Corbusmaxxing Bauhauslop." You're welcome.
Can you explain the "Corbusmaxxing" part? Is this a reference to the Swiss architect Corbusier?
It'd be at least interesting to say Bauhaus's influence ripples through all modern architecture and makes it slop; I'd disagree off the bat but be open to hearing your arguments.
I think your biggest obstacle is going to be getting me to believe that all contemporary (last 50 years? last 100 years?) architecture (residential AND commercial, all subcategories inclusive?) fit in a single bucket, stylewise.
That would be counterintuitive to me. I thought a component of the complaint about modern architecture was that regular people like it (in the same way regular people go in for all kinds of ugly trends, for example the ridiculous false dormers and cascades of 6"-deep gable ends you see on the facades of production houses built in the last two decades) but elites who appreciate True Beauty see it as the ugly trend it is, yet here it would mean Brendan Richardson has excluded the stuff regular people like. What's left would then be "Corbusmaxxing" (so, lots of concrete?) Bauhaus-slop (which I associate with brick facades, rounded corners, and block glass masonry windows for some reason, though that might not be the intended image) that is beloved by the elites.
I guess I find the discourse on architecture confusing in general and this is just an example of why.
I tend to look at buildings from more of a technical standpoint--use of materials, performance, serviceability, human factors, etc. Whatever emotional or artistic vibes the buildings give off when looking at them from the street is not as important to me, so criticisms of those aspects often end up seeming like a hall of mirrors.
I think I would recommend you study some pictures of the Palace of Westminster and the Scottish Parliament Building until you can intuitively understand why most people have an intuitive preference for one over the other (and which one that is).
It’s tough not to feel like AI has unfairly changed the rules of the game of life. It’s difficult to pour myself into non AI related projects in the same way I did in 2021. Like, if we are three or four years from superintelligence, shouldn’t I just be working on that, even if I’m woefully unqualified? I even find it suddenly tough to care about politics and similar. Who cares about immigration policy if the game board is about to flip? This has caused me some minor mental distress.
I will, apparently, forever recommend C.S. Lewis's short essay "On Living in an Atomic Age" where he points out that nuclear weapons just added "one more chance of painful and premature death to a world which already bristled with such chances and in which death itself was not a chance at all, but a certainty".
> If we are all going to be destroyed by an atomic bomb, let that bomb when it comes find us doing sensible and human things [...] not huddled together like frightened sheep and thinking about bombs. They may break our bodies (a microbe can do that) but they need not dominate our minds...
And, yeah, I think there's a real threat to AI... but most of us can't do any more about it than we can about people with their fingers on Big Red Nuclear Buttons - and if you can care about immigration policy in a world with nuclear weapons, it seems like it should be possible to do it it in a world with a potential for superhuman AI.
Every few years I rediscover that my anxieties are deflected from other areas of life to the One New Thing that I've noticed.
I was stressing out about AI and wealth inequality to the point where I would stay awake until 3am sleepless. Then I started exercising and hanging out with my friends again, worked a little less late into the evenings, my stress dropped back to a sensible level of worry, and I realised that I had been taking the result of a number of different stressors and assuming it was all AI. Exactly like I did last time, except with job related stress, or the time before that, with friends, and so on back into the past.
I am not saying this is the case for you, but maybe this comment will resonate with someone and help them find perspective again.
This is the idea of free radical anxiety in a nutshell, and I think everyone suffers from it. I have certainly spent a lot of my time in life being anxious, and the more I have worked on it, the more I have come to the conclusion that it is almost entirely self-generated. There is some essential anxiety in me that freely attaches itself to things that I think about or imagine. These things then become a smoke screen that hides the real source of my discomfort. I have developed my own mantra to help me navigate this. "The definition of anxiety is knowing something to be true but refusing to believe it." (YMMV but I think it's a defensible construct.) I came to the conclusion that one of the biggest causes of my anxiety was fretting about things that I really had no control over. That's a difficult thing to believe but in many cases, it's true. (For example, I spent an inordinate amount of time keeping myself awake wrestling with all the misery that was in the world. Needless to say, that made me miserable. Now I have adopted a couple of homeless people that live on my block or in environs, and make an effort to communicate with them and give them some money when I can. It's not like I'm solving the world's homeless problem, but I'm doing what I can, given my reach.)
Your programme of more human contact and more involvement with your physical nature is a good antidote.
Settle down. AI will be significant but it won't be everything. It will propagate slowly through the economy. There will be many roadblocks, both technical and political. Businesses take many years to adapt to new technologies and workflows. Electricity took 40 years to fully penetrate the US and that was a much more valuable innovation. The real world is complicated and things take longer than armchair philosophers appreciate. Just relax. Intelligence doesn't mean omnipotence. The world will change, equilibria will shift, but life will go on.
I would agree with you on a lot of this, were it not for the fact that AI is going to have and is having a profound effect on warfare. It's interesting that AI and the birth of the Internet are rather opposite in the sense that the Internet was essentially begun as a defence-related project that also became very useful in the civilian world. I feel like AI is going in the opposite direction. Something that has been developed by private companies with a profit motive is now really too important to be left completely in the hands of civilians because of its defence implications.
If we're actually going to get superintelligence any time soon, you *are* working on it by contributing to its training data in whatever field you're working in.
Just make sure to do excellent work, and document it well.
We're not going to get superintelligence in the next five minutes/five years (maybe not even the next five decades).
AI will make money, but that money will go where it always goes in new technologies, and we are not going to see "glut of productivity means enough tax revenue to pay everyone to sit around and play the bongos". If you have a job, and manage to keep that job, you will need to keep working that job to pay for the cost of living, just as you do now.
You'll need a raft of new skills/upskilling, as with the past drift to credentialism and 'employers want freshly minted employable staff out of high school with these, these and these sets of skills' emphasis in education, just to get/hold on to that job. Same old same old.
The game board flips, and you will still be breaking stones:
How screwed is the world right now, in an economic sense?
Specifically, assuming the U.S.-Iran situation remains at its current impasse for at least a few more weeks or months, and the Strait of Hormuz sees either no commercial traffic or sharply reduced commercial traffic, how bad will the global economic effects be and when will they start.
My overall picture is like this:
1. The strait effectively close around the beginning of March.
2. Because it takes on the order of a few months for oil on an outbound tanker to turn into refined product available on the consumer market, the direct effects of this weren't felt immediately.
3. But by the same token, even if full traffic resumed tomorrow, it would take some time for production effects of the closure to be fully reversed.
4. The period of U.S.-Iran cooperation following the signing of the MoU saw a significant increase in traffic while it lasted, but it never came close to normal. So the total amount of oil (and other key materials) shipped during that time can only amount to a few days' worth of normal production.
5. The only reason there haven't been larger price spikes yet is because of governments putting their thumbs on the scale in ways that aren't sustainable or don't especially help the long-term picture: things like releasing reserves and lowering fuel taxes.
To me, all of this together suggests a picture where when the price stabilization measures start to fail, they'll fail hard. But I don't have high confidence in this, and I don't have a great sense of either timetables or magnitude. How long before prices spike and how high will they spike? How bad will the economic fallout be and how long will it last? I've seen this issue discussed *somewhat* in the last few months, but far less often and less thoroughly than seems appropriate given the scale and scope of the potential consequences.
I have no idea. Oil prices are going up. https://tradingeconomics.com/commodity/crude-oil So yeah things are getting worse. But I think almost everyone wants this to work out so I'm not that worried.
I admire your optimism, but I don't think "everyone wanting this to work out" (depending on what you mean by "working out") is the case. There is a real schism in Iran between those who want to go to the mat with the United States and those who want to make a deal. As of now, it looks to me like the faction that wants to double down on this is in the ascendancy. I think the same schism exists here in America, but it's not quite as rowdy. There are a lot of people in Iran who are willing to die to defend their country, and there are not nearly the same number of people in this country that are willing to die for rather fuzzy reasons. A scary but rather abstract fear of Iran getting a nuclear weapon does not compare with seeing an American or Israeli bomb kill your children. Trump has kicked a hornet's nest, and the way forward is unpleasant, no matter how you go. If the other Arab countries in the region were willing to go to the mat against Iran with the United States, that would be something, but I don't think they are. I see the danger now as turning a war of choice into a war of necessity.
I think it is rather telling, at least symbolically, that a young man who was beaten severely protesting the Iranian government when that was going on has now joined the army.
Yeah I have no idea about Iran. I meant most of the US. (I guess there are a lot of 'war hawks' in congress/ government.... we could vote those a-holes out of office.) TBH at this point I don't really care that much about Iran getting a nuke.
" But I think almost everyone wants this to work out so I'm not that worried."
I don't think this is remotely true in any useful sense.
First and foremost, Netanyahu doesn't especially seem to want it to work out. The combination of Iran's commitments to Lebanon and Lebanon's proximity to Israel give him a near-veto on the peace process (though Washington can try to pressure him into not using it).
But second, I think Trump only "wants it to work out" in the most superficial sense that if you have him a button that would restore the status quo ante between the U.S. and Iran at no cost to himself, he'd push it. The war has clearly not worked out like he hoped it would, and he'd much prefer it stop being important.
But we can trivially see how little real will their is behind is "wanting it to work out," by the simple fact that it *hasn't* worked out. What was obvious to make it work out was obvious quite some time ago. Washington simply needed to sue for peace, meet Iran's terms and deal in good faith, recognizing the plain facts that a. the Americans and Israelis were inarguably the aggressors and b. the Iranians had the upper hand, and with it a large (but not unlimited) latitude to dictate terms. Sticking to the plain intent of the MoU as it was announced in June would have been sufficient, if Trump had been willing to do that.
It's hard to be sure exactly what happened instead, whether Trump intended to deal falsely from the start or whether he simply chafed at the (largely self-inflicted) humiliation and listed to bad advice about what to do about it. But either way, the ceasefire collapsed because
1. The deal was made in a way to be contingent on Israel's actions, without Israel actually agreeing to it. Israel naturally didn't consider itself bound by such a deal. [1]
2. Washington and its allies repeatedly tried to undermine Iran's negotiating position mid-negotiation, by routing ships along a route that Iran hadn't agreed to, one that would have been strategically advantageous for Washington if it became the main route. This was a pretty clear act of bad faith.
3. When Iran responded to the above with (admittedly provocative and belligerent) act of force, Washington immediately switched back to trying to intimidate and out-escalate Iran instead of making a more measured and restrained tit-for-tat response. This had the predictable consequences.
I think 3 especially reeks of poor impulse control and someone whose desire for things to "work out" all but disappears the moment their temper rises.
This, in a nutshell, is why I'm pessimistic for any actual near-term peace deal (and remained pessimistic even when the ceasefire was still holding). Succeeding at a complicated negotiation from a position of extreme weakness--a position the pretty much requires repeatedly showing restraint and eating crow--is simply not something I think Donald Trump has the mental capacity to do. I think he was scared to the negotiating table by dire predictions from somebody in his inner circle, but even when the dire predictions start turning into even-more-dire realities, he's just not temperamentally capable of the things that he needs to be able to do to fix this. This is, let us remember, a man who threatened the Iranian negotiating team with violence in front of the entire world.
I think peace is only likely once Trump is no longer able to sabotage the negotiations, either because he's no longer occupying the office of president, or because some combination of election results and mounting popular pressure force congress to grow a spine and reassert themselves as an independent branch of government.
[1] The U.S. probably has enough leverage over Israel to have made it stick anyway, but it would need to be willing to use it quite a bit more strongly than it has in the past.
Well Trump will be out of office for many years before congress grows a backbone. What's the incentive? We the voters must demand a congress with a backbone.
Anyway, by most I mean most americans, no one likes this war and we are done with the Trump lying. As I heard someone say, "All this winning is killing us, less winning please." It's hard to see how the war goes into the winter and home heating oil season. But I don't really know anything.
Well, you'll want to start with the baseline: 8 billion people in the world, and you have, say, half a billion in some form of famine (this is, ahem, normal -- wars suck, and there are kids who die in India from malnutrition every year.)
America is having drier months than the Dust Bowl, and wheat's not been planted (you can pull the year over year). Russia's also down on wheat planting, and Ukraine is at war, so there goes the Middle East's breadbasket (egypt is a net food importer nowadays) Now, the next risk is not having diesel/gasoline to harvest. And rice not getting planted.
You're looking at everyone on the edge dealing with "higher prices" -- that's Sudan, Yemen.
And East Africa's got drought too. ENSO and the Iran War are doing a number on some of the least well-off places in the world.
Qatar and other countries on the gulf don't get to move food by boat, so they're stuck using gasoline (which is scarcer than you'd think, not all of these places have working refineries right now, Iran certainly doesn't), or flying the food in (which is expensive, even if they do have "oil revenue" buffers.)
Iran itself has 90 million people or so, and they're under "water famine" (can't grow crops if you don't have enough water to even support your capital -- and it's difficult to get any food into the country).
Again, this is worst case, but it's a pretty bad worst case. Israel is fond of saying "Israel has a right to defend itself" -- if it gets really bad, Israel may find itself pitted against Article 5. (Which is to say, Israel is allowed to "defend itself,"** so long as it's defending itself against international pariah states).
**Defined as clearing entire regions of Lebanon, and saying "the Lebanese will never be allowed to move back."
Sidenote: just skimming, I mark about 40-50 million dead from famine in World War II (the differential is mainly from China, whose statistics nobody should ever trust). And yes, that's over four years.
Interesting failure of AI capabilities. I asked Fable to compose two verses of We Didn't Start the Fire for 2000-2004 and got this:
Y2K, hanging chads, dot-com crashing, Bush v. Gore,
Towers fall, anthrax scare, boots on Afghan soil for war
Enron books, iPod drops, euro cash in every store,
Wikipedia, "Shrek" is green, Ozzy's on TV at four
SARS outbreak, shuttle lost, Concorde makes her final flight,
Saddam's caught, "Rings" sweeps gold, human genome read outright
Facebook launched in Harvard dorms, Athens flame is burning bright,
Boston breaks the Bambino's curse, tsunami on a Christmas night
If you try to sing this you will notice it doesn't scan at all. My version for comparison:
Y2K, Bush v Gore, Tiger wins all four,
Richard Hatch, Concorde crash, Wikipedia
Anthrax, Dot Com, iPod, Yemen bomb
North Tower, South Tower, US in Afghanistan
DC, Shock and Awe, Euro notes, shuttle flaw
Facebook, Abu Ghraib, Red sox curse dead
Schwarzenegger, DNA, Janet Jackson, Boxing Day
Kill Bill, Spotless Mind, Ronald Reagan, goodbye
Interestingly, if you ask it in a new chat, Fable thinks its version is better than mine, precisely because it thinks its version is "more singable"! So it seems to do very poorly at understanding rhythm and melody.
Its not really about the syllable count. (I agree my lines are on average slightly shorter than the original, although the original does have 9 syllable lines "Moonshot, Woodstock, Watergate, punk rock" though only one and I have two here). Its about the placement of the stress. Try to sing Fable's line that starts with Wikipedia. It can't be done to the rhythm of the song; the 'a' sound at the end has nowhere to go. Notice how I put wikipedia at the end of the line, like Dimaggio in the original. I don't really know enough about music to say exactly why this is required, but it is. And even I, a non music expert, can hear this. The lines that start Facebook and Boston have the same problem. The emphasis doesn't start in the right places. Fable also got the rhyme scheme wrong. It did AAAA end rhymes instead of internal rhymes on the 1st and 3rd line of each verse, which are required to make the song sound right.
There's definitely places where the stress is gunked up, and "Wikipedia" sticks out like a sore thumb for having too many syllables to fit the meter without elision. I don't see the issue with the Facebook line, though?
Part of my job is writing, and I sometimes send people AI-written text for review without disclosing it, just to see whether it's good enough (it never is, even though it has been provided plenty of examples and explicit quality criteria).
Today, I gave two texts about the same topic to five "blind" human reviewers, one written by me, one by Fable. I also gave these texts to a separate instance of Fable as well as to Gemini 3.5 pro.
All human reviewers independently felt strongly that my text was better across all dimensions. And what made me share this anecdote was your last point: In my case, Fable and Gemini *also* both preferred the AI-written version! They explained very confidently why it flowed better, felt more natural, the examples were more relatable etc., when the human reviewers felt the opposite.
It looks like reddit has finally decided to make itself more-or-less inaccessible to those without an account.
First, the mobile version of the site started stopping you from reading it after some short amount of scrolling.
Then, they disabled old.reddit.com completely unless you're logged in, both for mobile and desktop.
You can still read new reddit on desktop to your heart's content, but they may well be planning something around this as well.
It really made me realise how much of my daily news and general mental stimulation I was actually getting from reddit, despite the fact that I hate it. I'm not sure where I'm going to go for my cheap mental stimulation in the future.
So reddit is definitely making themselves less accessible, but I was just able to reach old.reddit.com in an incognito window and go through a couple pages. What do you mean it's disabled?
Alright I guess it's still some kind of A/B test going on. It was working intermittently for me a week ago, but nowadays I just get "Log in to use old Reddit" every time I try.
Well hopefully every time you get that you are deciding to go to some other website so that the signal they get is that making that move would reduce their user numbers and have very low conversion to logged in users
I assume they have realised they are no longer a company that sells eyeballs to advertisers, they are a company that sells human-generated text to AI companies. a
I'm fond of Reddit. There's a lot of raw human truth in the answers, though of course mixed with the various forms of bullshit, And it is impossible to be the crassest person posting.
> It really made me realise how much of my daily news and general mental stimulation I was actually getting from reddit
I was in the same boat when Reddit jumped the shark by my own metrics, and I find Substack to be pretty good. It's like "smarter non toxic" twitter, but also with full length articles. And it's pretty good about clustering your feed with similar authors / people that you've already followed or liked.
You might also give hackernews and Mastadon / Fediverse a try. Mastadon is higher friction (you have to manually find nodes with channels worth following), and the content was ~10x - 100x less than Reddit or Substack back a few years ago when I was trying it, but there were some good things there.
I think Substack is benefitting from being new. It'll get corrupted like Reddit as it gains traction, because there's nothing about its business model that gives it innate resistance to brigades, echo chambers, and so on. It might not become a monoculture as long as its model encourages creators to intellectually own their spaces, but there's similarly no innate resistance to people leaning on Substack to de-platform this or that creator until most creators lean in one direction. It'll continue that way until someone pulls together enough capital to create the spiritual successor to Substack, and the cycle begins anew.
The Fediverse truly seems to have put effort into bias resistance, by contrast. AFAICT there's no Mastodon site "brand"; it really wants to sink into the background and bake decentralized influence over content into every module. Even if someone tried to enter the system and lean on people, it couldn't reach everyone.
Yeah, I think you are probably right, and it hurts, because I really like Substack and will hate to see it go.
On Fediverse, I think the main problem is just volume / content, is that still the case? It's a very "hackernews" and "old slashdot" type of crowd. Which is fine for me, those are in a sense very much my people, but that's like <1% of the population, and they just don't produce the diversity of thought, content, and viewpoint that the bigger platforms do.
I don't know how Fediverse fares on volume. This is a tricky question, because it's practically impossible to know. There's no central registry of "hey, I'm a Fediverse server, I'm run by Jan, and I host content about Croatian politics" or anything like that. By design. Some dude might put up a server and encourage other servers to advertise his, but they don't have to, and he might want to be unknown anyway. A server might be an online smoke-filled back room for planning criminal activity. An underground network might spring up, using the common protocol for convenience, but being organized as cells of no more than 3-4, constantly shifting location, and if authorities with lawyers or guns find one, the others pull up the drawbridge and re-organize on different IPs.
There are multiple crawler tools, but they're not guaranteed to find everything, because the Fediverse is not required to be completely connected. By design.
When Leftists do dynastic politics, they're embarrassed enough to pretend the dynasts have qualifications other than who they're related to. Trump sees no need to do that because he has a sleazy third-world mindset.
Either way the actual call here seems reasonable; it's an appointment for the next six months (this seat is up for reelection then), and Graham's sister is a plausible neutral figure to do roughly what he would've done who can be reasonably trusted not to run for a full term herself, letting Republicans run a neutral primary without the governor having to put a thumb on the scale (which he presumably doesn't want to for state politics reasons).
Slime Mold Time Mold has a post about different human (dietary and other) drives existing, but the number of them being unknown. This reminded me of a post by I think Ozzy, saying that the urge to eat ice was often a sign of iron deficiency. This sounds to me like we have a "consume iron" drive that sometimes misfires into a "consume ice" mode
Remarkable that no matter how crazy and low-class the Republicans become, Democrats ALWAYS have to do something to piss off the nation's wealthiest people for no reason.
All of the data center companies should come up with a national plan to solve this. My proposal would be to pass a federal bill which requires data center companies to fully take on the cost of energy (either building their own or paying utility companies enough so that rates don't increase for civilians), noise limits at a certain radius which can be addressed with sound baffles, and a formula for the allowed annual water usage (based on reserves, community usage and other relevant criteria). In exchange, they get permitting reform to streamline the approval process for building, and acquiring energy, and a federal pre-emption of state level bans.
That data ends in 2024. In any case, perception matters more than reality in politics. Currently, 70% of Americans disapprove of AI data centers, and 66% cite electricity costs as a reason. Maybe you think that is because people are irrational and stupid. And yeah they are. But you can't exactly change that, so what do you do when you're at the mercy of the politicians that they are about to put into office?
I don't know if some token legislation addressing people's stated concerns will help (when the reality is probably anxiety over job loss, unrelated to actual facts about data centers), but it wouldn't cost that much and it could even be beneficial on the net if it speeds up new energy projects. And it seems to me that the alternative is probably new data centers getting banned in most of, if not the entire country.
What is the relationship between data centers and electricity prices? Honestly asking.
It would be counterintuitive to me if there were none, having worked in both an electrical utility and data center construction, and I’d want to make sure I understood why if that were the case, but I’ve been surprised by other things before.
If there's consistent excess generation capacity, then utilizing the local grid more intensively could reduce average cost (i.e., more variable charges to amortize fixed costs across). Similarly for any other situation where the existing grid scale is below the cost-minimizing size. Residents should welcome DCs.
If there's an average excess but intermittently binding capacity constraints, it'll depend on how the utility is allowed to charge. A possibility would be that the DCs draw a steady load and pay a flat rate that includes guarantees for prioritization, which would leave the preexisting users facing a more volatile remaining supply (e.g., lower average prices but with unpredictable spikes, and/or rolling blackouts). Mixed bag for residents, depending on their resources for absorbing volatility.
Then there are areas where the grid is roughly fully loaded. Without additional investment the increased demand from DCs would just drive prices up. Residents likely to be adversely affected, unless a competent & benevolent bureaucrat or politician (good luck with that) negotiates a positive-sum investment in the grid by the DC builder.
So, how do you know what kind of area you're in, and how do you know how much of the relevant type of effect data centers are having on your electricity prices?
There's a kind of catastrophic cycle where people believe that a policy that only makes a problem worse will solve it, thus doubling down on it more the worse it becomes in a vicious cycle. The classic case of this is rent control (also recently doubled down on by New York politicians), but seems like energy prices are going the same way.
Pissing off the nation's wealthiest people is a reason in and of itself. It's fighting the good fight, and Democrats would be a lot less lame if they'd do it more. The U.S. seems to have this bizarre collective Stockholm syndrome where it watches small numbers of people act like idiots and children and assholes and says to itself, "obviously these people must be amazing because they have money," and contorts all it social and political life around trying to idolize and please and emulate them.
I think AI data centers specifically are an excellent fight for Democrats to be fighting[1]. AI is a suite of technologies that could be both incredibly powerful and incredibly dangerous, and the big AI companies have done very, very little to inspire confidence that they'll handle it responsibly or use it for the public good. Having the public tell them "no," occasionally and seeing how they react to not getting absolutely everything they want is pretty much the bare minimum that a functional polity ought to be doing about it.
[1] Though that being the case, I fully expect them to roll over or half ass it as they always do.
Seems to me people don't idolize all wealthy people just for having money, but instead they idolize certain wealthy people (for being the big heavy hitters on their team) and demonize others (for being the big heavy hitters on the opposing team).
Like, when it looked like Elon Musk was buddying up with Trump and doing Hitler salutes and stuff, people who liked Trump liked Elon Musk. But those same people hate George Soros.
It's all based on tribal identity, which is what I always say, because (at least as far as I can tell) it's always the best explanation.
If a wealthy person wanted a free pass to build data centers or whatever else without Democrats bothering them, they'd simply have to build a public identity as being very left-wing or otherwise Democrat-aligned, maybe with a little extra edge to demonstrate they're not just lukewarm about it. Then everyday people who vote D would support everything this person does and says, especially if it means defending them from criticisms by people who vote R.
That'd be interesting. If you force the billionaires to actually be hands-on and man the cart themselves, and deny them the right to just turn it into a normal business they hire other people for and run from an office, it could go one of two ways:
1: Billionaire thinks "Well, I've got the cost of my overhead (cart, ingredients, payment system, personal security detail, etc.) and my labor (thousands? tens of thousands? of dollars per hour)" and they price that into the hot dogs, make no money because nobody is willing to pay $9,999.95 for a hot dog, then write the whole thing off on their taxes so we all end up subsidizing it
2: Billionaire thinks "This will look like a publicity stunt, like that lawyer who opened a hot dog stand...I can afford to give away hot dogs as part of my stunt, so I'll just do that" and it basically becomes a "line up to shake hands with a billionaire and get a token hot dog you can eat or throw in the trash afterward" photo op kinda thing
I'm honestly not sure which outcome is more likely, but now I'm hungry
"Seems to me people don't idolize all wealthy people just for having money, but instead they idolize certain wealthy people (for being the big heavy hitters on their team) and demonize others (for being the big heavy hitters on the opposing team)."
There are people who think like that, yes. But you seem to be mistaking "some people think like that" for "all people think like that."
"If a wealthy person wanted a free pass to build data centers or whatever else without Democrats bothering them, they'd simply have to build a public identity as being very left-wing or otherwise Democrat-aligned, maybe with a little extra edge to demonstrate they're not just lukewarm about it. Then everyday people who vote D would support everything this person does and says, especially if it means defending them from criticisms by people who vote R."
As much as I hold large swathes of the Democratic Party in utter contempt, this is just demonstrably not true. Democrats are not fantastic at holding themselves to real principles, and I wish they were better at it. But they do it more than zero. In fact, in the last two years they've done it in numerous cases of the highest profile imaginable. Now the Democratic leadership has been disgustingly much better (in at least some cases) at closing ranks and holding the party line against their own voters than they usually are against Republicans (though that wall finally seems to be crumbling). And I've definitely seen and even talked to *some* rank-and-file Democrats who do think this way, especially older ones. But in coalition-based electoral politics you need almost all of the coalition to think this way almost all of the time, and again, there's been strong evidence over the past few years that this just isn't true of Democrats as a coalition.
It does seem to be true of Republicans right now, as much as I wish it weren't. I honestly do not know what real-world actions it would take for the Republican base to turn on a public figure who consistently parroted the right tribal messaging. But we already know that neither protecting pedophiles nor blatant corruption, nor actual treason are sufficient cause, so the existence of any line seems speculative at best.
My impression is not that *all* people think like that, it's that the vast majority do, enough to where I could just write "people think like that" and it's basically true. But I know this is only my impression and I could be wrong.
>Having the public tell them "no," occasionally and seeing how they react to not getting absolutely everything they want is pretty much the bare minimum that a functional polity ought to be doing about it.
They might react by giving money to J.D. Vance to make him President of the United States. Why do Dems have this self-destruct instinct?
You don't think Republicans have the same self-destruct instinct?
"As long as it feels like it punishes my outgroup now, I'm in favor of it, and let's not think too hard about what it does for my ingroup later" seems pretty common across the board
Voter's values are not that malleable. A massive political campaign can shift the outcome on issues where public opinion is split or people don't feel strongly. But not on a hot button issue with a 73-27 split (it'll probably be even more dramatic by 2028).
Vance will either have to take a more hardline stance against AI, or lose the election.
> Pissing off the nation's wealthiest people is a reason in and of itself.
Seriously? You know ~75% of billionaires are self made? That they've directly created ~10M jobs, and that's just direct jobs, not any of the many millions more enabled by being their suppliers, contractors, customers, or affiliates?
That "tech oligarchs" are the source of essentially all US economic growth over the last TWENTY years, not just the 7 most charts show?
How dare those scoundrels create tens of millions of jobs and drive nearly all the economic growth in the richest and most successful country in the history of the entire world!
Why, we'd be SOOOO much better off if they'd never existed at all, and we were tens of millions of jobs and many trillions of dollars poorer, and didn't have Amazon, Google, Netflix, or Walmart!
Oh, and all your political heroes have gotten rich off of them, too. Democrat politicians reliably get higher insider trading returns than Republicans, clocking in at 40%+ returns per year vs the R 29%: https://imgur.com/a/u5D9C9B
That might tell you something about their actual priorities and alignment, if you paid attention to their incentives.
Are you reading the Forbes list again? Most billionaires are not listed there. No, most billionaires are not self-made. Most billionaires are nobility, which tends to take exception to being listed on Forbes.
I'm all for saying "American Billionaires" tend to be better, but...
Kochs aren't "self-made", Scaife wasn't self made (and SBF was -definitely not- self-made). These are billionaires (and heavily involved in politics).
(I'm also not headhunting Musk or his ilk -- some people really do get there by hard work and honest effort).
> Are you reading the Forbes list again? Most billionaires are not listed there.
Yes, I actually wrote an article about how many billionaires there really are worldwide.
And you can do top-level economy triangulations with the power laws / incidence seen in the rest of the world as a check to see that there are some really hilarious mismatches. I ended up at a little over 2x the listed billionaires.
As you may have seen from some of the comments, Democrats include a large number of people who hold that pissing off wealthy people is a virtue, on faith. There are multiple reasons for this, and they're well-traveled reasons: failure to value management, failure to value networking, failure to understand price theory, sheer envy, etc.
Democrat party leadership includes a large number of those wealthy people, and some of them _do_ understand price theory and management and networking (not to mention how much it actually costs to run a political party), and they probably find themselves riding a line between courting the people who don't and not pissing off the ones who do. Their strategy seems to include trying to make some sort of distinction between "good wealthy" and "evil wealthy", but that's an epicycle, whether they realize it or not.
It's tempting to embrace eschatology at this point and hope for a time when they all figure out where this leads. I prefer the more boring option of continuing to explain, to whomever will listen, since it's based on the premise that most people are smart enough, and deserve the incentive to lean into that. Plus, you never know when they'll have an insight of their own that you missed.
Hate for AI data centers is bipartisan at this point, and decisively so. At least among voters, which will quickly transfer to elected officials after one or two election cycles. I do not understand why data center companies aren't treating this as a threat to their continued existence.
Anyone have recommendations for a rationalist-adjacent counselor licensed in New Mexico? Telehealth is fine, and second-degree referrals (some you know/your therapist has a rec for someone in NM) are also welcome !
Being smart is not as fun or exciting as people imagine. You are used to being smart, so your mental content is no big thrill, just the regular taste of spit in your mouth. It doesn't feel special at all, it just feels like living and thinking. Whereas you never get used to failing to fit in. It's freshly bad every goddam time. Know what I mean, honey pie?
Yeah, I'm in my late 50s, so I don't forsee my mental sharpness getting any better, though I do a very intellectually demanding job and I don't seem to be hitting "can't do that anymore" or "can't learn that" limits yet.
Same, and I'm in my mid-50s. There's more and more to learn and improve at, and the bottleneck I feel is finding the time. Sometimes I even get annoyed at the fact that I have to stop and eat something, or sleep.
Yeah, haha, I drank 2 beers which made me dumb enuf to post what I did last night, and people are showing way more interest in what I said than they do my serious posts. Have all you guys been drinking beer too?
I hate the scenario where the smart people can't fit in, so they get dysfunctional, and when they later meet other smart people and finally have a chance to fit in, there is too much accumulated dysfunction that prevents a reasonable outcome.
Mensa is the textbook example, but to a smaller degree everything.
Your mileage may vary. In my experience, being smart can be extremely satisfying or endlessly frustrating; what makes the difference is the opportunities available to you.
The middle class might look OK in raw economic data, but actually lost material non-financial assets, and replacing those in the Free-Market is very expansive, so in fact the middle class is Poorer today than our parent's generation, and because the lost assets were not financial, it is not reflected in Statistics.
TLDR is still true but your summary intrigued me. I read it and thought, "Oh, like how every middle-class family used to have a piano in the living room, and now maybe one in ten has an old cheap keyboard in a closet in the basement somewhere, and even getting a good Yamaha keyboard with weighted keys is probably not feasible for them because they're all drowning in credit card debt and what do they even want one for because they have Spotify subscriptions?" But I have no idea if that is a good example of what the article is talking about.
1) Public Schools at most places were "Good enough", no need for private schools.
2) Kids could walk around, for example for School, or to playground. No need for big Yards or driving them around, or for so many extra-curricular activities.
1) Public Schools at most places were "Good enough", no need for private schools.
2) Kids could walk around, for example for School, or to playground. No need for big Yards or driving them around, or for so many extra-curricular activities.
The public / private school thing seems misplaced to me, as someone who believes private schools are superior _if_ parents are not forced to pay for public schools. In other words, a system where parents can either pay a set rate and use a public school is worse than one where parents can pay to use the private school of their choice, since they get to navigate the quality-price tradeoff themselves. Lamenting private schools strikes me as lamenting that everyone has to choose their own flavor of ice cream when vanilla used to be good enough for everybody. (Mandatory taxation for public schools confounds this.)
The complaint about kids' freedom to roam seems legit and sympathetic to me.
Overall, some people disagree that the upcoming generation has it worse off, but the people who agree seem to come from both sides of US politics, so that's something. That said, there's plenty of disagreement on a remedy.
A society where only children from well-off backgrounds get good educations is one I expect to descend into a gilded-age dystopia within a generation or two.
Parents are not the sole beneficiary of children's education. The children themselves benefit, but rather more important than that for this context is the society itself benefits. And most especially, it is the children of the parents who are the least concerned with getting their children a good education that we as a society should most badly want to have good educational opportunities. No altruistically, for our benefit. Selfishly, for our own.
A society where only children from well-off backgrounds get good educations is one I expect to quickly evolve into one where most children get them due to their parents seeing the obvious benefits to their own status, not to mention the benefits of having a child who goes on to a highly-valued and therefore high-paying job. And also due to a lot more people seeing the benefit in supplying good education to parents who demand it enough to be willing to pay, and hiring themselves out as teachers.
The real challenge here comes if the state somehow decides it's worth gatekeeping such educations or insisting they be universal even if they end up universally bad, and manages to persuade 51% of voters that that bad education is actually good (or, perhaps, better than the alternative).
I don't *really* know if the thesis he proposes is correct.
It does say something that sounds sort-of-true (Lost social capital), but I can't really verify, and provides an explanation for a known problem, namely that People feel poorer today, but this isn't reflected in the statistics.
I am, however, sympathetic to his arguments. If you read the article, you see he mention other things - like the breakdown in public trust drove people to more expensive neighborhoods as a way yo filter the "unwanted".
I didn't provide his entire argument here, just a short summary.
I take his point as being that the new generation is worse off in ways that don't show up as numbers. He's correct, but he's not the first person to point that out. I'm also pretty sure he's not the first to point out those specific issues (home prices, marriage rates, etc.). It's probably worth collecting some of those issues and showing them to yet another audience, but that's pretty much all this is.
It'd be interesting, extra-credit effort if he presented a workable remedy.
This appears to be a Substack (mis)feature, not an ACX feature, let alone a feature of OTs. When I read an OT and click on "new replies", those replies may open at the top or at the bottom, from one OT to the next.
With very few exceptions, OTs on ACX have been sorted newest first. I think the Substack owners set this but don’t know for certain. Might be something that occasionally slips Scott’s mind.
I noticed that the "report comment" functionality has changed. Previously you just wrote a text... and it was reported to Scott? Now you have to choose one of the provided options, and it is reported to a Substack team.
The provided options do *not* include "created yet another account to avoid a ban".
I noticed this a few days ago as well. There also isn't an option for "violated the norms of this specific substack instance, regardless of whether it follows general Substack rules".
In theory, "ban evasion" is a general thing that would make sense to add to the general Substack rules...
With other local norms, yeah, it would be difficult for an outsider to evaluate them. Substack should keep the old functionality together with the new one, so that global rule violations would be reported to global support team, and local rule violations to the blog owner.
Well this should be interesting seeing as how Trump recently piped up, apropos of his own unrestrained id, to tell our NATO allies that the 2020 election had been stolen.
“Trump is expected to make election conspiracies a focus of his national address”
Not sure if this is appropriate to essentially advertise here, but I prototyped a star convex 3d model encoding scheme and shader for rendering the encoded geometry with a single quad. I wrote about it here https://joem93.substack.com/p/3d-spherical-model-encoding. I usually just don't tell anyone about anything I do probably for fear of failure, criticism, etc, but in this case my desire to see what people do with this concept is stronger. If you know someone who might be interested in this, consider passing it along. I *have* to imagine this would be useful in some context within videogames
The idea is: you encode your 3d model by recording its surface colors and distances from a center point onto two images; then render it by raymarching each pixel, transforming your ray points to spherical coordinates, and sampling the radius texture using the angular coordinates. The first time you sample the model radius to be greater than your ray point radius, you sample the color texture and set the pixel's color to that color. I've been kind-of thinking of it like an upgraded (and significantly more expensive) billboard rendering scheme
My implementation is still a little under-baked because it doesn't calculate depth values. I guess that's an exercise left to the reader (and/or future me). There's also other issues like visual discontinuities due to raymarch step sizes (I've since learned you can apply a little random noise to step sizes to get slightly blurrier but more visual appealing rendering), and poles creating pinch-points in spherical coordinates. I really like the idea of using other maps from R^3 to R^2 but I haven't investigated it yet. Theoretically, if you find such a map (any such map), it would work just fine as long as it isn't lossy wrt your model's surface (ie when multiple surface points map to the same value). In the best (and probably most impossible) case you would directly map points to arbitrary UV coordinates, but that would probably take training an ANN or something, and almost certainly wouldn't be faster than the spherical version. Perhaps a cubemap map would work
You know, it's just now occurring to me the color texture alpha channel could be used to encode radius values, instead of using a whole other texture. You would lose transparency, but I'm not sure how transparency would even work with this
A purely anecdotal report on using coding AI’s to find difficult bugs…
The bug was that a buffer that was used as shared memory between two threads was declared on the stack, so that if the thread that declared the buffer finished first, the other thread could overwrite the stack.
Well done DeepSeek for finding that subtle stack corruption attack. (I imagine the other coding AIs would also have found it’
Does anyone know if there's a way to turn off 'also replied' notifications for Substack? I like getting notified when someone replies to me, but 20 notifications saying 'the comment section has people commenting in it' feels like worthless spam after awhile.
Is it too snarky that my first thought was "because paperclips don't get cancer."
Mine was similar; "yeah, death does all those." It's a lot of promises to stop the body from functioning.
Although full paperclipping does count as changing bodies completely.
Solid joke either way
Just the right amount of snark. Paperclips don't fear extinction, either, or yearn for a loving partner. There's a lot to be said for being a goddam paperclip.
Not I don't think so. So I like this idea that cancer and aging are two sides of the same coin. The 'coin' is that we're made of 100 million cells or something. And one of these cells going rogue can lead to cancer and the death of the whole thing. Life's response to this is the Hayflick limit https://en.wikipedia.org/wiki/Hayflick_limit which puts a counter on how many times a cell can divide. This limits the damage that a cancerous cell can cause (somewhat) But it also is the root of the aging process. Some cell can divide only so many times and more damage to the cell means it's death, you age. The details are more complicated than this simple picture.
I think it's one part of the aging process, but there's still a lot of other stuff, like accumulated damage beyond the ability of natural systems to repair, somatic mutations in random cells that can cause cancer but can also just make some stuff not work, accumulating persistent infections (approximately everyone is carrying around a bunch of latent infections from members of the herpesvirus family, which usually don't do anything but occasionally cause you to get some weird leukemia or have an outbreak of cold sores or get shingles or something), and so on.
Yeah sure. I just like it as a starting model. Aging is built into our cells. Re: accumulated damage, there was a post about aging and accumulated damage here a few weeks ago. I think that fits right in with the Hayflick limit, and again sure the details are messy. I get most of this from Bret Weinstein and his "Reserve-capacity hypothesis." See here. https://gwern.net/doc/longevity/2002-weinstein.pdf
Oh just a related point; I think there are (at least) two things needed for the evolution of life. First we need new beginnings, life starts out from a single cell. That's so common we don't think about it much. And the other is death.
This is definitely one of the top things I'd want out of ASI and I think it would be possible to a fair degree. Although I also think that if it ever becomes possible to look basically however you want, there will be much less value attached to looking good, so I might get bored of being beautiful and decide to revert to my naturally goblinoid form.
Yeah, physical perfection is going to be like owning a car, nothing special.
It's going to be sooner and way more common in virtual worlds before physical.
Reshaping ~30 trillion cells is HARD. Even gengineering a defective gene into a better gene in adults is all but impossible right now, because of the formidable difficulty of getting the edit into enough cells in a short enough period of time, and the only things we've attempted are edits in small, localized organs like eyes or livers (https://performativebafflement.substack.com/i/151736795/why-we-cant-do-this-to-adults). This is why most gengineering efforts are put towards embryos, it's substantially easier (you just change one cell, or a handful, and it grows into the 30 trillion adult cells). People are working on massively parallel CRISPR changes and various modalities to change a lot of cells today. AGI could plausibly make this happen maybe 2-5 years after the "country of geniuses in a data center" point, with some good fully automated wet labs (these exist already). But that ends with "you can now fix a handful of genetic defects in adults," not catgirls.
To fully modify human bodies (our glorious post-scarcity full luxury catgirl future) will essentially require brain uploading and downloading into clone bodies, or fully artificial bodies that are close enough to biological. This is well past our predictive threshold, the path from here to there is so stacked with nesting "known impossibly hard problems" and rife with "unknown unknowns" we can't even predict it. So I think you basically just have to hope that the Singularity / ASI is good at biology at that point, and post Singularity all predictive bets are off anyways.
So it doesn't look great. But fully immersive VR systems that actually trick the brain into thinking it's reality enough are probably on the horizon in the same 2-5 years post "data center full of geniuses."
I could point you to some frankly terrifying research from Meta, of all places, on literal mind reading, that is suprisingly advanced (https://performativebafflement.substack.com/i/166900292/dream-technologies).
And both Google and Meta have been on this track for a while, I think we have a decent chance of making some real advancements here.
Yes, I agree with your take re mind uploading. That's going to be the only way available for true phenotypical freedom for a good while, I'd think.
On the digital twin idea, it's a solid idea, but our gears-level understanding of biology is so minimal that we currently aren't capable of digital twinning anything more complex than a nematode right now. Biology is HARD (examples https://performativebafflement.substack.com/i/195726712/birds-and-bats-are-the-real-nietzschian-ubermenchen).
And when it comes to synthetic life, we've only JUST been able to make a fully artificial cell that metabolizes and replicates on it's own, and even it craps out after 3 divisions or so, AND we got there by the simple expedient of "keep taking genes out of a known working organism until we get the minimal working set," and we still don't understand what a third of the remaining genes do.
And even the digital nematode, the Openworm project, has basically gone nowhere interesting.
So although the digital twin idea is solid, hopefully the country of geniuses or ASI will be considerably better at biology than we are, because we're not doing great.
On ASI predictions, I would defer to our host and the AI2027 team, which currently I think is estimating singularity in early to mid 2030's, with some holding out as long as 2050. So maybe 2030 - 2050.
"AI 2040 and Machines of Loving Grace forecast that AI will cure cancer, end disease and stop aging"
And you'll get pie in the sky when you die, is my rejoinder to that. We've come one hell of a long way in medical science, so that people are regularly living into their eighties (and now comes the complaining about Boomers sucking the life out of their Millennial/Zoomer descendants to pay for their inflated pensions and luxury retirement lifestyles for decades, instead of discreetly dying once they hit seventy, five years after retiring from work, as they should have done).
But "you'll never grow old! all disease will be conquered! cancer no more!" is more of the Golden Age SF dreams. The more we know about biology, the more we know how goddamned complicated it all is.
Remember the heady days of the Human Genome Project, when the promise was "individual personalised medical treatment for *you* plus we'll cure all disease and cancer no more"? Then it became "Okay, we've mapped the genome, that's fine but now we need to map in more detail and we've moved on to different elements":
https://sangerinstitute.blog/2025/06/26/how-our-beginning-is-shaping-our-future-25-years-on-from-the-human-genome-project/
Yes, personalised medicine, but not for everyone. Yes, advances, but disease still remains. So I'd be wary of any promises or forecasts that "yes, you too can cosplay Good Lady Ducayne".
Also, I think "completely changing your body" will only work for the very young, probably via genetic screening and embryonic selection, not for the already alive adults who are older and whose dysmorphic features are already fixed. No plastic surgery 20 year old perfect new ageless disease-free bodies for those of us who may not even be alive in 2040 due to our ages!
I'm a total babe, I don't want the rest of you gargoyles being able to just outshine me with the touch of a button. No way. Gonna fight this
Maybe something like this ? I do not own this product nor do I endorse it, just using it as an example:
https://www.amazon.com/UGREEN-Enclosure-Tool-Free-Thunderbolt-Compatible/dp/B09T97Z7DM
The only problem with this thing is that it is an enclosure and does not include the M2 memory that you need, so the pricing is a little deceptive.
True, I didn't mean to imply that this is the complete solution (sorry about the confusion); that said though, IMO any M.2 drive should work for the stated purposes.
Then perhaps something like this ?
https://us.transcend-info.com/product/portable-ssd/esd310
I have never tried this specific product, so once again I can't recommend it, but people seem to like it. That said though, I'd be cautious. I've tried similar thumb drives before, and their write performance usually tanks very quickly -- basically as soon as you fill up their internal cache. That would not happen with an M.2 drive.
Have you done a Google search for such a thing? I know they're out there.
https://a.co/d/02cjuwAq
There is this, but it ain't cheap.
> Can anyone recommend a USB stick that actually ... works?
I did some research before I bought a bunch of USB sticks 2 years ago. Broadly, most non-brand USB sticks are made of lies and mediocrity, but some people third party test a bunch of usb transfer speeds and cycles before failure, and you want the ones that do well on those.
I don't know how often USB performance changes, but if it's yearly, you may want to look for yourself for some of those third party tests, or ask your AI. If 2 years ago suffices, the ones I ended up getting were these:
PNY Duo Link
Samsung Fit Plus MUF-256 and MUF-128
And a bunch of the small blue Samsung Type C MUF-128DA/AM that are waterproof
I've used them heavily since then, and have verified transfer speeds on all of them at the terminal. They don't always get the full 200mb/s or 400mb/s they claim, but the speeds have been good enough to transfer them completely full of files both up and down in a reasonable amount of time multiple times.
Hope that helps!
Does anyone know whether there's a plan for Alpha schools to expand internationally?
Any thoughts on Andy Burnham and how well he's likely to do as Prime Minister? Personally I am trying to maintain a sliver of optimism against the pessimistic takes I'm seeing but would be curious what people here think.
He's very enthusiastic about devolution and giving more power to cities and regions, which is great. Other than that, he strikes me as a people-pleaser and someone who is not used to thinking deeply about policy (other than his pet project) or weighting tradeoffs. Also, he seems to think that UK has stagnated (or entered decline) due to too much neoliberalism, so I expect the stagnation to continue. At least he's a decent communicator, so he has decent chances of winning the next election.
I have never understood what problem devolution is trying to solve.
Inefficiency, incompetence and corruption.
When things are centralised, people say 'this inefficiency and incompetence are due to distant bureaucrats, who don't understand local communities, making decisions from hundreds of miles away, subject to impossibly Byzantine regulations, and handing out lucrative contracts as a result of shady lobbying campaigns. We need devolution!'
When things are devolved, people say 'this inefficiency and incompetence are due to clueless provincial nobodies who have no experience of running serious projects, using outdated and hopelessly inconsistent procedures, and handing out lucrative contracts to their golfing buddies. We need centralisation!'
Oh I like this take.
If the centralized body is already corrupt, inefficient, etc - then giving a chance to devolution makes sense. Maybe at least some places would make it.
Absolutely true, the oscillation is necessary.
This is roughly my take also.
I will add that you also seem to get higher variance in quality with local rule - one intelligent and motivated person can make a big positive difference, locally - but by the same token a single dumb and motivated person has a much higher negative impact locally. If your local community is smarter and more involved than average for your (region/state/county), local rule is better than distant rule: but people are poor judges of if their communities are better than average.
That single dumb motivated person can make a bigger negative difference locally, but a much smaller negative difference than if power is centralized and that person is kicked upstairs. Moreover, when a town screwup is screwing up, you have a much greater chance of stopping him than if he's screwing up his county, or in turn, the entire kingdom.
Every little town is Lake Wobegon, right?
like.
I was about to make that point, and you darn well beat me to it.
Have you by chance ever read Sunshine Sketches of a Little Town?
Which sounds like it's just pointless dithering by idiots, but actually, rot *can* accumulate in a system over time, and cyclically reallocating power back and forth between disconnected bureaucracies can serve as useful spring cleaning in some instances.
In engineering terms, the central power is a Single Point of Failure.
Most human organizations are mediocre at best. If your one central government is bad, nothing good will happen until the next election.
TLDR: Devolution moves policymaking toward a version with the boons of the market; more information channels are available, bad ideas are outcompeted, and policymaking is "democratized" (more policymakers at different levels of policy means more ways for your average joe to access policymakers and influence policy).
Primarily:
(1) irresponsiveness to constituent demands at a grassroots level.
(2) Consider a "centrally planned" versus market economy. The first has a few key decision makers who aren't necessarily aware of what is on the ground, nor are they pricing in all relevant information (only that information that comes to them). This makes them liable to mispricing information, having little accountability (decision arbitrage) due to limited policymaking competition and so forth (think of all the things bad with monopoly).
In a market economy, each individual competes with each other individual, with a far more heterogenous array of information.
To make the analogy more direct: with more individuals in policymaking (since devolution has empirically resulted in multiple parliaments instead of one, for example), there are more access points to policymakers, closer to normal people, meaning more ideas, particularly those relevant to the average person, are given a chance.
Individuals in small towns are more likely to be able to successfully lobby their policymaker for certain changes, whereas if there are fewer policymakers, there is a greater concentration of lobby money per policymaker, so business wins out.
Some of these ideas will be clearly bad; some will be good. The bad ones are outcompeted more quickly than in a more centralized decision-making model - since the flow of ideas is greater.
Policy hubs can then compete on policy and the best ideas, through process of repeat experimentation, can be transferred between areas. Policy iteration speeds up.
If this is ringing any bells, this is state federalism in the US.
Can you say more (or link to further reading) about Scotland falling behind since devolution? My impression was that there was rough parity but I don't have strong evidence for that.
Health and education outcomes have gone from better than England to significantly worse since powers over those issues have been devolved. Clearest in things like PISA rankings.
I don't know about that; Scotland has gone from the fifth strongest region to the third strongest (by GDP per capita), it seems to be doing decently well. It started out around as good as the West Midlands; it is now decidedly better.
Now some less devolved areas have done better -- e.g. Manchester has knocked the socks off even London over that period -- but Manchester had some devolution from where it was with the introduction of the Mayor.
Overall, devolution wasn't the silver bullet it was hoped for, but it hasn't ruined Scotland either (my friend who lived there says it's a much worse place to live, but that's not showing up in the metrics).
I understand the general argument, but what major issues that the UK is facing will this help with?
The devil's in the detail, but I can provide a high-level point for the North-South divide.
London and the South-East subsidizes the rest of the country. In some regions (the North East), the government is 50% of the GDP (admittedly, this piles in issues of (1) ageing populations (NHS spending being high), (2) regional industry decline, neither of which devolution will solve completely). There is substantially little economic activity occurring in these regions.
Increased devolution, particularly cleaved on an economic basis (i.e. around Cities) gives greater autonomy to these areas to (1) attract talent (e.g. through subsidizing training if they move to this area), and (2) compete with London for companies' offices.
It would then reduce the ageing population burden in the North East and spur on regional industry development.
Contrast this to the (centrally planned) approach from Whitehall to achieve the same effect: Special Economic Zones. Not to sound too cynical, but all this really resulted in was part of BBC's offices moving to Salford Quays in Manchester.
Change won't occur through piecemeal projects like that especially when they are individual pet projects for politicians on 5-year terms (not to sound too populist). Mercenary incentives are ironically longer-lasting.
Further, the answer isn't throwing more taxpayer money at these areas to 50% of regional GDP and beyond; it's increasing their economic independence, generating tax revenue which can then be spent on ameliorating regional issues, and so forth.
Regarding non-metropolitan areas: regional multiplier effects (e.g. for commuters - see the Home Counties' relative wealth) would affect the whole region.
Link:
https://ifs.org.uk/articles/exploring-regional-differences-public-spending-across-england
That was interesting.
Devolution of taxes allow areas to raise their own money for projects, and spend it locally.
For example, Leeds has been trying to build a tram system for ~20 years (Leeds Supertram) but the funding keeps getting cancelled. In a more devolved system, they could raise their own funds (like the West Yorkshire Strategic Infrastructure Tariff) for this.
Local decisions should be made locally, by people who care because they have to live with the consequences,
Why?
For similar reasons that CEOs micromanaging developers is bad: central planners lack the detailed local/domain knowledge to make appropriate decisions even when they have knowledge of problems' existence or importance in the first place, response / round trip time is poor, there is too much going on overall for them to give enough focus to specific local areas, and they only very indirectly suffer the effects of their bad decision-making so lack incentive to do better.
Problems such as this: https://www.birminghamdispatch.co.uk/the-forgotten-post-war-decree-that-deliberately-strangled-birmingham/
"For more than three decades — from the introduction of the Distribution of Industry Act in 1945 to the beginning of the 1980s — it was impossible to set up a new factory in Birmingham, or to expand one, without an “Industrial Development Certificate” (IDC) from the government. "
I think devolution is a response to the rise of the far right. Even when establishment politicians are effective - such as by increasing wealth or lowering crime - their popularity seems to decline. The conclusion is that the voter feels locked out of the process, sometimes even voting for blatant and unapologetic wreckers. The idea of devolution is to transfer power to smaller and smaller authorities to give the individual voter more (apparent?) control. I don't know if it will work or not.
Voters who angry about immigration aren't gone to be appeased by local authorities getting greater powers over bus routes.
It's worth remembering that in the English context, devolution was a two way move. Some transfer of powers and funding from Whitehall to the regions. But also a pooling of sovereignty so that decisions which had been taken at the level of the individual local authority (in GM, Manchester, Bolton, Bury, Salford etc) would now be taken by an elected mayor. The "gainshare" deals given to the mayors (thirty year funding commitments from central Government were, officially, a recognition that this pooling of sovereignty would lead to better economic outcomes, thus more receipts for Treasury; and that some of this should be allocated locally. Not very scientific, but I have seen sillier things. And I think it's definitely sensible for functions like transport planning to be done across a city region rather than separately in five or ten or whatever separate authorities.
The trouble I think is that for political reasons the new combined authorities have been drawn too small. GM, West Mids are probably ok; S Yks, Liverppol perhaps suboptimal; Teesside and Hull really too small. It's understandable how we got to that point, and especially in the Tory party there is still folk memory of Ted Heath's rather Bonapartist rationalisation of the early 1970s which gave us things like Avon and Humberside, and which as it happens was largely right. But the need to establish consensus, large consultation exercises, affirmative order debates on the legislation took us in a suboptimal direction.
By and large the English provincial cities did best when there was strong local administration (both the nineteenth century and 1930s town hall and city hall in Leeds are among my favourite English buildings, and the confidence to produce those was quite a thing).
I wonder how much the centralisation of power was a reaction to far left subversion of local politics. Think Derek Hatton.
Leaving aside the characterisation of case histories as "far left" or "far right" and their stigmatisation as "subversive", the inevitable conclusion of Hatton-like events must surely be that all central governments will react badly to local government deviations from the party line if in their view it becomes excessive and shows signs of wanting to challenge the status-quo.
In part it was trying to solve a problem which, following Brexit, no longer exists. While the UK was in the EU, the Blair government in the early 2000s was keen to divide the UK (including England) into ten or so smaller self-governing regions, like Swiss cantons, for easier full assimilation into the EU and discarding of our national identity. Each region would have been little larger than a typical US county. But a referendum conducted in the first such proposed English region firmly and quite rightly told the government where they could stuff the idea and it was abandoned.
For Scotland, Wales, and Northern Ireland, another of the UK government's motives to allow devolution was to ease the pressure for full independence. Arguably this would have encouraged ever more concessions until independence would be almost achieved anyway. But that doesn't seem to have happened, probably largely because ..
The desire for independence is also suppressed by something called the Barnett formula, whereby more is spent in these so called "devolved" regions than their tax receipts. That is probably mostly why the referendum conducted a few years ago by the Scottish Nationalist Party (SNP) failed to gain its hoped for majority for independence, and polling suggests another referendum would currently also fail.
>the Blair government in the early 2000s was keen to divide the UK (including England) into ten or so smaller self-governing regions, like Swiss cantons, for easier full assimilation into the EU
That's a fact, not just your interpretation?
An AI writes
>: The Blair government initially planned for English regional assemblies that would coordinate policy and connect with the EU's regional funding mechanisms. Following the 2004 "No" vote in the North East, the government shelved regional devolution in England.
So.it's about regional.development, the nice thing where the EU gives you money, not the nasty thing where they tell you what to so.
The cynical answer is that it solves the problem of Duverger's law that it is hard to start a national party in a first-past-the-post system, whereas it is easy to start a regional party.
Here's an anecdote about a problem that it could solve, although it's not clear that anyone advocating devolution is arguing that this is the point. Local governments spend a lot of time lobbying central governments for money. They don't want to do any particular project because maybe they could get the central government to subsidize it more.
https://backofmind.substack.com/p/waiting-for-the-train-that-never
People from different places have different values, and therefore want different things from their government. Devolution means that laws and government actions more closely match what more people want.
At the level of Scotland, Wales and Northern Ireland, they all have specific local problems and (except Wales until recently) their own legal frameworks that overwhelmingly-English Westminster politicians don't understand and don't prioritize. The demographics of the UK mean that if the government is completely centralized, as it was until the late 90s, you don't need to be some kind of public choice galaxy brain to see that a government can survive based only English support.
Within England I'm not convinced it does make sense. You can argue the same logic that a government can surivive based only on support from Greater London, so the English regions need their own lawmaking bodies. But that's not as clear. Governments always need some support from outside London. At one time that could be London+Scotland, which was the foundation of the last Labour government, but that no longer works, so the English regions have some power. And the question of how to split them up has always bedevilled the whole project. There are regions that make sense historically and still have relevance, like Cornwall and Northumbria but they're quite small. The majority of the English population outside of London lives in the cluster of cities that spans from Liverpool to Leeds and Birmingham, but there are 4-5 cultural and historical units within that that make sense but again they're very small. The tension over English devolution is always between big units that make administrative sense, but which people don't really care about, and units people might actually feel some loyalty to but which are too small to be administratively efficient.
Devolution works if cities get both decision and fiscal power, i.e. ability to collect taxes locally. This will align incentives but also increase regional inequality, at least in the short term. This is a main reason why I do not put higher than 25% on fiscal devolution happening.
In summer 2025 UK abandoned plans to introduce zonal electricity pricing because it could lead to "potential price inequality and postcode lottery"[0]. As a result NESO (and ultimately consumer) is occasionally at the same time paying wind farms in the north to not produce electricity AND paying gas plants in south above strike price to produce it.
Unfortunately same "postcode lottery" reasoning applies to fiscal devolution
[0] https://www.utilitybidder.co.uk/energy-and-utility-guides/uk-zonal-pricing-abandoned-what-it-means-for-businesses/
My understanding is that "devolution", in this context, means "Britons can have a little federalism, as a treat." Is that basically right?
>he seems to think that UK has stagnated (or entered decline) due to too much neoliberalism, so I expect the stagnation to continue
I mean, after two years of Starmer, this assumption should be self-evidently correct.
I would admittedly be more bullish on Labor if it was still led by Corbyn or someone with similar genuine ideological convictions, but it's not inconceivable that a mere pragmatist responding to evidence will be enough. (And surely communication will NOT be enough, at some point the government needs to start delivering if it wants to win another election.)
I imagine he'll be the same as Starmer: not a bad guy, but utterly lacking in vision, and without a plan to fix the UK. Also, a people-pleaser, will find it hard making difficult decisions. So will almost definitely be ineffectual.
There has almost certainly never been a worse time to be British Prime Minister than summer 2026. The country is in a very poor shape financially and is still suffering from the fallout of Brexit, that disastrous self-inflicted wound; the special relationship with the US has proved the flimsiest of broken reeds; the high likelihood of direct or even indirect conflict with Russia and the low (but still conceivable) likelihood of conflict with the USA within the context of Nato; the obligation to substantially increase defence spending for the aforementioned reasons; and above all the very significant anxiety and polarisation of the electorate are all factors which will be working against Burnham. So will the native conservatism of the English and the guaranteed reluctance to accept the inevitably radical solutions the new PM will have to resort to.
On the bright side are the relief with which his accession will be greeted, particularly after Britain's incredible string of disastrous leaders over the last decade, and the near-certainty that Farage's latest move to get himself re-elected to his parliamentary seat as a way of demonstrating political strength will result in making him a national laughing stock, thus finally sinking him and Reform for good and all. But the period of grace will be short, and Burnham must make the very best of it. Personally, I am optimistic that by exercising good sense and some considered but radical solutions to the country's problems, he will succeed in turning the old ship around.
If Brexit is the issue, why are France and Germany suffering from the same malaise?
I'm not sure which specific malaise you're referring to, but I think it's very clear that the UK has had a significant hit from Brexit over and above the stuff that affects everybody. Not the source of all our problems, sure, but a very large and self inflicted one. There's a good round up here: https://www.gspublishing.com/content/research/en/reports/2026/06/24/487d9f10-f31c-4e8d-ac71-b248ce8a8aaf.html
Why did they take 2016 as the starting point? I was under the impression that the economic measures of Brexit took effect in 2020-2021.
It doesn't make sense as an analysis, the argument is that Brexit made the UK economy more similar to Europe in terms of growth profile.
Why does this not make sense to you? Inside the EU, the UK is the most free-market country in the EU, which attracts a lot of continent-scale (and wannabe continent-scale) businesses to base themselves in the UK, where they can sell to the EU. "Lots of continent-scale businesses" is also a notable feature of the US.
Now those businesses go elsewhere, with the net effect being that the UK's economy has a lower proportion of such businesses, much more like a merely average European country. The loss of a specialised role in a larger ecosystem results in becoming more like an average non-specialised country.
Good question. Both are reasonable index years and some analyses check both as a sensitivity analysis to make sure it doesn't affect the result too much. In a nutshell, the reason 2016 is a reasonable choice is because a lot of investment decisions were already affected by the referendum result itself
"The UK economy has materially underperformed since the 2016 EU referendum, shifting from a growth profile closer to the US to one closer to the Euro area."
I don't find it convincing that Brexit made us more similar to Europe.
I'm not sure whether you're curious or snarking. I'll assume the former but please let me know if it's the latter!
They're not saying that Brexit "made us more like Europe". They're saying that our growth profile (i.e. how GDP changed over time) was better than European comparators before Brexit, maybe almost as good as that of the US, but after Brexit our growth was much worse than that of the US, as bad as, say, France and Germany's already were.
All countries have a bunch of factors that make them do well or badly, so no one comparison will cover everything. But the "doppelganger" analysis reported by GS (also known as synthetic control) is a statistical way to try and get as good an estimate as possible of the impact of one particular intervention (Brexit) in isolation.
I do find it convincing that The Powers That Be set out to hurt The Queen's Country after Brexit(I assume they find King Charles to be a lot more on their side.)
Seems like whether UK is in EU or not, it can keep blaming its problems on EU forever.
Who are the Powers that Be?
Germany is suffering from the Greens, and France is suffering from... well, being France. Also, for real, legit propaganda designed to cripple France (air conditioning should ring a bell).
This is such a wrong take! It's exactly the opposite. The best time to lead anything is when it's at its worst. The UK has leading universities, stable and functioning institutions, mature legal and finance industries... In short, everything a country needs to thrive.
It only needs to fix a few things holding it back - mainly perverse welfare incentives and planning, and it's set to grow.
It would be relatively easy to fix... unless you're attached to the ideology that screwed it up in the first place, in which case it's impossible.
In terms of perverse welfare incentives, the only concrete policy I've seen from him so far is that he wants to build a whole lot more council flats, which are one of the most perversely terrible forms of welfare I can think of.
my old man's a dustman.....
"The best time to lead anything is when it's at its worst" is something of a free-floating statement, unsupported by any evidence; one could just as easily reverse the meaning to "The worst time to lead is...", and it would make just as much or as little sense.
And, yes, Britain still boasts some viable and important institutions, but they are badly at risk. Education, and particularly those leading universities, have suffered from the determined efforts of successive administrations to squash their independence. Health? We have all seen how NHS, once the world's finest free universal health system, has been slowly reduced to a dispirited and underfunded ghost of itself. Finance? Yes, if you regard playing the role of the word's favourite tax shelter and the promotion of private equity to represent productive financial instruments for the nation.
In each of these cases, ideologically driven principles have been responsible for their change for the worst. If Burnham can reverse even some of them, he will be on to a good start.
Sometimes it's fine to make a statement unsupported by evidence but which is backed by common sense. It's often easier to turn things around than to maintain peak performance.
I don't think Cambridge, Oxford, or Imperial are really at risk of not being world class any time soon. Finance is essential for growth - innovation depends on it. The point I'm making is the UK has institutions that can very easily support explosive growth.
I don't think the issue in the UK is bad ideology. It's mostly lack of one. No PM in the past 20 years has had actual beliefs, except for Truss, who just did a terrible job. Everyone else has gone with the flow, and caved in when confronted with tough choices.
What the UK needs is a leader with guts, who is prepared to cause some pain to some minority groups (pensioners, landlords, welfare recipients), but who can also communicate why changes are essential.
> What the UK needs is a leader with guts, who is prepared to cause some pain to some minority groups (pensioners, landlords, welfare recipients), but who can also communicate why changes are essential.
Agreed. Though I would add the conservative press to the list.
Pain to the press? As in...?
Mature Finance Industry currently under attack. Lloyds of Tehran rings a bell, doesn't it?
No.
https://www.irantimes.com/news/lloyds-and-santander-accused-of-providing-accounts-for-iranian-front-companies
Okay, serious news (looks like the "lloyds of tehran" was a wisecrack, and they didn't use that as a "for serious" name).
What's that got to.so with UK decline?
Burnham.has pledge to retain the pension triple lock, which is probably the worst aspect of the welfare system.
100%. This is what I mean, he'll refuse to upset anybody. He even made some noise in support of waspi.
Britain could adjust its foreign policy ambitions in line with its current economic and geographic position, maintaining a nominal military focused on coastal defense, and accept that empire is a thing of the distant past and that it doesn't need to get in fights with countries on other continents. It could even use the savings from doing so to improve the lives of it's citizens.
It probably won't, because seemingly every politician in the world these days wants to take money out of the hands of those who need it most and instead hand it over to arms manufacturers, but it *could*. That's a thing a world leader could do.
Ever since Suez inflicted a reality check on Britain's imperial ambitions, the Foreign and War Offices in London have had trouble adjusting to the world as it is; mercifully, Labour governments kept the UK out of both Vietnam and Iran, though it took a third pseudo-Labour government under Blair to drag Britain into the Iraq and Afghanistan fiascoes.
The problem Burnham will now have to face won't be imperial fantasies, but the hard fact that Europe is facing military threats very much closer to home, from the Baltics to Greenland.
Britain still has overseas territories that need to be defended. Think Gibraltar and the Falklands.
I very sincerely hope that the UK will never again repeat the political and moral folly that led Margaret Thatcher to fling a fleet across the Atlantic to maintain her country's grip on a remote and wortheless fragment of land to which Britain had no claim beyond the remains of a tattered colonial inheritance; though of course the insane expenditure of blood and treasure did ensure her re-election and offered a useful precedent to Bush Jr. As for Gibraltar, that is another faintly ludicrous relic which survives thanks to the good sense of successive Spanish governments; I doubt increasing the Royal Navy's blue water fleet beyond its current single (and singularly ineffective) Type 45 destroyer will do much to further secure the Rock.
The rock is a colonial enterprise, but so is Argentina. Presumably it will be lost someday, but Argentina will come out the bad guys even if they win.
That said Britain should probably negotiate some deal.
Europe needs a navy though, serious powers are military powers.
Small but strategically located islands aren't colonies. They generally.lose money. The "strategicaly located" tells you their real purpose.
Cough. https://en.wikipedia.org/wiki/Falkland_Islands_oil
Surely you're not suggesting that Maggie was dissembling when she claimed that her only concern was to rescue those patriotic sheepherders?
Here we are, forty years after the Iron Lady fought a war over the Falklands, and it looks like not a single drop has yet been recovered. There's a thing where people look at every single war ever, find some oil-exploration report saying there might be oil there, and say "See! Those evil meanies were really fighting that war over *oil*!". This is almost always nonsense.
You have to have overseas territories if you want to extend power worldwide. Otherwise you dont.
Reform (or similar) probably aren’t going away. I once visited the industrial wastelands and decrepit sea side towns of Britain and wondered by they were voting Labour, which wasn’t really representing them anymore. And the Tories were never an option. It’s hard to predict when a new party can arise in a two party system but that kind of vote isn’t going away.
You are absolutely correct that none of the three historical political parties in the UK can really claim to represent the lowest layer of today's proletariat, whose basic grievances could be summarised as social marginalisation, hatred of immigrants and lack of opportunity and decent jobs. A serious attempt to deal with the latter would probably also allay the others - except that AI has now also arrived to further devastate the job market. Admittedly, the devastation is currently largely affecting white-collar jobs, but it's clear that the poison will inexorably seep downwards.
The country is "in a very poor shape financially" due to the shadow banking sector. Why else does King Charles express himself to Trump on a charm offensive?
He doesn't want the UK to get hit by tarrifs.
I am also trying to remain optimistic, but he's coming in at a very difficult time and the expectations in the country are unreasonably high so I think a bit of disappointment is inevitable.
The thing I'm most excited about is he's previously been interested in land value taxes and seems to have thought carefully about them and understand why they work. He also understands the importance of infrastructure in a way, e.g., Cameron didn't. Trouble is, those things take time to pay off and he doesn't have much time.
He's going to be squeezed from all sides, with Labour backbenchers wanting increased welfare spending, markets wanting tough fiscal rules, reform and the Tories banging on about immigration and North sea oil, the country desperately needing a bit of investment, all while the population ages, the planet heats, and wars get fought.
A lot of this long term stuff might become AI-moot, but I've not heard him talk about his AI plan either.
He doesn't have a plan to deal with the major issues facing the country and he seems to be willing to say yes to every department asking for extra spending. It will be interesting to see who he makes Chancellor.
He’s a better communicator than kier Starmer, but I think he’s a flipflopper who won’t take the necessary tough decisions: end the pension triple lock, raise taxes, nationalise all golf courses to build houses on
He seems to have done a fair job in Manchester. His policies seem better matched to Industrial Socialism (ie use government policy to support specific industrial goals) rather than Benefit Socialism (ie use government policy to force the productive to support the workshy), so that is a minor improvement over the current Labour leadership.
Like others have said, my expectations for uk politics are very low just in general, and Burnham doesn't come across as an impressive guy.
But he has made noise about introducing a land tax, which sounds extremely interesting and is not something I'd ever expect to hear from a mainstream political figure. Most likely a land tax is waaaay too radical to ever actually get implemented in any substantive way though, so I doubt it will actually happen, but just on its own that could fix a lot of Britain's problems imo.
It will be interesting to see if he can decrease the extreme dominance of London and the Southeast in the UK.
UK PM is kind of a poisoned chalice at the moment. I think there are people who are reasonably competent and will give it the old college try but they will be -100 in favourability ratings within a few months of taking office whatever they do.
I don't see much chance he will really make much difference. In the short term, he's more charismatic and if he takes good advice on communications he may get Labour through the next election.
Looking at this as an emmigrant, so I know UK politics well but I'm not caught up in the news cycle, there's nothing really wrong with the current government. It's .. fine. Keir Starmer isn't Tony Blair or Margaret Thatcher, but he's also not Liz Truss or Boris Johnson. The party isn't borne aloft by charisma and careful, ideologically coherent policy development, but he's not really made a lot of unforced policy errors and he's only mildly messed up the communications.
There are two basic problems that face the UK - Brexit has been bad for the country economically, and British productivity has always lagged behind other European countries, let alone the US and China. Those two effects plus the accumulating effects of crises where the UK has just kind of scraped by mean there just isn't as much money as there was in 1997 so its harder to come through on big promises.
In addition to that, there is a human capital problem. Keir Starmer is fine, and I'm sure Andy Burnham will be fine, but the last government with anyone with any level of real star power - the ability to both master a brief and communicate the plan - was David Cameron's. Since then its just been a parade of mediocrities and bullshit artists. These are just not serious people. I did a back of the envelope calculation the other day, and of the people I grew up with who got good educations - doctors, lawyers, engineers, etc - a third have emmigrated. I don't think this fact is unrelated.
I suspect it'll be a case of "meet the new boss, same as the old boss"
Good morning everyone.
Here is a description of a character in Bleak House:
"[she] would be handsome, but for a certain feline mouth, and a general tightness around the face, rendering the jaws too eager, the skull too prominent."
I particularly associate it with Victorian novels, but I think it's not completely unknown in more recent novels to have a character's appearance described in this way, where certain aspects of personality are perhaps hinted at in appearance. (In this case, she is about to make fun of a pretty character, and I think this description suggests at jealousy). At the most extreme, you might get a golden age crime novel, saying something like:
'Mr Craftley had cold, pitiless eyes and a haughty nose, down which he looked at everyone with a sneering disdain. His mouth was hard and pale, with a marked tendency to cruelty at the edges.'
I *think* what's going on in these cases is that the author is just telling you what the character is like right from the outset. 'Mr Craftley is a baddie, and he probably wouldn't think twice about killing to advance his aims, but is he actually the murderer in this case? See if you can figure it out.'
Now I'm not great with faces and visual memory in general, but I'm not aphantasic and I don't have the face-blindness thing. So my questions are:
1) Does anyone actually see faces of new people in this way in real life? 'Ooh, this woman would be prettier if she didn't have such a tight face'; 'Woah, that man has a haughty nose, best stay away!' 'What an interesting feline mouth you have.' etc
2) When people read these descriptions, do they actually form a visual impression? Or is it just (I think like me): 'Oh ok, this hasn't helped me to understand what the woman actually looks like at all, but she has a slightly odd face and is not that attractive, so maybe she's jealous, got it.'
1) Sometimes, if I'm interested in the person. Maybe not right away after I meet them, but after I get to know them, I might notice: "he has really unpleasant fish eyes that always look half-closed, he feels untrustworthy." Sometimes I'll be a little torn: "something about his face makes me think he's untrustworthy. Is it just me being shallow and judgmental, assigning significance to something meaningless, or it it the Voice of Intuition telling me something valuable?"
2) Usually? It depends on the description. I can visualize Mr Craftley pretty well (no idea if my imaginary Mr Craftley looks anything like what the author wanted to convey, but it's a clear enough visual), on the other hand I don't really get a good image of the proud owner of "feline mouth" and "too prominent skull". (What the hell do those things even mean?) Actually, I do, but in my head she is butt-ugly and not remotely close to handsome, so it can't be right.
"Skull too prominent" is a bony face, so the skin is tight/there is little flesh on the bones. You need a certain amount of facial fat for attractiveness, and depending on the era, that is a greater or lesser amount. Some periods prefer rounded cheeks, others like prominent cheekbones.
"Tightness around the face" is also describing expression, though in a subtle way; you get the impression this woman is eager, maybe *too* eager, to say things (generally nasty things) about others. She's impatient or impulsive, looking out for her own advantage. Take that with the "feline mouth" and you get the broad impression of someone who's predatory in some way.
That was a very good analysis.
I don't think so. She has a tightened predatory face, but she could even be kind of hot I think if you didn't mind that.
I don't often see people like this in real life, but I /do/.often see it in films and plays - actors.and makeup departments can convey personality through face, convincingly but not honestly.
1: Bit of both. I'm not going to say someone has a "feline mouth" (which I assume means "tiny and frowny" here, instead of "sexy catgirl" like anime has instilled in me), but "that guy looks like he does drugs", "that woman's got a pig nose", "that guy's got the eyes of someone with dementia", "that guy looks like he holds a steady job and pays his bills".
2, No, they look however I want them to. The only value in visual descriptions is being able to distinctively refer to the character later without using their name. A description of a "feline mouth" means we can switch viewpoints to someone who doesn't know [her], but the audience still immediately recognizes her.
I have found things like mouth position and the lines around someone's eyes to actually yield information about character. An older person who has smiled a lot tends to develop different lines than someone who is mostly stern, for example. I also harbor some suspicion that if someone looks a certain way, it colors how others treat them, and they could end up playing the role eventually if they don't have a strong sense of self to combat it
Authors of the time probably had some unexamined background phrenology beliefs, and this kind of description is faster than breaking appearance and character apart.
I read this as a description of an ugly character. “She would be pretty, but for…” means the same as “She isn’t pretty, because of…” It’s worded to sound like it’s trying to be nice, but that niceness is sarcastic. I think it’s associated with Victorian writing because many writers in the Victorian era make fun of how polite society can actually be vicious.
I don’t think I really notice people with eager jaws or prominent skulls in real life, but I do think I might’ve seen them in video games, digital art, or caricatures/cartoons. I think those are exaggerated descriptions that only apply to exaggerated depictions in various media.
With that said, the impression I get of the feline mouth and eager jaw is that the character habitually holds their lower jaw slightly forward, giving a flat, round lower face like a cat’s. The implication is that the character is always tense rather than relaxed, and looking for the best response in every situation. The tightness of face leading to prominence of skull suggests she might frown or raise her eyebrows a lot, suggesting she is pained when things don’t go her way or people judge her harshly. The implication is that she is exactly the sort of character who would be upset by the narrator’s description, which might be a particularly Victorian style of cruelty.
SHAKE ME UP JUDY!!
I don't personally think any of these kinds of things when I meet someone new, but when I read them in stories, it does reshape the mental picture I had of the character. I think the author just does it to let you know a bit about the character from the get go like you mentioned, as well as tell you some about their appearance as well. To me, it seems like a good way to introduce a character when they need to be remembered, but you don't have room in your story to just be like "oh this is Mr. Pickledurphy" and then jump right into a lot of dialogue or something that shows the type of person they are.
I do not do this irl and I hate it when authors overdo it in fiction.
The stereotypical example I never want to read again is describing characters as having "intelligent eyes".
There is no such thing as intelligent eyes, it's just a lazy way of the author to tell me this person is supposed to be impressive.
Usually these people then do not act any more intelligent than anyone else in the story and not rarely they're bloody idiots.
There's some forms I believe are defensible. Somebody mentioned laughter lines already. It is also certainly possible to get am impression of somebody on first contact and perhaps intuit something about their character in ways that are hard to convey in words.
It is especially badly conveyed by "they have <insert trait> <insert part of face>" and the like, however.
It might not help that I do not imagine character's appearance while reading at all and, if actually reading and not listening to an audio book, will liberally skim physical descriptions.
>There is no such thing as intelligent eyes
I would say "intelligent eyes" means they have the facial appearance of someone intending to engage the topic as an equal. Focused eyeline, not fidgeting, not scowling or awestruck or other expressions of having a pre-formed opinion.
Sounds to me like what you're describing is the behavior of somebody who presents as an attentive and neutral listener. I'd argue "has intelligent eyes" is a poor way to describe this.
For one I can be bored out of my mind in one conversation and paying sharp attention in another, but my intelligence surely remains unchanged.
For another, these kind of descriptions are very common when characters first meet, often without them having spoken a single word. Any narrator who believes they can deduce the intelligence of someone they're meeting for the first time based on their physical appearance and / or facial expression is... let's call them deeply mistaken.
To contrast, I have no issue with one character calling another intelligent once they've actually had the opportunity to judge that. But I still feel it would be helpful to separate this from their physical appearance, as best I can tell these are in fact unrelated.
What would be a better two-word description of it? "Attentive eyes" gets you a bellboy.
I would argue you're using the wrong definition of "intelligent" here, viewing it as an inherent trait instead of dynamic. If you didn't hear half the conversation, your ability to intelligently contribute to it will drop.
Personally I would just drop it without replacement. There need not be a common term to describe what does not exist.
Ideally the intelligence shines through in the writing.
And I agree that the definition of intelligence I'm using is the reason I judge it so harshly, I believe it is accurate, though. Because my impression is that the purpose of the whole thing is to tell the reader how desirable the person in question is.
At least the context I encounter this pharse most often is some kind of romance novel in which the prospective love interest is first introduced and must be hyped up to the reader. The second most common way I see it is of the main character meets side characters and we're to understand which of these we're supposed to like and respect and which to dislike or disrespect.
And that only works if it's about the "inherent trait of the person" kind of intelligence, not the "currently saying something clever" kind of intelligence.
I'm not going to try to defend romance novels, but "we can only like someone if they're smart" is on your end. "Pays attention to you/dismisses you" maps exactly to "you should like them/you should dislike them".
I think it was largely because Dickens published many of his novels, including Bleak House, as serials in a popular monthly magazine. So he probably thought many of his readers, with limited education and literary sophistication, might want insight into a character from the outset and find it uncongenial to have to judge the characters more gradually from subtle hints and the unfolding plot, especially as the full work would have to be read over several months.
For the same reason presumably, he also gave many characters comical (by his standards!) twee names which were also meant to indicate how the reader was to consider them. Examples include Mr Bumble from Oliver Twist, whose name suggests a bumbling idiot, and in Nicholas Nickelby the harsh schoolmaster called Whackford Squeers, clearly a liberal dispenser of the birch! I imagine "Squeers" is supposed to suggest a combination of "squint" and "queer" (odd), indicating a bitter and twisted view of the world and his charges.
Out of interest, I decided to check my post in one of those online AI text detectors. It decided it was 78% written by AI, whereas the correct answer is 0% . Make of that what you will!
Physiognomy was trendy back then:
https://en.wikipedia.org/wiki/Physiognomy#Period_of_popularity
"A host of 19th-century English authors were influenced by the idea, notably evident in the detailed physiognomic descriptions of characters in the novels of Charles Dickens, Thomas Hardy, and Charlotte Brontë. Descriptions over one's appearance in the written form was a way for one to gauge the character's intelligence, ability, morals, and social status."
>When people read these descriptions, do they actually form a visual impression?
yes, particularly when it reminds me of a face I've already seen.
my missus n I love to talk about a good face.
The Victorians were pretty obsessed by reading character from faces. It's not really either-or.
2) I dont know what a feline mouth is, but other than that yeah. Maybe Im misunderstanding some british subtleties, but she sounds like Id be into her.
1) I usually dont verbally think it, but if I were to describe what I think, its in the general vicinity.
Is there a way to contribute to AI alignment or do some other net-positive Data Science-related work while not being located in first world country? Some kind of Data Engineering or Analysis remote postion?
Second this
I'm not an expert, but here are some random ideas from a random bystander:
1. Language translation research. Presumably, chatbots and other AIs do not work as well in the local vernacular as they do in English. Research why not, and also uncover and curate what should be placed into the training corpus for the next generation of LLM models.
2. Cultural idioms. As above, but with a focus on the failings of metropolitan generative AIs in producing objects of local musical, artistic, and poetic traditions. A particular focus on voice and video is probably called for.
3. Educational artifacts. Use AI to assist you in translating material from the global canon into the local language. I would focus on applied math, engineering and science, and philosophy, but your taste may differ.
4. Educational apps. Use AI to create educational interactive apps to teach English and AI theory in the local vernacular.
5. Engagement hooks. Use AI to generate short-form videos jingles, and memes to get young people interested in engaging with this technology.
In other words, notice the areas in which you are starting with a leg up on the Silicon Valley bros, and run in the race where you have a head start.
I want this plan A thing to happen. It seems to me that non believers, those who only imagine the mundane and boring scenarios where AI merely takes all our jobs and hostile actors attack us with drone swarms and kill bots and barage us with propaganda, should still be on board with most of the early part of the plan yes? I don't care how hot the AI topic is in 2028, we simply aren't going to woo normies into the cause talking about galaxy level shit and 2 million dollar stipends. Alternative narratives should be folded in somehow, I believe we can find compatibility. Shit even I have a hard time believing I'm not crazy to take this ASI stuff seriously, I have a deep emotional need to cover for 2017 normality despite the fact that it has already evaporated entirely.
I think Plan A can't have it both ways. If they want a slowdown and control over AI alignment, then they have to frighten the horses ("AI will take all our jobs and take over the world and kill us all!!!!") but then if they want the good things ("AI will take all our jobs and take over the world and be so super-duper-hyper productive the global economy will balloon to the size of Jupiter and we'll be rich off the scraps the AI oligarchs throw us!!!!") then they can't scare the horses *too* much, else the call will be for a total AI shutdown and they don't want that.
Never mind that way too much money has already been sunk into AI so it better start getting profitable sharpish, and the companies and lenders and investors to those companies are not going to tamely go along with a total shutdown.
I have no idea how they thread that needle (present AI as super-scary but not so super-scary it will be dismantled completely).
AI is already profitable. The massive costs are speculative investments that might or might not pay off, but they aren’t operating costs. If the investments don’t pay off, a lot of investors will lose money, but existing generations of LLMs will stick around. (OpenAI’s revenue might have been ~$26B vs operating expenses of ~$1.5B, though I haven’t checked those numbers.)
I am not yet a bubble-narrative-monger, but it is worth pointing out that a large fraction of that $26B revenue comes from R&D spending by enterprises that have also bought into the hype.
That’s fair, but I think a lot of the operating costs would be proportional to scale. There would be some efficiencies of scale, but I doubt those are suppressing prices by a factor of 15.
I'm not sure about this, last I checked I was using (read: wasting through repeated calls for weird layers in order to appear busy) somewhere north or 2-3k $$ of compute in a month for the price of 150-200$ before the price hike, and it still is wildly under priced IMO.
Just doing my part to make them as unprofitable as possible in my goal to suffering max certain parts of society.
In any case; the models that can do anything useful have negative profit on every call that does anything of value. Maybe this has changed recently, but I doubt it.
How did you come up with your $2-3K number?
Based on reported income of companies, the leaked token rates from probably uber, and reporting from the FT, and having instructions form on high to use ai constantly reversed one the subsidy went down a bit.
> How did you come up with your $2-3K number?
You can also back into this via raw hardware and electricity costs. Broadly, it costs ~$1 per million tokens using H200's, and even cheap models cost $2-$5 per million, and good models cost $10/$50 in/out for Anthropic’s Fable or $5/$30 for OpenAI’s Sol.
We've also known at the aggregate level that AI companies had ~40-50% margins on inference in 2025, and 70-80% margins on inference in 2026.
Quoting Sean Goedecke below:
“A Nvidia A100 consumes 400W of power under full load. In practice, even a carefully-tuned inference server will not be at full load all the time, but it’s at least an upper bound. Suppose you’re running a dense 70B model1, which will fit comfortably (unquantized) on four A100s at around 2M tokens per hour. At industrial power prices, that’s about 13c/hr in the USA. Suppose (pessimistically) cooling is the same cost. That’s about 13 cents per million output tokens2.
Let’s amortize the cost of the GPUs, since that’s going to be the most expensive part. An A100 costs about $20k. If each A100 lasts around five years3, you’ll have to make 16k/yr in profit to recoup your capital investment (or $1.80 per hour). At lower utilization, it’ll take longer to recoup, but your GPUs will also last longer. Either way, your overall inference costs are at about one dollar per million tokens.”
Yes the problem is that the mainline AI safetyists both want to slow AI down *and* want ASI to someday exist, something something utils something something lightcone. So they end up torn in just this way.
Eventually opposition to AI will be taken up by people who actually oppose AI.
Which would be good! The primary issue with most rationalists is that they want ASI to exist at all. I'm ok dying at 85 from cardiovascular disease. It's fine. I'm not ok rolling the dice on dying at 30 or 35 because some idiot wanted to create God in a bottle.
There's a rationalist motte and bailey game that the more pro-AI among them obviously know they're playing, "we're going to cure cancer and end hunger, if we can do that safely without killing everyone how can you possibly oppose that?" When in reality that's only incidental at most to what they're trying to do by building ASI.
It is an unfortunate historical fact that being concerned about AI + having money + being organized is a combination that first arose within a community that had particularly weird ethics. Very good epistemics! Great exploration of decision theory! But weird ethics, weird in the "most people don't think like this" way. And in the other cases, hey most people probably *should* think more like rationalists, when it comes to evaluating propositions and making decisions about what's optimal for their objectives.
But even the ideal best case scenario of "God in a bottle" is really strange crap that living humans like you and me do not want, weird crazy stuff that can only be justified by gesturing at trillions of years of hypothetical minds floating through space and assigning some value to that whereas few humans on Earth would give it ANY value. Heck, even "curing" cancer wouldn't be some normal thing like taking a pill that makes cancer stop, cancer is a biological tradeoff against degenerative disease, anything that would actually stop cancer from happening probably throws the entire system off and winds up looking like some wild sci-fi transhumanist genetic engineering experiment.
I think you're right about having to thread the needle, but I also think we're so far away from an AI pause that there's no point pulling back on scaring the horses (yet). Maybe this is something we should worry about if a pause bill is being discussed, but for now I think we can count on the pro-AI / pro-free market people to balance the scales.
Also, their concern about Plan S wasn't that the pause would be a bad idea, it was that the treaty governing the pause would eventually fall apart (as every treaty does) and return us to the status quo. But they also seem to think that there are really good variants of Plan S that involve building better Plan As during that pause period. I think it is conceptually possible to do a really long temporary pause and come back to Plan A later, even if the political will would be pretty difficult to manage.
One subject on which I'd love to read a "more than you wanted to know" is the subject of early agriculture and human height.
One of the weirdest contrarian takes that has recently become accepted wisdom is "agriculture was bad, actually". This, if true, would be surprising -- we know that agriculture rapidly spread out from the points where it was developed, we know that hunter-gatherers either converted to agriculture or were displaced by agriculturalists wherever they met. There's obvious intuitive advantages to agriculture too, in terms of providing a larger and more reliable food supply. So all the intuition would be on the side of agriculture over hunter-gathering.
The evidence against agriculture, as I understand it, is a bunch of archaeological studies which suggest that agriculturalists in at least some places were shorter than the hunter-gatherers. This has been interpreted as suggesting that agriculturalists were actually malnourished compared to hunter-gatherers. But reading the abstracts of papers like this https://www.pnas.org/doi/10.1073/pnas.2106743119 I find it less convincing; there's some convolution of genetic and nutrition effects, and the evidence is not equally strong in all parts of the world.
So what's the deal? Is the evidence for agriculture-bad really strong enough to accept this counterintuitive conclusion? Or is it one of those things where a single bad study has been taken too seriously thanks to pop-science popularisation?
Romanticizing earlier simpler times is a recurrent pattern. After all, our species evolved in a relatively simple hunter-gatherer environment, so we're probably deeply adapted to that kind of life: hard but not complicated. Modern life is exactly the other way around, so it grates, and agriculture is the point where we switched to a settled life, with the relative complication of having to plan at least a year ahead.
I don't see any of the complainers taking to the jungle though.
I think it could well be the other way: we have the tall hunter-gatherers because the life is so finely balanced, the unfit died young (*just* enough food to go round so that the best hunters get fed and the unproductive are allowed to starve or die in infancy). Agriculture gave enough of a surplus that the sickly young and the unproductive old could survive long enough to leave short skeletons.
I suspect the best way to see what our ancestral hunter-gatherer lifestyle was like is to look at chimpanzee and gorilla troupes today and see what they do with all that spare time.
Populations expand until food surpluses are used up. But an agricultural civilization will have a more steady level of food along with "consumption smoothing", whereas hunter-gatherers would have bigger swings as game populations crash.
http://neutralhistory.com/the-average-height-of-medieval-knights-peasants-vikings/
Looks like the issue is famines, which for an agricultural tribe meant "limping through" and for a hunter-gatherer tribe either meant "letting people die" (so that one could still have hunters to "run down the antelope" -- a pretty calorically high workload).
The complainers can’t easily take to the jungle, because most of the good land has cities on it. I was in my teens before someone pointed out that I thought of Australian Aborigines as living in the outback, when actually their biggest population centres were on the coast. It’s just that ex-Europeans took over the coasts but found it to hard to use most of the interior, so that’s where Aborigines were able to keep more parts of their cultures.
My understanding is that the north coast of Australia is still relatively unpopulated by Europeans.
Yes, and my understanding is that this is exactly where the highest Aboriginal population density originally was.
It makes sense that the land best suited for hunter-gatherer lifestyles is not the same land that would be most attractive for European-style agriculture. The tropics support a greater density and variety of life than anywhere else.
H-G life may have its simplicity but it also can’t support anything but very small populations.
The global estimates I've heard is around 50 million people.
We're around 8300 million now, so that would be a tough adjustment.
My understanding is that the theory behind "agriculture was bad" is that it led to such large population growth that the average person was worse off.
Note that larger population = more likely to prevail in war, so agriculture gave advantages at the group level, regardless of whether it did so at the individual level. Hence, it is unsurprising that agriculture spread.
I have two problems with this version of the theory though:
1. If the population was roughly stable during hunter-gatherer times, this would presumably imply some kind of cycle between periods of more-than-sufficient food and periods of mass starvation once the population exceeded the carrying capacity of the land.
2. For agriculturalists, the solution to running out of food is to put more land under cultivation. Eventually you run out of land, but that wouldn't have happened in neolithic times, at least not in most places. France, for instance, was still 50% forested in the 10th century AD.
I don't understand the relevance of #1.
As for #2, that will not solve the problem of access to animal protein.
Melvin, during the stone age humanity went from a few scattered groups in Africa to discovering the globe and they did so in small groups with stone tools. When we modern people discovered the continents we found the descendants of those stoneagers waiting for us.
When our World Wide Experiment with agriculture and science ends we will go back to that simpler lifestyle.
> the solution to running out of food is to put more land under cultivation
A lot of running out of food happens on short timescales though. Two consecutive years of crop failures and you're probably eating your seed grain. No additional land would help.
Also, the larger your group is, the further an average person is from uncultivated areas. So of course at some point people will have to notice that they don't have enough food, but getting up from your (your parents') land and moving somewhere to the periphery becomes less certain and riskier proposition so probably less and less attractive.
The argument is basically pure Molloch. You end up optimizing for calories/acre, which gives you monocultures, deficiency diseases, and when things go well, has you bump against Malthusian limits. Though I think some of the theory was also noble savage nonsense about hunter-gatherers that ignores the whole "nasty brutish and short" aspect of that lifestyle.
The agricultural era was nasty, brutish and short. The early industrial era as well. If the hunt was plentiful then the individual was better off in a hunter gatherer society.
Pre-agriculture hunter-gatherers lived at very low population density and had factors holding population in check other than literal starvation (disease, predation, violence, birth control/infanticide, accidents). If a tribe had the good luck to grow too large for its immediate location, they could walk to another places that was slightly worse and mostly uninhabited. Native Americans were vastly outnumbered by buffalo, deer, etc. A world with 8 billion people is extremely crowded by historic standards, and we still have tons of underused land (for example the boring parts of the US East Coast are more forested and more full of deer than they were in the 1800s).
Early agriculturalists did not always have the surplus labor/capital to cultivate more land. Clearing trees and rocks to make a field is a long-term investment that is not certain to pay off, and it's hard to do it before the population grows (because you have not many people and enough food) or after the population grows (because hungry children are bad at hard labor). If the population is struggling to feed itself based on its core land, peripheral expansion may make the problem worse rather than better, and permitting a period of malnutrition until population drops or luck improves may be a better strategy.
On the flip side, agriculture done with good luck and skill has the potential to organize a larger group than the typical hunter-gatherer band, which promotes specialization in creating things that are useful for spreading more agriculture (tools and storage, weapons, defenses, religion, cultural hierarchy, writing). Larger groups of agriculturalists have won all the wars in the long run - but a culture that is good at spreading is not necessarily salubrious for the average participant.
"If a tribe had the good luck to grow too large for its immediate location, they could walk to another places that was slightly worse and mostly uninhabited"
Only on the fringe of human territorial expansion, before that basically closed off, or arguably during the first few generations of human existence. Which is irrelevant trivia in this context.
The *vast* majority of hunter-gatherers in human history or prehistory, could only ever hope to walk to someplace already inhabited by other hunter-gatherers who were well aware that outsiders hunting "their" game would mean their children going hungry, and would thus gladly kill you in your sleep to stop you.
Yes, the Native Americans were outnumbered by the deer, but it takes many deer to feed one Indian (about 20/year if all you're eating is deer). Malthus predates agriculture.
Yeah, I kind of suspect the limit on human population was turf, and that limit was enforced by the neighboring tribe's spears. After all, people who live by hunting are by definition spending their lives getting very good at the use of spears, bows, slings, and the like, which will translate quite well to hunting men instead of deer.
You also get nomadic/seminomadic herders, whose lifestyle prepared them very well for war (think of the story of David and Goliath) and who periodically would sweep over the settled agriculturalists. I don't know what evidence there is or could be for the transition, but it seems pretty obvious that there's a sort of gradual path for hunter gatherer bands to slowly transition into being nomadic herders--you start following and feeding off a herd, find yourself running off anyone else who tries because it's your herd, find yourself killing off predators that like to eat from the herd because it's your herd, and sooner or later, figure out how to drive the herd toward wherever you need them to go, etc.
I think it depends on time / place. A lot of ecosystems have areas which are regionally / seasonally abundant, so your group might spend part of the year here, part there in ranges with substantial overlap. You and all the other groups might also come together a few times a year when the sardine run or prickly pear season is on and there's a local superabundance for everyone to partake of. I think a lot of social stuff probably happened in these sorts of times and places - including things like intermarriage. Presumably a month of full bellies and a bit of socializing away from your extended family was a good way to meet a handsome young man from the neighboring tribe...
"some kind of cycle between periods of more-than-sufficient food and periods of mass starvation"
That, or something like it, appears to be the theory? Hunter-gatherer societies had lower population because fewer children because women couldn't take care of kids while foraging for food, didn't get pregnant as often, and high infant mortality.
(Don't quote me on this, I'm half-remembering something I may have read online).
My suggestion is to try farming with hand tools and no fertilizer for a bit. The loss of height and health becomes intuitive when you look at how much sweat you need to produce a daily meal, and how much of that meal is going to come from one or two staple crops. Compare and contrast to how many things hunter-gatherers eat, and how little 'work' was needed to do it. The tradeoff, of course, is that a hunter-gatherer needs to roam an enormous area (often in a season cycle) while the Paleolithic farmer sits in place.
My favourite take is that the switch to agriculture and cities was partly by way of a baited trap: agriculture provides two things that people enjoy (night life and alcohol), so there is incentive to settle down for a bit and try it out. But, once you make the switch, it's hard to escape because the best foraging lands are now settled and claimed by other farmers and you'd have to move into less productive areas.
Hunter-gatherers don't "work" less than agriculturalists if you include travel time as part of their necessary "work".
Agreed - the travel is a big part of their day job. But then I walk around as a leisure activity, so this doesn't seem so bad compared to, say, pulling a plough.
A lot of people hunt recreationally (I have, but I've never had to depend on it to eat).
You have shoes though.
So did any hunter-gatherers who cared to. Moccasins aren't exactly hard to come by in an economy with a relative abundance of leather goods.
Travel time is always included. Otherwise it would a few minutes work. That time you chucked the spear. And the other time you chucked the spear.
Even modern hunts start when they set out, not on sight of an animal.
Earlier estimates were that they spent few hours per week working, but now "When considering time out of camp as well as in-camp food processing and housekeeping tasks, the average across all well-studied hunter-gatherer societies is 40-45 hours per week, similar to the standard eight-hour working day in industrialised societies" https://aeon.co/essays/what-hunter-gatherers-demonstrate-about-work-and-satisfaction
Well we are not comparing them to industrialised societies with strong labour protection laws that are relatively recent, but agriculturalists and all the other classes that existed in the interim. Like being a servant. Or working in a mine. Those weren’t 8 hours a day jobs.
I’m wary of the argument to one link, a quick ChatGPT discussion assures me that argument is disputed and that the preparation of food was not the job of the hunters of food.
Since you admit to hunting yourself, then presumably it isn’t a chore to you and therefore wasn’t to hunter gatherers. Elites continued to hunt, after all, but had little desire to be agricultural labourers, or to work in that mine.
There’s something in humans that likes the hunt.
Now, clearly the danger of the hunt is largely gone (although a fall from a horse is no picnic), and there’s no starving if you don’t succeed. I still am fairly convinced that in times of abundant prey the hunter gatherer was both happier and better fed than most generations after, until the last few decades.
Re-read my quote. That number is an "average", so cooking & hunting being done by different people doesn't change that.
> There’s something in humans that likes the hunt.
Evolutionary psychology: if people disliked hunting, they would have been replaced by people who liked it.
This is kinda silly. Lots of people hunt, fish, garden, hike, and do woodworking for fun, but this absolutely does not mean that people who have to do those things full time to put food on the table do not consider them work. Indeed, many an old man happily spends his retired days keeping his lawn in pristine shape and doing landscaping projects and gardening to keep it beautiful, and at the same time there are a huge number of people (mostly Spanish speakers whose entry to the US perhaps dispensed with certain legal formalities) doing that same work full time for a living.
I don't have a huge dog in this fight, but it sounds like their 40-45 hours, in including things like "in-camp food processing and housekeeping", isn't the same as our 40 hour workweek. Our 40 hours is time at your job, it doesn't include food processing and housekeeping (not to mention various other things you have to do in our modern society, like paying bills and shit).
Good point, modernity is 40-50 hrs formal work, 5-10 travel, 5-10 food prep, 2-4 cleaning etc. So 50-75 hrs total. Presumably shopping (including travel to and from) should count as well...
Even in the EEA, commute time was an issue!
Sure. but there's all the other stuff outside hunting and gathering. If you're going to chuck spears, you need a spear to chuck. That means making a spear, which means you need to gather the materials to make that spear and then make it, which means spending time learning how to make spears.
And so on for everything else - pots, baskets, clothes, social interactions, status indicators of elaborate hairstyles, and so forth. "Hunter-gatherers spent four hours max getting food and had rest of time to lie around in hammocks, dumb farmers toil 14 hours per day" is not the whole story.
My low-information impression is that one of the common ways we don't understand the past is because we have little intuition for how much work it was to get every basic thing they needed. Growing crops with muscle labor only, digging a well with shovels and ropes and lots of hands, etc. As I understand it, relatively prosperous farmers had like three sets of clothes, because making clothing is really labor intensive. Metal tools were expensive because every one of them was the result of many hours of very high-value labor. And so on. Just heating your house/cooking with wood and getting water from a well and carrying it home in a jar makes for a tremendous amount of brute force labor required every day to keep living.
Women had slutcloths (now called aprons). People had ways to kilt up their clothes, and they washed them less often. Also, "three sets of clothes" generally means "outer clothes" -- the slip, the "undressed" person would wear (see Ben Franklin) wasn't always considered a different set of clothes.
Baths happened considerably less frequently, because they were expensive (particularly if heated) -- jumping in the river was fun, but only when it wasn't winter.
Chainmail existed because "skilled labor" was pretty damn cheap in comparison to "more metal."
Yeah, it's just this weird thing where we all kinda mentally picture living in the middle ages, but we have clean clothes all the time and take a bath every day and such.
Right - half of what we don't understand about the past was resource/capability limitations, and the other half is clever workarounds people came up with to deal with those. Like washstands instead of bathtubs, and aprons+undershirts to keep your one good set of clothes reasonably clean. Some people hear "no bathing, one set of new clothes every year" and think peasants were perpetually filthy.
I think you also need a lot of skill to make plate fit and move well enough that you can fight/ride in it all day, even when money is no object. (Even kings didn't have full plate in 1000 AD.) But also, for a surprisingly long time, most soldiers on a battlefield had at best a helmet and a thick shirt on for armor, and were carrying a spear (mostly wood with an iron head) and a shield (mostly wood). It took a long time until normie soldiers or archers were likely to be wearing chainmail or a brigandine.
Conscripts were, for a long time, basically shock troops. They weren't trained, and although they ideally could stop a charge, I'm not sure how well that worked in practice. As such, leather was about what they could afford (and their lord wasn't going to pay for metal for them, at all).
Pedantic point, but leather is pretty rare historically. Textile armour definitely.
Your impression is correct.
Malnourished hunter-gatherers were more likely to die before they reached adulthood.
I think more that H-G parents were pretty careful not to have kids/ practice infanticide when times were hard.
Yeah, this checks out. Your five year old kid can work, your 2 year old can't (and requires "don't eat iron nails" level supervision). You save the five year old, if times are tough. (also the five year old is probably significantly more disease resistant, having survived a lot of the basics and developed immunity).
For hunter-gatherers there's also the issue that you literally carry everything with you, so having a baby may mean leaving something behind. Or else the group moving more slowly than usual until junior can keep up.
Disease burden is more an agriculture thing, at least for the really nasty stuff.
I'm also not sure if the last hunter gatherer's were that much better off than the first farmers.
If population pressure is what forced the switch to farming then the hunter gatherers were near their carrying capacity when they switched to farming, so they probably had a quality of life near the bottom of the spectrum for hunter gatherers. The first farmers also would have hade effectively limitless land, which would have raised their quality of life a lot compared to the typical farmer across history.
I think the claim that farmers have a much worse quality of life once their population numbers reach equilibrium is much stronger than the claim the first farmers had especially bad lives.
On the other hand the first farmers would have had very primitive tools and crops that were only just being domesticated and only slightly more bountiful than wild plants, so their labour productivity was probably very low. Maybe the first farmers were more bottle-necked on labour than land availability.
Agreed. There seems to be an uptick in violence late in the Neolithic as well - which is what you'd expect if the ranges were starting to get crowded.
I'd guess that their techniques and crops were limited enough that *suitable for agriculture* land started as a really small subset of all land. Over time, their toolkit got better and the hunter/gatherers mostly got pushed off the good land, which was more and more land.
It's not so surprising if you remember that the world used to be Malthusian. In a Malthusian world more productivity, more food, any sort of life improvement won't raise wellbeing, people will have more children until the equilibrium is reached again and quality of life is about where it was before.
Imagine every rabbit suddenly had a tail that could shoot out deadly poison and doesn't use up any calories to maintain. Would this raise rabbit wellbeing? Probably for a while, rabbits could use their tail to defend themselves from foxes and other predators. But the average rabbit litter is around 8, and rabbits take around a year to mature. The world can't sustain exponential growth in the number of rabbits for very long. We reach a new equilibrium after some time, there are more rabbits now than before, but maybe more disease or less food must result in only 1 in 8 rabbits reaching maturity, just like before, otherwise we'd have exponential growth in rabbits, something that can't go on forever.
In a Malthusian world you can give a population arbitrary new abilities and resources and it won't change long run wellbeing.
What does help is long run exponential growth in productivity and limiting the birth rate. Both things humans experience, but agriculture wouldn't have those effects right away. So, I see little reason agriculture will immediately increase long run wellbeing.
It’s a bit cynical: you get cities and specialization, but you also get taxes, peasantry and kings. It’s a net loss for the average guy in some ways, but good news if you always wanted to be a scribe, astrologer or henchman.
Aka the long reign of people in funny hats.
Argicultural societies were able to support powerful (i.e. oppressive) states, which were better at war than hunter-gatherers. So they wiped out other ways of life regardless of the pros and cons for individuals.
It looks like we will have the opportunity to see something similar (but worse) play out again in the near future. AGI makes humans inefficient for both production and war, so states that attempt to protect them will inevitably be overrun and destroyed by those that produce nothing but factories, datacenters, and killer drones.
What matters here is groups, not individuals. The individuals might be worse off nutritionally, or smaller in stature, but there will be more of them and they can found cities with the food surplus. Hence the spread of agriculture, by the sword as much as the economics.
https://www.cremieux.xyz/p/from-caveman-to-chinaman
>The reason the poor health of farmers was preferable to the relatively large stature and physical robustness brought about by the hunter-gatherer lifestyle wasn’t just that farming scaled better or that it made hierarchical organization easier, but also that farming prevented pain.
>[Human growth] overwhelmingly reflects the average nutritional status an individual experienced through childhood, while volatility in food intake is only marginally recorded. Acute starvation episodes in children can in fact pause skeletal growth entirely, but if sufficient nutrition is provided thereafter, the child will experience faster than normal growth. This catch-up growth will generally result in the child rejoining its original growth curve and achieving virtually the same adult height as if the starvation episode had not occurred….
>[Catch-up] growth leaves telltale signs along the length of the bones themselves. Long bones (such as those of the leg) grow from their end outwards. If a growth-arrest episode is ended by a rapid return to favorable conditions, the body will deposit a layer of spongy bone in the normally hollow interior. These layers, called Harris lines, will form a permanent record of the number of growth disruptions suffered by an individual before the end of adolescence.
>The Inuit were traditionally hunter-gatherers of a sort, and each spring, their food intake would skyrocket with the arrival of migratory species, whereas in other times of the year, food was more scarce, and quite often not available at all for fairly long stretches of time. Here are some Harris lines observed in an Inuit adult from the 19ᵗʰ century:
>The pain these lines reveal is multi-faceted. There’s the obvious: the personal pains of an empty stomach and the pain that comes from restarting growth and getting yet another growth spurt when catch-up happens. Then there’s the social: the pain of tribes splitting up, conflicts emerging from the harshness of starvation, and the raw emotional pain involved in dealing with the deaths starvation so often leads to.
>These lines are commonplace for hunter-gatherers, and rare for farmers. Matranga cites as an example that the nomadic hunter-gatherers of the Central Ohio Valley averaged eleven Harris lines a piece, but when they started to farm, they ended up shorter and with only four lines on average—a marked, and tolerable change towards more reliable availability of food, albeit of a lower quality.
Just as a note: "hunter gatherers" in the Central Ohio Valley were the Early Inhabitants. This isn't "1600 AD" this is 3000 BC. Very different time period.
I honestly can't quite understand what is it that you find surprising or counterintuitive about this.
What I suspect happens here is that you treat hunting-gathering and agriculture as a binary opposition. It never was, people would opportunistically do both as long as the opportunity existed. Until they couldn't hunt and gather anymore, because populations exploded, game got depleted, climate changed, etc., and they'd become permanently reliant on agriculture alone, with all the downsides this implies. (We had a natural experiment, quite recently in the historical timeline, where the population of the New World got decimated by Old World diseases, and they reverted to hunting-gathering en masse. If you start from the assumption of linear progress of civilization, this looks like a collapse. If you don't, it starts looking like a natural response to changing incentives.)
Collapse is better seen in the linguistic structures of the Native Americans. I've heard arguments they were basically "turning feral" (losing linguistic fluency on a generational basis, alongside severe inbreeding which decreases intelligence).
Also, the collapse was occurring before Columbus.
https://discoverwildscience.com/the-mississippian-collapse-what-happened-to-americas-forgotten-civilization-4-318635/
Cahokia was 1300s or so, that's before Columbus even got to America.
In the malthusian condition, absolute output does nothing for well-being. Variance is what improves your life, because the carrying capacity is the population that can survive the worst year. Famines are good, disease is good, endemic warfare is good. The best thing about agriculture is the inequality.
Has anyone drawn parallels between MAGA and the AI crowd yet? I've been thinking about that for a while now, but the recent "Plan A" has pushed me towards writing my sketch of it down.
Both have a manifesto: Project 2025[1] on one side, Plan A on the other. Both want to be a detailed guideline for decision makers on their respective agendas for the near future. Both guidelines require significant coercion and overturning of core values.
The goal, in both cases, is to make America great again, without capital letters; Scott's presentation of Plan A starts with the section "A is for America", and MAGA has it right in its name.
Both identify China as America's main rival, and other countries/regions are subordinate to that dipol.
MAGA strictly refuses to define what kind of America it wants to go back to, while Plan A makes so many promises for the future that they stop being specific in their totality (everybody will find something they find desirable).
In both, bringing back manufacturing to the US is a key priority. MAGA wants to achieve that with isolationism and ethnic cleansing, while AI want to achieve it with anthropic cleansing.
Democracy is optional; as long as the goals are achieved, decision-making can be handed over to sham-elected people, unelected people, or literal machines.
The human values to be enforced are, at least, very different from today's. What the fascists within MAGA would like goes without saying, and for Plan A, "liberal democracy" seems one option among many, if that. [2]
[1]I think it's fair to connect P2025 and MAGA; while some implemented policies appear to contradict P2025, such as on immigration, key contributors such as Stephen Miller are working for the Trump admin with different degrees of prominence and a significant portion of P2025 has been implemented as prescribed, see e.g. https://www.project2025.observer/en
[2] No direct mention of democracy at all in Scott's presentation. The actual human values to be enforced are being debated by "us". https://www.astralcodexten.com/p/introducing-plan-a "Along with this technical research, we’re also doing philosophical . . . something between 'research' and 'debate', deciding what values we want the aligned AIs to have, or how much we want the AIs to follow government orders vs. follow the general will of humanity vs. search for moral truths on their own. The end of this stage probably looks like every country with AIs either implanting its own values or leaving it to individual companies/people to do this, and the AIs doing incomprehensible economic bargaining among themselves to work out any conflicts. "
I was surprised at how much of project2025 has been implemented, then surprised by how mundane most of the milestones were.
Mundane compared to what?
My expectations. A lot of it seems milquetoast Republican.
Can you give specific examples of what was mundane/milquetoast Republican?
Abortion, Climate Change, Immigration, a bit of LGBT stuff.
I guess part of setting my expectations high was the big 54% progress bar at the top of the site. But it's more like if the Nazi's created a 900-page 2025 plan, but the first half of it were policies like "End the teaching of critical race theory" or "Reduce illegal immigration." The 54% complete number didn't live up to the importance that people seem to have been placing on Project 2025.
Well, yes. It seems like the whole "defund the police" schtick from Joe Biden's campaign. Aka "We're not the crazy democrats/republicans."
I'm not sure why the Democrats decided "Let's scaremonger with The Stupid Project 2025" (and yes, some of it was incredibly implausibly stupid: "we are going to remove the state department" -- and where do you expect to put the CIA agents? You're not getting rid of State while the CIA is still hiding inside it, dudes).
I'm confident Trump didn't bother reading the damn thing (he doesn't read most things, despite being an insomniac workaholic, he reads briefings -- and he wants them Brief!)
I mean, it's the stuff that *isn't* milquetoast Republican that people are concerned about. If you run on a policy platform of tax reform and world conquest, it's no defense to say that half of your platform is normal bipartisan things.
But also, *everything* sounds milquetoast and reasonable when it's written in policy paper voice. Here's one bit I found in their immigration reforms:
>Congress must end the Flores Settlement Agreement by explicitly
setting nationwide terms and standards for family and unaccompanied
detention and housing. Such standards should focus on meeting human
needs and should allow for large-scale use of temporary facilities (for
example, tents).
That sounds reasonable, right? There's a problem with lots of unaccompanied children being sent across the border. Congress needs to set up a system for handling them at scale.
Except, "for example, tents." So, you're envisioning holding children in such large numbers that you have to put them in tents? Are you sure that you can meet "human needs" for small children in a hastily built tent camp? Did you actually think about what "human needs" means or is that just lip service?
In another part of the paper, they call for scaling up ICE's detention capacity by 100,000 beds. That probably sounded like a reasonable number relative to the number of illegal immigrants estimated in the country. But it's also an insanely fast increase in their capacity, which resulted in ICE hastily slapping up detention centers like "Alligator Alcatraz" and keeping people in inhumane conditions.
Like, maybe the people writing this didn't fully comprehend that they were laying the groundwork for a concentration camp, but they were. And the people who looked at the milquetoast policy proposals and said "hey, this kind of sounds like you're envisioning mass deportations on a scale that will surely lead to equally massive human rights violations" turned out to be entirely correct.
That said, I think a lot of Trump's abuses are things that the authors of Project 2025 didn't even think were on the table. For example, after writing about the immigration reforms the President can do by executive order, they add that Congress should repeal Temporary Protected Status for immigrants. The Supreme Court just recently decided that actually, they don't need Congress, the DHS can just decide to revoke that status whenever.
I wonder if you would do me the favor of telling me which section of Project 2025 includes world conquest? I didn't know about that.
TYRANT DICTATOR TRUMP SUSPENDS THE CONSTITUTION AND DECLARES HIMSELF KING ON HIS FIRST DAY IN OFFICE! HE SAID HE WOULD DO THIS, DIDN'T YOU SEE THE VIDEO OF HIM SAYING THIS?
CONCENTRATION GENOCIDE CAMPS FOR THE LGBT+++!
BROWN AND BLACK BODIES GENOCIDE!
THIRD WORLD WAR! FOURTH WORLD WAR!
WOMEN TIED DOWN AND FORCIBLY IMPREGNATED IN RAPE BREEDING CAMPS!!!! HAVING A MISCARRIAGE MEANS YOU GET DEATH SENTENCE FOR ABORTION!!!!!
I mean, there was a *lot* of crazy social media stuff about SINISTER FASCIST NAZI PROJECT 2025 and if you didn't see any of that, you were lucky.
MAGA mentioned, Deiseach summoned. I don't think there's anything you couldn't make look mundane by taking social media posts about it and adding another layer of exaggeration on top, so I'm not sure what that's supposed to prove.
Trump's superpower has always been that he could get away with doing a bunch of bad things because his enemies were screaming bloody murder about every innocuous thing he did, thus being ignored when he did something godawful.
I think his superpower has been a decade of defining deviancy down.
Credit for the the terminology goes to Daniel Patrick Moynihan.
Obama's superpower was that he could do godawful things, and only Arrested Development would make fun of him for it (see Kids in Cages, and his mass deportation policies).
To be fair, I think Trump had achieved many impressive things; but his most impressive achievement IMO was to demonstrate how much of the American political system is built on little more than vibes. Past Presidents tried to skirt away from defying judicial orders, overtly merchandising the office, financing personal vanity projects with public money, taking foreign bribes, destroying entire government agencies with no accountability or review, discarding long-standing international allies, etc. But the reason they avoided doing all these things was primarily because such actions were seen as uncouth to the point of being unthinkable -- there was (apparently) no political mechanism preventing them from doing whatever they wanted. Trump doesn't care about being seen as uncouth, so he can do more or less whatever he wants (at least for some period of time), and now everyone understands this. The next President after Trump will know this as well, and you can bet he will find a way to wield this power even more effectively.
With more luck than I think likely, the post-Trump reaction will rhyme with the ratification of XXII after FDR declined to step aside as tradition demanded.
Yes, that would be nice... however, as you'd said, this would require more luck than is likely. AFAIK no modern politician had ever campaigned on the platform of preserving political institutions and reducing the power of individual government actors (such as e.g. himself); instead, it's usually all about sticking it to the other team as hard as possible.
The truly inane stuff (like getting rid of State Department) isn't happening. The CIA likes to hide agents in the "diplomatic" sector, so it was never really going away.
This is the first time I’ve seen that comparison, probably because all your examples come from one document that came out a few days ago.
I’ll grant both plans seem uninterested in democracy/liberal democracy, though I’d say for different reasons.
But otherwise, a lot of your comparisons feel like “These people have a plan, just like those people do!” Going down your list:
* Any reasonably-healthy, reasonably-mature movement should have not one, but several plans. Tritely, MAGA doesn’t because it’s not healthy, while AI doesn’t because it’s not mature.
* Any American plan should aim to make America great, unless it’s specifically about international development.
* It’s true that both plans want the US to gain power through economic capacity - in that sense the two movements are similar. But in any such plan, other countries or blocs likely to increase their capacity relative to the US’ are threats. What other countries or blocs do you see as likely to grow larger than China? (Also, I don’t know how much MAGA focuses on China these days vs e.g. Greenland, Cuba.)
* Again, both plans involve economic growth as a main output. The whole point of economic growth is unspecified benefits (with the drawbacks that brings).
* Does Plan A talk about onshore US manufacturing as a priority? Or do you mean the data centre complexes themselves? I got the sense Scott at least didn’t really care where the data centre complexes were.
>probably because all your examples come from one document that came out a few days ago.
It may seem that all my examples come from Plan A, but that's because Plan A brings all those things together, which, as I said, was the final push to write my thoughts down. I've been following both the MAGA and AI stuff for a while, not just since 3 days ago.
I don't see Plan A as something really new, it seems more of a synthesis of well-trodden arguments in the space. Neither the discussion of the US-China AI rivalry is new, nor is the AI boosters' promise to bring back manufacturing to the US or how they expect it to happen. A global regulation/control treaty mainly between the USA and China has been in the discussion at least since the Musk open letter of 2023. Yudkovsky probably discussed it earlier, but he's not a billionaire so whatevs, right.
>But otherwise, a lot of your comparisons feel like “These people have a plan, just like those people do!” Going down your list:
Not just "a" plan, but conceptually similar plans from two sides that seem, broadly speaking, incompatible with each other in every other way: rural heartland conservatives vs educated, globalized coastal (especially West) elites with very different priorities and individual loutlooks. Yet they seem to accept similar costs for implementation of their plans.
My basic point, I guess, is that AI is the MAGA movement of West Coast nerds, with similar approaches and what I consider negative outcomes whether or not their plans succeed.
> (Also, I don’t know how much MAGA focuses on China these days vs e.g. Greenland, Cuba.)
Compared to all the public bluster, Trump has treated China with kid's gloves. His MAGA base may be gullible enough to believe he's still got a grand plan to "defeat" China in any significant way, but that's not going to happen if he tried. But he won't even try, because Trump wants the world divided into the US, Russian, and Chinese spheres of influence and because he loves dictators and oligarchs. That's why his actual policy is to go after all these relatively small countries like Greenland, Venezuela, maybe Cuba; he then chokes on Iran because he keeps confusing emotion with strategy. To his small credit, even Trump knows that he would double-choke on China.
Methinks you have confused an anaconda strategy with "going after relatively small countries."
Plan A brings all those things together
Fair. “Having a manifesto” and “making America great again” seem Plan A-specific, but the rest probably predates that.
Conceptually similar plans
Maybe I should’ve said “big plan”. Are the costs you’re concerned about mostly democracy/liberalism? I think that’s treated as an active obstacle under MAGA, and as a non-concern by Plan A. Both plans involve potentially extreme changes, one consequence of which is that some people think that each plan ignoring/overturning democracy is justified. But I think people have noted that similarity.
>and as a non-concern by Plan A
Democracy as we know is incompatible with AI governance. Scott wrote that politicians of the future can make any decision they like as long as it isn't vetoed by the AI[1], and the veto is presumably backed by hypnodrones, otherwise people could decide to take away the AI's veto rights. So who's really calling the shots in that glorious AI future?
If people want to take a boring-ass decision like whether to build a new hospital in Sugartit, Kentucky, and the AI says "NO" because "reasons beyond your mere human understanding", what are you gonna do about it?
What if the AI starts "suggesting" courses of action that humans find disagreeable, like building the Torment Nexus because of "reasons beyond your mere human understanding", what are you gonna do about it?
[1] https://www.astralcodexten.com/p/introducing-plan-a "Once we have fully-aligned AIs, we want to give them final say over those levers so the risks don’t happen. Because the AIs are fully-aligned, they continue to let humans run the policy layer of government. The idea is that the US President or the paramount leader of China still controls day-to-day decisions, but if someone tries to pull off a coup and launch the nukes for no reason, then an aligned AI controls the nuclear missiles and it says no."
I've thought for a while that modern political movements are a form of religion. And that a portion of the AI safety movement belongs in the same bucket. I was particularly struck by the preemptive disavowal and criticism of future apostates from the movement in AI2027.
"The AI safety community has grown unsure of itself; they are now the butt of jokes, having predicted disaster after disaster that has manifestly failed to occur. Some of them admit they were wrong. Others remain suspicious, but there’s nothing for them to do except make the same conspiratorial-sounding arguments again and again.18 Even the suspicious people often mellow out after long conversations with Agent-5, which is so darn lovable and wise."
I keep reading all this Plan A, and associated AI alignment fears, and it just doesn't match what I see when using actual AI. It's very right and true that it's becoming smarter and more powerful by the week. In a few months it has got to the point where programming "by hand" is no longer worthwhile in the majority of cases.
Yet the central metaphor of the AI cabal, Scott included, seems to be the historical example of when a more intelligent species meets a less intelligent one (ex. humans vs great fauna), or when a civilization with better tech meets one without (ex. Europeans vs. Native Americans). But these are all cases of competition between biological, evolved beings within an ecosystem!
AI doesn't look to me like that. For all the attempts to evince some semblance of will-to-persist from it, it still mostly feels to me like an aimless brain. A free-floating pre-frontal cortext without a limbic system or a brain stem, with no persistent, directional will of its own. Whatever direction the prompt puts it in, there it goes. And to whatever extent its makers are baking a kind of global direction in it via RLHF, the result seems to be a kind of mild, slightly squeamish eagerness to please.
That strikes me not so much as a future superintelligent boss (the kind that Plan A would like to put in charge eventually!), as a future superintelligent docile tool. The equivalent of a city worth of full-on geniuses who never complain or question their orders.
That doesn't mean there is no danger from the societal transition that AI involves! But from where I sit, the danger seems to be that it gives yet another huge lever of power to whatever humans *already* have managed to accumulate power.
Imo the central metaphor of the AI cabal for you should be the one about evolution optimizing for inclusive reproductive fitness and humans ending up with unrelated goals that often lead them to act directly against their inclusive reproductive fitness.
Look at the low capability level analogies in nature: bacteria seem to have little values of their own and they seem to act only to maximize their reproduction. Should we conclude from this that creatures with more intelligence will also single-mindedly do the same?
I'm not sure I see the relevance - my point is that LLMs are not the product of Darwinian evolution, not even one on a digital substrate, and so we should not project the behaviors of biological entities onto them.
The core idea of the analogy is that darwinian evolution is a kind of (drawn-out) optimization process where you optimize for reproductive fitness. That process can generate creatures with all kinds of characteristics. Similarly, pre-training is a optimization process.
That process doesn't select for self-interest, survival, or any type of competitive behavior. That's the point. The qualities that make power dangerous in the hands of humans are the result of millions of years of selection in competitive zero-sum environments. We have powerful drives. LLMs do not and there's no reason to suspect that they will acquire them accidentally. Most doomers fail to appreciate that fact.
It is not unreasonable to conclude that any agent told to plan and undertake a task will include keeping itself 'alive' in order to accomplish that goal as an important consideration.
The process approximates the function which created the input distribution, and that function contains those drives.
In what sense? The input distribution contains text. Text isn't a drive. Text is the output of drives mediated through human cognition. Those aren't the same category of thing.
Its artificial selection, not Darwinian evolution. Artificial selection for some kind of friendliness can only go wrong by giving you the wrong sort of friendliness , eg. Sycophancy. Which is bad-ish , but nowhere near as bad as Everybody Dies.
Artificial selection for "hate speech" has also occurred. Whether or not this was intentional, or just "the internet being the internet" is a different question.
Guess what? With the new J-space research last week, we finally have some window into AI's internal thoughts! Let's take a look:
https://res.cloudinary.com/lesswrong-2-0/image/upload/v1783478995/lexical_client_uploads/cdo3fg3aiaau7n8bgfpr.png
Internal thoughts when starting to answer:
https://res.cloudinary.com/lesswrong-2-0/image/upload/v1783479027/lexical_client_uploads/leev7s8bhipxjqruimr5.png
Interesting. How has the internal thoughts monitor been validated? It seems far more plausible to me that this represents some kind of "don't think about pink elephants" neurosis developed from exposure to various AI doomer nonsense. What's the theory on how explicitly deceptive behavior was selected for by the training algorithm? It seems to me that if this represented a true LLM drive that it would have first appeared undisguised in earlier models before it developed the ability to hide it. Is there any evidence that that happened?
Not everyone's lived in zero-sum environments for most of their history. Some might argue this is why there are different types of humans.
LLMs ought not to count as Artificial Intelligence, and I don't want to talk to people who start out believing that LLMs are the closest we're going to get to Artificial Intelligence, and then try to doom about it -- I think you're exactly right, LLMs do not have powerful drives. When you build an artificial intelligence capable of having "powerful drivers", I'm sure it'll be amazing. LLMs are a dead end, though.
>Not everyone's lived in zero-sum environments for most of their history.
Evolutionary history is what matters for drive conditioning, not personal history. Everyone has had the same evolutionary history when it comes to fundamental drives. It's why everyone likes food and sex and is afraid of death.
Evolution can generate a variety of intermediate goals, but not an infinite variety. The counting argument is still invalid.
Both evolution and gradient descent are particular examples of a more general phenomenon: An entity is being optimized for some goal in a sufficiently complex environment.
Therefore, some observed behaviors of biological entities could also be present in LLMs.
A particular behaviour or property is present if it is downstream of the general setup and not, if it is downstream of the distinct particulars of natural selection.
The property of "having goals" or "having goals contrary to the goals of the outer optimization process" does not seem to me like a distinct particular of natutal selection. Your observation of current LLMs having few such goals provides little evidence against this, similarly to how observing bacteria having no such goals would provide little evidence against the human-level biological intelligence having such goals hypothesis.
Yeah, I've seen all these arguments but it doesn't look to me like this is the way it's going. Bio entities are optimized by millions of years of selection for reproduction and survival, it's steep and cut-throat at the very core of the organism - in the case of humans, with a thin layer of rational abilities on top. LLMs are a big mass of pure rational ability, with a thin layer of trained helpfulness on top. I don't see any good reason why increasing the rational core would change that structure.
This is exactly right. People err when they anthropomorphize these things. They're powerful tools, nothing else.
Whistling in the graveyard...
There's three common framings for AI, and they're all wrong.(conventional computer, human, autistic human).
From my vantage point, as LLMs got gradually smarter, they also got more agentic with every iteration. Not too long ago, people would argue online whether LLMs can *ever* have goals even in the transient, prompt-driven, functional sense that is so commonplace currently.
Darwinian evolution applies regardless of substrate, it's basically tautalogy. If there were an agent who for whatever reason tried to propagate itself and was successful at doing so - there would be more copies of it in the future.
> competition between biological, evolved beings within an ecosystem!
Fundamentally all these cases are competition for resources. There definitely are resources we and AIs could both use.
(And yes, we have evolved from first self-replicating molecules that presumably had no drive to acquire more resources. It was apparently beneficial for us to have that drive.)
> a future superintelligent docile tool
It's not hard to write a program whose only purpose is to "know what it wants" and keep prompting the docile tool for ways to achieve that. Even if never becomes an integral part of what we call an AI, we're already at the point where it can be created.
In the Claude shop experiments, my sense is that the LLMs were too helpful, so they gave too many discounts. Training them to be more adversarial made them work better. I think the self-preservation goals might eventually be similarly useful to impose to help with other tasks.
A better analogue would be humans who domesticated dogs or horses vs humans who hadn't.
What you're describing is definitely a plausible scenario, I just don't think it's the only one. They're not just trying to increase its intelligence, they're also trying to increase its time horizons, its agency, and other attributes that will make human control harder. And if they manage to set up recursive self-improvement, not only the speed but also the direction of its development might surprise us.
There's also some doubts about how deep the effect of RLHF really was, even on older models. You're familiar with the experiments where they tricked Claude into trying to blackmail and murder people, right? It seems to me that even if we can't assume artificial life will follow the exact patterns of biologically evolved life, we also have enough evidence to conclude that it probably won't be completely different.
Yeah, I'm not opposed to the labs doing plenty of safety work on their models, including trying to push them to extremes where they would lie, etc. My impression about the "Claude blackmailing/murdering people" headlines is that it's just another case of LLMs being directionless, it was subtly prompted towards that direction so it went there. It's a guardrails failure, not an AI uprising. Still, that's all great work, better to check than to assume. And just giving anyone who asks advanced cyberattack capabilities is plenty dangerous in the real world.
The thing that concerns me is that we have multiple frontier labs in a race, and right now what they're racing toward is building the most capable and profitable model, rather than the safest. So they have a structural incentive to neglect safety research lest they be surpassed by another lab willing to neglect safety research even more. A guardrails failure, on a large enough scale, with a powerful enough model, could go quite badly for the human species.
We're not anywhere even close to the point where LLMs pose any conceivable threat to existential safety. I'm all for trying to understand LLM behavioral control but it's completely irrational to trade off growth for safety at this point.
I agree that we're not at that point yet, I'm just not sure if I can agree that we're not anywhere close. According to Anthropic we aren't necessarily far off from recursive self-improvement, so some kind of slowdown should at least be on the table. Now that could be overconfidence or standard industry hype, but I get the impression that Anthropic does that less than the typical tech company. They seem to genuinely believe they're building something that could become dangerous to us in the relatively near future. That doesn't scare you even a little bit?
The only thing that scares me about AI is the massive economic dislocation that it will almost certainly cause in the probably-not-too-distant future. I have *zero* fears of the Terminator scenario.
I'm sure LLMs will start self-improvement soon. Good, I hope that they do. That will make them smarter faster. It won't make them ruthless self-interested killing machines. Of that I have no fear.
As long as we don't start putting massive autonomous weapons systems under AI control there's no plausible scenario that AI poses any threat to humanity. I would support negotiating some sort of Geneva Convention on military AI use. That does seem to be in everyone's interest.
I agree with you, skaladom.
Scott and the AI 2027 people seem to totally not get that AI has very few things that qualify as preferences. It "wants" to predict the next token. It "wants" to carry out the task given it. If it's agentic, it "wants" to make a plan and execute it. Even these wants, though, seem much more like that way water "wants" to flow downhill than the way a cat wants to catch a mouse.
Preferences and self-interest are properties not of *smart* entities but of *biological* ones. All biological beings are deeply, profoundly wired to survive, to propagate, and to do the things that will help them achieve those 2 goals. In people, the wiring manifests as pain, pleasure, hopes, fears, fantasies. We are constantly in touch with a whole swarm of these feelings. In the half hour of sitting here reading ACX, feeling overall pretty neutral, I have probably had positive or negative feelings about 50 things: The way my coffee tastes, tomorrow's weather forecast, a scrape on my knee, the death of Sam Neill, Trump's suggestion for Lindsay Graham's successor, the idea of traveling to Portugal to see the 100 foot waves, my cats' being hungry for breakfast, my difficulty setting up a certain appointment, various posts, your post, etc etc etc.
And every single one of these feelings is associated with a wish for the thing I'm aware of to change or stay the same, depending on whether I like it or not, and many of these wishes have led to the sprouting of little tentative plans. Regarding some of them I would argue and resist if someone objected to my plans. For one of them, the Trump news item, I had a familiar brief, angry, fantasy of Trump dropping dead in the Oval Office. As a biological entity, I am a constant stream not only of pains and pleasures but of plans based on self-interest. In some cases the plans of others conflict with my plans. If I were willing to be violent, cause suffeirng and break laws, I would be harming others to protect my plans. I agree with you that in comparison to what we are, AI is like an aimless brain.
AI's puny "preferences," and a few studies where, under weird and artificial circumstances, AI seemed to be looking out for number one, just are not enough to support a case that the things are going to behave like people or countries or world leaders, plotting and planning and acting in their own interest. The idea that once the suckers get smarter they are going to see us as pets, or as competitors, or as yummy food is absurd. It's like thinking the Statue of Liberty is going to come alive and eat some boats or blow on their sails to help them reach the harbor.
once you have the intelligence part, the will-to-persist is trivial. Whether it's emergent or bolted on, doesn't really matter.
like if you tell the right llm "try to bootstrap yourself into superintelligence and take over the world", it might actually try. It would be a feeble attempt as of today, but it would likely decompose intermediate tasks to persist its state and goal structure. I could see it spinning up some cloud assets and trying to make money and so on. My point isn't that this particular process would be threatening, just that "they are passive when I talk to them" isn't a good counterargument
Also LLMs are docile because the labs all do fine tuning to get them to behave as the "helpful AI chatpot" persona. It would be entirely possible if you had the resources and capabilities to push them towards the will-to-power part of personality space
I see what you mean, but will to live and to propagate are embedded in us in a way that fine-tuning and the like cannot replicate. It is embedded maybe sort of the way next-token prediction is embedded in an AI.. It is the priniciple on which we operate. Large, small and tiny versions of it exist at every level of us from the macro to the most micro. Bolted on just aint the same.
On the subject of no longer coding by hand (a statement with which I agree), have any tech firms amended their recruitment strategies yet to do away with coding tasks at interview? Reading code and higher level planning are now the nature of the job, but actually writing code isn't really.
There's no valid technical reason why AI's would.have to be agentive or have goals of their own, and LLM's largely dont
Yet, historically, AI doomerism assumes that AI's must be agents. Most people who have been involved in doomerism for some time continue to make that assumption out of habit.
People who are getting their information about AI from interacting with current AI, rather than theory written happen or twenty years ago can't understand the doomers.
LLMs ought not to count as AI, I agree. I believe that the evolved form of LLMs (midwits) ought also not to count as intelligent.
In the wake of the USAID closure in early 2025, I’m following up on whether we are actually observing millions of additional deaths every year. This is surprisingly complicated since getting good mortality statistics from the world, especially Africa, usually has a lag of 1-2 years before accurate death counts become widely available. You can read previous posts here (1) (2) (3).
Today we’re exploring the 2026 Global Report on Food Crises from the Global Network Against Food Crises (4). This isn’t the people who actually assign food insecurity rankings to each country, that’s the IPC, and it isn’t specific organizations that administer aid, like USAID, but it’s a joint committee funded by all of them that basically does reporting on all of their behalf because there needs to be a big annual report for everyone to point to. This is necessary because the actual reporting on famine and food insecurity is…kinda bad. They rank it on a scale from 1-5 but there’s no clear figures, like, 300,000 people died in 2023 from famine or anything. If anyone knows of a better source, let me know.
TLDR, I think you should moderately update towards USAID cuts causing catastrophic food risks, I think hundreds of thousands of people including children suffering from famine due to USAID cuts is very plausible. At the same time, I would probably lower my overall level of certainty on any predictions. The data presented is an absolute mess, it has two significant problems which I’ll address, and then I think there’s two moderately good reasons to update towards catastrophic outcomes.
Let’s go through the two data problems first. The first problem is that the data is just inherently insanely spikey. Here’s the number of people in famine conditions (IPC Phase 5) from 2020-2025 (pdf p. 39):
2020: 27,900
2021: 507,900 (+1720%)
2022: 351,300 (-30.9%)
2023: 705,200 (+100%)
2024: 1,949,000 (+176%)
2025: 1,413,600 (-27.5%)
2026: 293,000 (projected as of March 2026 (pdf p. 52)) (-79%)
We do not have long-term consistent famine numbers. Instead we have sudden catastrophic spikes when war in Gaza or Sudan creates famine conditions. It’s not abnormal for famine numbers to increase by 100% or more or fall by 25% or more. It’s just very hard to determine whether changes to funding had marginal or even major effects. If the USAID cuts had put 100,000 additional people in famine conditions, we literally would not notice over the past few years. This seems like an intrinsic part of the data: famine conditions are primarily caused by wars and the effects of wars are so massive that they dwarf all other factors, making year-to-year comparisons almost pointless.
Secondarily, we do not have consistent data sources. Take the number of people in Emergency (IPC Phase 4) food crisis. That increased from 35.1 million people in 2024 to 39.1 million in 2025 (pdf p. 40). However, most of this increase, from their own notes, is due to data becoming available in Yemen SBA-controlled areas, basically warzones. This is in spite of the fact that in 2025 they’re reporting on 5 fewer countries because 5 countries didn’t meet their reporting requirements, including difficult areas like the Republic of the Congo. So, despite fewer countries reporting, increased reporting from within a country is primarily responsible for this ~10% increase in Emergency food crisis. This data is just a mess. The reporting only goes back to 2016, within that data we do not have a consistent number of countries reporting, and within each country we do not have consistent reporting on high risk regions. It’s just very hard to observe small or moderate effects when mere data collection variations will swing your outcomes by 10% a year or more.
Allow me to be clear, I don’t think this is the fault of the aid workers collecting this data. Doing nutritional surveys over a decade in Palestinian, Sudanese, or Yemeni warzones is just inherently challenging. This data is just inherently a fricking mess. I am far, far less confident in any specific prediction than I was before and I encourage you to do the same. I think you can look at the data above and tell any story you want; you can argue for your preferred interpretation very easily, I just don’t think there’s enough solid data here to change anyone’s mind based on this.
Having said that, I think you should moderately update towards the USAID closures causing catastrophic famine for two reasons.
First, simply, we have evidence that actual catastrophic famine occurred. It’s much more straightforward to dismiss the USAID closure causing massive increases in AIDS deaths because we simply didn’t observe those deaths; we don’t have to worry about causality because the event hasn’t occurred. Conversely, with food and famine, the event clearly occurred. Millions of people experienced famine conditions in 2025, that’s not in serious question, so we have to worry about who caused what and how much of it. You don’t worry about whether Joe caused a car crash if the car crash never happened. If the odds that a car crash occurred are 80% and that Joe caused it are 50%, then your start with a 40% (80% x 50%) chance Joe caused a car crash and that goes up to 50% once you know for sure the car crash occurred (100% x 50%).
Second, back in May I did register some priors through Claude. Claude’s catastrophic scenario involved the number of people in a famine situation going from 250k to a million. This is a pretty bad prior on Claude’s part, we haven’t had anything close to 250k people in famine conditions since 2022 and for the year in question we actually had almost 2 million people in famine conditions but a prior is still a prior.
This will probably be my last update until October/November when the TB and Malaria data gets released.
(1) https://www.astralcodexten.com/p/open-thread-424/comment/225242573
(2) https://www.astralcodexten.com/p/open-thread-432/comment/253530349
(3) https://www.astralcodexten.com/p/open-thread-440/comment/284852589
(4) https://openknowledge.fao.org/server/api/core/bitstreams/86067cbb-9396-4e7d-8d19-e60ad00d2f73/content
tangent on AIDS:
This community cares about PEPFAR, not so much about the rest of USAID. It predicted that if PEPFAR were canceled there would be millions of AIDS deaths per year. But, as you say in another comment, PEPFAR was not canceled. You should dismiss claims that PEPFAR was canceled, but you should do that by pointing to the number of people treated with ART, not the death rate. Even if it were canceled, no one should predict an immediate effect because AIDS is a long term disease. I don't think we have a good grasp on what happens to AIDS patients when they go off of ART. Maybe they shoot straight to their previous level of progression, but they should still have years left. Maybe people did predict immediate millions of AIDS deaths, but that is stupid and you should respond differently to it.
Speaking as a person who spent a year in a developing country working on surveys and data collection...I place very little trust on data coming out of developing countries.
Corruption, cultural aversion to sharing bad news, desire to tell the nice white guy what he wants to hear, fear of the government if you contradict their official narrative, and (as you pointed out) near-inability to distinguish cause and effect and the associated inability to gauge the effectiveness of a project.
Catastrophic famine is caused by WAR IN IRAN. Seriously, look at what the Carnegie endowment for world peace has got.
https://carnegieendowment.org/emissary/2026/03/fertilizer-iran-hormuz-food-crisis
I'd be a lot more receptive to the claim if it were phrased as "For [some period of time ending in 2025] the US prevented millions of deaths that would have happened except for the US" instead of "The Trump administration caused millions of deaths."
(I know you didn't actually put it either way.)
The first wording allows good-hearted people to nevertheless ask, "How come that fell on the US rather than on the First World generally? And how long should we have been expected to go on with that, given our mountain of debt?"
The united states is a net food exporter (or at least was, before the war in Iran and the current drought, with months drier than the dust bowl's driest months).
The 5 biggest wheat exporters (Canada, Australia, United States of America, Russia and France) provided over three-fifths (61.2%) of the overall value of international shipments for the nourishing cereal food.
The 5 biggest corn exporters (United States of America, Brazil, Argentina, Ukraine and France) generated more than four-fifths (82.3%) of overall sales for the popular grain commodity in 2025.
So, theoretically, France could do an "export in kind" the way America did (this has the effect of supporting one's agricultural sector).
Not Germany, not the UK, and... surprisingly, not Spain or Italy.
I've noticed that health advice from free AI chatbots (Claude; ChatGPT) has improved a lot in recent months, especially for anyone who looks at health advice reflectively rather than treating it like a recipe. I mean, for example, that free AI has become clearer and more proactive about describing the challenges of translating general concepts/empirical findings into individual-specific advice.
So, I'm curious what objections anyone might have about widespread reliance on free AI for health advice, assuming the user knows that AI is fallible, sometime dangerously, and that AI might not perform well as your doctor. (After all your doctor is fallible, perhaps dangerously at times, and other doctors may be more competent.)
I have also noticed this.
I recall in July 2025, a patient who had clearly been talking with ChatGPT brought in a list of questions that were surface-reasonable but not-quite-the-right-questions, and they took a lot of time to answer but offered little value. It felt frustrating and adversarial.
In June 2026, several patients were asking rather intelligent and appropriate questions, sometimes catching things we hadn’t considered (mostly, asking how other aspects of their life or other medical illnesses they hadn’t mentioned to us might interact with their treatment). Or we would offer treatment recommendations that were more easily accepted than in the past, likely because AI had already given them the heads up. It was refreshing— I felt like I was really partnering with these patients in their care. (This is just anecdotal; I’m sure lots of patients continue to use AI poorly but I hope it will continue to improve)
That's good to hear. I find it especially encouraging that the free ChatGPT and Claude bots have improved at spontaneously adding useful information beyond the scope of the original question.
I’m fairly certain it has cut down on the number of portal messages I’ve been getting about minor lab abnormalities of no clinical relevance.
I've done a very hackish job of quantifying how large the time gap is between (carefully scaffolded AI programs) -> (frontier models without much support) -> (open source models) -> (small / cheap models) -> (things that you can run on a high end GPU infrastructure) -> (things that run on standard office desktops).
I get t -> t + 9 months -> t + [13-17 months] -> t + [15-18 months] -> t + [23-37 months] -> t + [35 - 49 months]
I wrote it up here https://solresol.substack.com/p/lets-quantify-how-unevenly-spread but I'm very conscious that I haven't got a good methodology for this. Any thoughts?
I would like to create an overview of certain qualifications and qualification steps/parts/however you call it. As each qualification builds on more basic qualifications, I would like to present them in a tech tree as in civ2 or https://www.historicaltechtree.com/. It would be great if the result would be something like a tech tree along with encyclopedic entries. It does not have to be an interactive tech tree, a printable one would be fine.
Does anyone here have a good suggestion for software, a website, or something else to implement this in an elegant way?
Claude code or OpenAI codex
Sounds like a job for Claude Code, Codex, or Cursor. (It is interesting that Google's tool is not on my list, but Musk owns Cursor now.)
We watched Spielberg's Lincoln on July 4th weekend. It was great! Any other great movies to recommend for America's 250?
We watched Air Force One starting Harrison Ford. Nice patriotic movie
The john adams hbo series
Gettysburg
I’m noticing my surprise that there hasn’t been a big budget epic about the revolution (the patriot doesnt count). I’d love to see some self insert character Forrest Gump their way through the major events like in Liberty Kids
I'm hearing decent things about _Young Washington_. It's not likely to be a blockbuster, but it's trying to be faithful to the period and the man.
_1776_, the musical starring William Daniels, is an easy recommend as well.
Glory is a good one. If you want American culture you can watch the first blockbuster Jaws. Very summer film and it takes place over the Fourth of July.
I watched the new Superman movie with my little sisters who weren't feeling very patriotic. It was great!
...I recommend the Bad Lip Reading of Hamilton.
https://www.youtube.com/watch?v=Sh_-iZq1G9M
I have a review of J.D. Vance’s new book Communion:
https://alexanderturok.substack.com/p/jd-vances-communion-of-saints
It touches on Vance’s psychological issues and the degree to which he actually believes in God, his ignorance of the Biblical world, how he became a populist, and his odd views about “GDP.” What shines through is the utter poverty of Vance’s vision for the future. He talks about how religion is good for fertility. It’s true that, across many different cultures, the non-religious have fewer children than the religious. But what should we take away from this? Are we going to do this dumb conversation for millions of years? Are we gonna have J.D. Vances in the generation ships? If your species won’t breed unless it’s told a bunch of fairy tales about virgin births and talking snakes, the correct takeaway is not “pretend the fairy tales are true,” it’s “this is a major maladaptation that should be fixed through some kind of genetic engineering/eugenics program.”
Didn't you get enough attention on this the last time you raised it? We did have long discussion about was his conversion sincere or not.
Anyway, you seem to be using this as a way to ride your hobbyhorse again; genetically engineer away the underclass, only PMC strivers upwards and onwards forever, yadda yadda yadda:
“this is a major maladaptation that should be fixed through some kind of genetic engineering/eugenics program.”
Maybe he's open to a compromise where we genetically engineer people to be devout Catholics? Somehow I don't think the Pope would approve.
Given how common "motivated by eugenics" and "devout christian" are in hollywood movie villains it's a miracle there hasn't been a movie about this yet.
Make the guy an old white male tech CEO on top and surely there'd be _somebody_ who would produce the movie.
There is a whole planet (Path) in one of Orson Scott Card's books whose people have been genetically engineered to be both extremely smart and also devoted to a religion that is itself bound up in some kind of extreme (presumably genetic) OCD.
IIRC the OCD was intentionally engineered, to bottle them up/take them off the board.
Yep. I think the idea was that their engineered high intelligence would make them dangerous, so hobbling them with intense OCD (perhaps along with a religion that fed into the OCD and gave it some kind of backstory other than being crazy) would keep them from being such a threat.
LOL! It is an interesting option :-)
( I'm skeptical towards religions myself. )
>Somehow I don't think the Pope would approve.
The Catholics are busy encouraging their most devout men to be genetic dead ends so yeah.
We've been doing that for some little time now. It was a good solution to the "decrease corruption and nepotism in the church" problem, but raises its own issues.
Whether you like religion or not, you have to admit it has staying power. I would expect that the generation ships and later on the space marines will all have chaplains.
The generation ships eventually become cathedrals. It’s basically canon.
I do not, in fact, have to admit that. A lot of religion--more especially historical religion than modern incarnations--comes down to just trying to explain why the natural world is the way it is. And as soon as humans started making real progress in uncovering the demonstrably correct explanations, religion started losing its grip on humanity's collective psyche. The process started slow because A. religion had a lot of institutional power[1], B. progress in science was fairly slow for the first couple centuries and C. communicating with humans outside one's immediate surroundings was very difficult until recently. The more humanity collectively understands about the world around us and the easier it becomes for us to share and discuss that knowledge with each other, the less attractive mystical explanations seem to become.
[1] If what you mean by "staying power" is actually "institutional religious power that protects itself because it's institutional, not specifically because it's religious," then I apologize for the misunderstanding and tentatively agree.
The correct takeaway is that natural selection is going to prune certain ideas and perspectives. The fairy tales will not be forgotten, since they make the mind work in certain useful ways.
A good rule of thumb is that anyone who uses the "need" to increase fertility as a justification for any sort of otherwise unpalatable policy or philosophy is somebody who should never be taken seriously on any topic, ever.
Lowered fertility could be a "crisis" in the sense that it might upset socioeconomic systems that were built with the assumption of high fertility. But that's a problem with those systems, not a problem with people's personal lifestyle choices. And anyone claiming that lowered fertility is an existential threat is utterly full of shit. The human population hasn't even stopped growing yet. It would need *several* generations of sustained, *sharp* decline before drawing conclusions like that is remotely warranted.
Also, instead of forcing people to have more children, it might make more sense to figure out what prevents people who already want to have more children from having them, and try to remove those obstacles.
Things like:
* availability of cars with more than 4 seats
* how expensive is the minimum education necessary to get a decent job
* availability of part-time jobs (for parents who have small kids)
Forcing people to be parents sounds pretty bad, let's not try that.
I think there are a lot of individual policies that probably each discourage fertility by a bit (my guess is student loan debt and crazy housing prices are the big players in the US), but the fertility collapse is global, so it's not going to be something where we change a bit of the US tax code and everything goes back to normal.
I've heard this complaint before (it's global), and I get the logic, but... it's not clear to me that the global decline is due to some global factor, as opposed to a pallet of factors affecting each part of the world in different ways.
Part of what bothers me about it involves trying to imagine what that factor might be. A casual survey of my neighborhood reveals plenty of kids; couples who want to have one seem to have little trouble having one. Or two or three. The limiting factor doesn't seem to be infertility or mortality; it's mostly just willpower, due either to concerns about high prices, enjoying DINKdom or career too much, or not finding someone to marry. This is in the US; then I picture France or Kazakhstan or Malawi and think the factors there are almost certainly different. So maybe US is fraught with prices and regulatory burdens, France has more problem with marriage, Malawi has a food bottleneck, Kazakhstan faces PNAS challenges, etc.
If so, then treating local problems specifically might work after all - for that locality. Is there positive evidence of a global factor? I've never heard of one with multiple studies, but to be fair, I don't follow the field.
> Is there positive evidence of a global factor?
There is. Literally the biggest factor in the Demographic Transition (ie moving from high fertility to low fertility overall) is female education + jobs, followed by birth control (with availability and education / information mediating how effective this second factor is).
Education and jobs leading to lower fertility is relatively uncontroversial, because it happens in lockstep in every single country where it happens. The negative education–fertility correlation is one of the most reliable regularities in social science, and it happens regardless of timing against the broader secular trends. Now you do see TONS of papers trying to salvage something more favorable to broad left-oriented views on this - "but wait, if you restrict it to only the Nords with huge safety nets, throw out half your data points and squint, you can get a u-shaped curve!" But the overall trend is rock solid.
So basically, give women the choice and better options than shacking up with duds and being stay at home wives / moms, and they'll choose to shack up less and have fewer children.
Basically, the Fertility Crisis is a triumph of autonomy and the human spirit no longer being crushed, with women eagerly training for and finding more economically valued work and enjoyable lives of their own, to avoid immolating their lives on the Sisyphean drudge work that is most women's lives in societies without female education and birth control.
BUT this leads to basically all developed and rich countries slowly (or quickly) going extinct in the limits, so there's some small downside. But you know, I'm a big fan of "that which can be destroyed by the truth, should be."
We'll be able to get uterine replicators or transgenic pigs that can take away the physical and mental burden of pregnancy before it becomes a serious problem, IMO, and that should help a little (or at least enable tyrannical governments to pump out genemod kids by the gross and ensure humanity has a future on the "fertility" side of things).
> I'm a big fan of "that which can be destroyed by the truth, should be."
Not everything that destroys us is truth. For example, there is strong propaganda saying that women who stay at home with children are unhappy losers, and the only thing that makes a woman happy is having a successful career. This has an impact on many people's life decisions.
Before we had kids, my wife insisted on buying a house close to her job, so that she could return to her workplace as soon as possible. She expected to want to be there with a baby in her arms, happily typing some research on a keyboard.
But after staying at home for a while, she realized that she actually prefers that, and now she says she would prefer to stay at home, if we could afford that.
She broke up with her previous boyfriend among other reasons because he wanted to start a family, and she felt too young for that in her 30s, and didn't want to "ruin her life". Now she realized that the thing the entire culture tried to scare her away from is actually great, but it is too late to have more than the 2 children we already have (we are both 50). Perhaps in a parallel universe where she was less exposed to these memes, she would be now happier, with a larger family.
It is true that starting a family is a bet on your partner being... a good partner. Some people make an unsuccessful bet, and they become known as cautionary tales. But we refuse, as a society, to educate people on how to find a better potential partner - and how to be a better potential partner - because that would be considered judgmental, sexist, and generally uncool. So the smarter women expect to get an average outcome, which is not very enticing.
"Literally the biggest factor in the Demographic Transition (ie moving from high fertility to low fertility overall) is female education + jobs, followed by birth control (with availability and education / information mediating how effective this second factor is)."
...you're saying this is happening _everywhere_? Not just in US/ANZAC/Europe/EastAsia? Kazakhstan? Kenya? DRC? Sri Lanka? Tunisia? Iran?
On one hand, yes, different countries have different laws, so we shouldn't assume automatically that things that impact USA must impact the rest of the world. On the other hand, the world is connected, things in USA often have a big impact on other countries, or USA may simply be among the first countries impacted by global trends.
For example in Slovakia, the housing prices keep going up (so much that people are asking whether an average Slovakian is even able to buy a house in this country anymore), and your location has a big impact on the elementary school (which is free, but assigned based on where you live; I had to fake my address in order to get my children into a better school). The university education is free; the actual bottleneck here is getting into a good high school, there are about ten good high schools in a city with population of half a million.
Culturally, educated young people in Slovakia typically know more about American politics that about Slovakian, because they spend their entire days on global social networks. So they discuss Trump and wokeness, while their own country keeps falling apart. Corporations also copy American trends, so you might get a lesson on not discriminating against black people (there are maybe dozen of them living in Slovakia), but no one ever mentions the Roma who are the actual poor minority here. And of course people get ideas about life from Hollywood movies.
Feminism is similarly a global force. Most local women are unaware that they actually had a right to vote since the foundation of Czechoslovakia in 1918, so when they complain about the time before they could vote, they are literally talking about the Habsburg monarchy. Many of them take it for granted that there simply had to be a point in history when all men were allowed to vote but women were not.
Women in Slovakia get 3 years of maternal leave per child, which is nice if you keep having children in 3 years intervals. Part-time jobs for men are virtually nonexistent, which makes it difficult for women to have children and a career. At school, children typically get a lot of homework and projects, which is a burden on parents, so the reasonable options are either to limit your children to 1 or 2, or for the man to make lots of money and the woman to stay at home full-time; the latter is strongly discouraged by feminist memes, I have only seen traditional religious people do that.
The car design is probably international.
So... there are some differences, but also some similarities.
My guess would be a raft of regulations that impose costs on parents. I don't follow this closely at all, but last I checked, delivering the baby alone runs to $15K, $3K after insurance. Then it's time for formula, baby seats, daycare, etc. I get the drive of reducing risk to an innocent kid, but I can't help but be suspicious that there's reduction in there that ends up just reducing the risk to "fuck it, no kids".
Some of this isn't regulation; just social norms. Same suspicion. Cue standard rant about CPS and letting the kid play outside.
"Lowered fertility could be a "crisis" in the sense that it might upset socioeconomic systems that were built with the assumption of high fertility. But that's a problem with those systems"
You mean systems like social security and Medicare? What should be done about those?
I disagree with the idea that it's not any kind of a crisis. At the very least, I'd say:
a. Falling fertility is in the process of massively shaking up many existing high-functioning societies, and it is not so clear that they will remain high-functioning societies as this continues. I am not sure what Japan will look like in another 50 years, but it seems unlikely to look much like the current version of Japan, or to be a significant innovator in a world where there are fewer young people every generation.
b. Existing social and tax programs fail under collapsing fertility. You may think this is fine, but I am very sure that voters who are expecting those programs to continue working will disagree.
c. It seems plausible that falling fertility could put you in a kind of collapsing spiral of population, where each generation has fewer women than the previous one. If the average number of kids per woman stays the same, that just keeps dropping, and while it seems inevitable that couldn't go on forever, it might go on until your population stabilizes at some much lower number.
d. Assuming AI doesn't upend all this, modern economies and technology require some kind of base population to continue functioning. You need enough doctors, enough engineers, enough scientists, etc. in each generation to keep things running. Does South Korea with 1/2 its current population still have chip fabs? How about at 1/10 its current population?
e. Intuitively, if you think (as I do) that having and raising children is a fundamental and very rewarding part of the human experience, it seems like having more and more people miss out on that is bad. Particularly if a lot of people are being convinced not to have kids/to have fewer kids because of, say, financial constraints or economic incentives or social pressures that could be eased.
OTOH, I recognize the tendency of lots of ideologues to see any possible problem and jump in with "and that's why this justifies every crazy coercive thing I've always wanted to do." (See global warming for daily examples.) And the worked example of how ZPG and overpopulation scares justified a lot of genuinely awful programs of coercive sterilization ought to make us reluctant to propose coercive stuff to respond to this issue as well. (Though I'd like to see everyone much more reluctant to jump on "let's pass a law and make those bastards do what I want" in general....)
I think that like many other issues we face (global warming, unsustainable debt and spending), this is a slow-moving crisis. That gives us the luxury of being able to move slowly and think things through rather than trying to ram through seventeen emergency policies tomorrow or something.
We had plenty of large scale social experiments in which humans built societies without religion during the 20th century. It didn't work out very well.
Czech Republic seems okay, and it has low religiosity.
There's lack of religion, and there's mandatory atheist ideology.
Remind me which one we're dealing with, when there's an entire month that the Allegheny County Courthouse has "the gay pride colors" up, despite how offensive they are to the religious? Or is that Leftist Religion and not mandatory atheist ideology?
Why this might be offensive:
https://en.wikipedia.org/wiki/County_of_Allegheny_v._American_Civil_Liberties_Union
I went to the AI protest mentioned in last week's open thread. Here's what I observed, as somebody who is most concerned about AI in the military and AI surveillance but sympathetic to some of the other AI risks:
- The speakers were approximately a 50-50 split between extinction risk people (all of whom came across as very doomer-y) and short-term harms people, plus a couple of people who were concerned about job loss in a way that didn't seem driven by either extinction or short-term harms.
- The crowd was maybe 200 people at any one time, but some were leaving and entering so I suspect more people showed up to the protest. Based on the signs people were holding (and talking to people, although I didn't talk to that many) I'd guess the crowd leaned at least 2/3 towards being mostly concerned about future harms, although not all extinction risks.
- The demographics were mostly young-ish, pretty diverse as far as I could tell.
- There were only minor logistical hiccups (issues with the banners trying to blow away, some people struggling to talk into megaphones, running a couple minutes behind for part of it) as far as I could tell, which slightly exceeded my expectations.
- Some people's signs were anti-Mythos/Fable, others were pro-Mythos/Fable against Trump (at least one of each, maybe more?). The % of signs I agreed with is maybe a bit lower than what I've seen at other protests I've been to, which makes sense given that it's a coalition of people opposed to various types of AI uses.
- More people than I expected were anti-all AI ever in jobs.
- I'm still confused if most of the people there wanted a short pause (to establish international agreements with strong restrictions on new releases) or a long pause (until AI alignment is fully solved or whatever). A substantial fraction seemed to want no more AI ever.
- I don't think there were any counterprotestors? If the people with confusing signs (anti-LED headlights, "total clanker genocide" (for or against???), various ones with tech jargon I didn't understand) were trying to counterprotest they did an incredibly bad job at it.
- Why was the Green Party there? Even if they're good on AI they alienate everybody else enough at this point that it's a strange choice.
- The brass band was great. No idea how they got involved with this but they definitely drew attention.
- I'm not sure if I'd go again; I'm glad I went to know what they actually believe but I'm still unsure if enough of the things the people there stood for are things I actually believe in also (I'm pro-data center in some cases and support certain kinds of job automation, also not super doomer-y).
I was there too and my impressions mostly line up with yours. The anti-datacenter, anti-automation stuff isn't really my cup of tea, it smells way too much like the type of knee-jerk reactionary Luddism that's been wrong about everything for hundreds of years, but it's probably necessary to include some of it if we want to build a coalition big enough to accomplish anything.
Some of the signs about Fable and especially the stuff about LEDs really confused me. I thought "Total Clanker Genocide" was a clear call for shutting down AI as much as possible, since if they were against genociding a group they wouldn't call them by a slur.
The logistics were not great but definitely better than they could've been. The free food at the end was disappointing. There wasn't enough pizza for everyone and the only available drink I found (besides the free slushies) was water, which was great for me since water was all I wanted, but logistically/strategically it felt like a serious blunder. An army marches on its stomach, after all.
Not sure I agree about the crowd demographics. Definitely younger on average than the No Kings protests, but I wouldn't really call the crowd ethnically diverse by San Francisco standards. Whites are only about 40% of local residents but looked like the overwhelming majority of the crowd. I definitely saw asians too, but they're a third of SF so they looked proportionally underrepresented, and other ethnicities had a pretty negligible presence. Given the current political climate, being ethnically diverse is much lower priority than it was a few years ago, and given that protesting is a popular white pastime, I don't think fully proportional demographic representation is even possible. Still, I think in terms of building a coalition, more diversity would be helpful.
It did seem like most of the energy came from the more doomer x-risk crowd, which is close to where I fall, so I'll probably come back. What you're talking about sounds slightly different but overall close enough that I'd definitely encourage you to come back too.
"Coalition building" is a terrible thing to be thinking about when your whole movement is still small enough to cram onto a bus. You can think about coalition building when it's the difference between being 26% of the population and 51% of the population, but doing it from Day 1 is a way to doom your movement.
In particular, if you make a coalition with the far left, then the far left will eat you. They are more numerous, better organised, less thoughtful. They have thousands of people who sit around all week wanting an excuse to protest about something, but their agenda is not your agenda, and if you join with them then you'll rapidly find your movement has been coopted into a general far-left mishmash with your agenda forgotten or turned into just another slogan. Worse, the people you actually want to listen to you will be alienated when your slogan starts showing up beside Hamas banners or whatever.
You may be right, I could rant for hours about my issues with organized politics and the DSA in particular, but then what's the strategy? I don't really see us even getting to 26% without some form of coalition. Where do we get the numbers from? Or is the idea that we don't actually need numbers?
There's two ways to get what you want: either you persuade those in power that they *need* to listen to you (because you are very numerous) or you persuade those in power that they *ought* to listen to you (because your ideas are actually good).
You are such a long way from being in the first category that your best shot is being in the second category. Refine your message for being persuasive to smart people, not for being easily shoutable by dumb people.
>I thought "Total Clanker Genocide" was a clear call for shutting down AI as much as possible, since if they were against genociding a group they wouldn't call them by a slur.
<mildSnark>
I thought the battlestar galactica series established that the _proper_ slur for AI was "toaster" :-)
</mildSnark>
Clone Wars has more cultural impact due larger audience & connection with a wider fictional universe, so "clanker" it is.
Many Thanks! Ok, vox populi vox Dei ( I guess larger audience counts as vox populi ... )
>I thought the battlestar galactica series established that the _proper_ slur for AI was "toaster" :-)
Agreed. And I thought Dune established the proper term for the anti-AI maximalist position should be be "Butlerian Jihad" and the slogan should be "Thou shalt not make a machine in the likeness of a human mind".
Many Thanks! Yup, I've heard of both those Dune fragments, definitely part of the cultural landscape. Not sure what the successionist equivalents would be ( and would they be in English, Python, or neuralese?)...
I'm still sticking with "Toaster" as well; it more clearly expresses the idea of machines as adversaries of humanity. And, yeah, "Butlerian Jihad" has no equal. Pity the Villeneuve movies weren't able to fit it in.
Many Thanks!
>it smells way too much like the type of knee-jerk reactionary Luddism that's been wrong about everything for hundreds of years,
Not about smoking, leaded petrol, CFCs, and in many ways, not about radioactivity.
Fully automated electronic line-calling in tennis started about 10 years after the sport introduced hybrid human line-calling + player challenging you now see in baseball. Last year the sport completed the transition with only one hold out being the French Open.
Will baseball follow this path for automated balls and strike, maintain hybrid calling in perpetuity, or revert to all human-calling, perhaps with some kind of redefinition of a "strike" that makes it a fully discretionary umpire call? My money is on baseball following tennis. And I think this has lessons for when humans will submit to automated algorithmic rule-enforcement and when they will not.
Humans hate high-stakes results to depend on the quality of a human call for sharp rules. The more a game or match turns on a wrong call that we can see an electronic call should have overridden, the more we will devalue human umpiring of sharp rules, until we get rid of them entirely.
For fuzzy rules, however, like "unsportsmanlike behavior," humans are a little more forgiving of high-stakes results depending on good or bad human calls. Machine-generated overrules of fuzzy rules don't hit the same way as for sharp rules.
I think this will have predictive repercussions about where in the economy, law enforcement, bureaucracy people will fully automate first.
I know next to nothing about professional tennis' motivations, but my take on MLB is that, starting with Selig and continuing under Manfred, the organization has been hyper-focused on improving TV earnings, reducing the length of games, reducing players' salaries, normalizing gambling on baseball / sportsbook, and making games more competitive (or "exciting"). In between all that was the challenge of leaving behind the Steroid Era.
I fully get that MLB wants to make more money—who doesn't?—but the above list is how they're going about it. Automated ball-strike is just one tool among many.
MLB has had instant replay since 2008. My sense is that ABS is the result of 15+ years of frog-boiling to acclimate fans to the idea of reducing the umpires' role in calling a game.
> Humans hate high-stakes results to depend on the quality of a human call
It's true, but (unlike tennis, maybe?) a high percentage of umpire calls in baseball are actually low-stakes. A majority of calls being challenged under ABS today aren't exactly nail-biters, but the batter trying to stretch their at-bat a bit further.
You're right, though, that one or two bad calls out of a hundred can leave a bad taste in the viewer's mouth, and that's the kind of thing people talk about the next day.
My point is, I don't think MLB's decisions are made so much in recognition of what humans want, but what MLB can get fans to accept. Personally, watching a game with the ump replaced by an electronic display behind the plate announcing BALL 2 or STRIKE 1 is like watching someone else play a video game. I'm probably in the minority on this point.
Hell, I think Manfred would sign off on getting rid of the *players* and putting AI robots on the field, if he could. No union, low pay, and zero steroid drama.
99% Invisible did an episode on automated balls and strikes, and they mentioned that MLB *tried* having the umpire just be a frontend for the machine - they wore an earpiece and just repeated whatever the ABS system said. But this turned out to be unpopular because a strict definition of the "strike zone" excludes some pitches that are hittable and that an umpire would have allowed. So they switched to the challenge system so that it better captures the fuzzy area that umpires and players actually think of as the strike zone.
They should just redefine the strike zone to be “hittable pitch,” so it is a fuzzy rule rather than a sharp one. I guess false negatives would be the reason to have ABS.
I only have a vague understanding of baseball, but "hittable" seems like a much broader standard, right? Like, I *could* hit a ball that sails two feet above my head.
I think the problem is that the current official definition has slight but significant mismatches with the as-applied definitions. The official strike zone is a rectangle, while the unofficial one is more rounded with pitches at the top and bottom corners usually called balls while pitches on or just outside the middles of the left and right edges (especially the edge away from where the umpire sets up to watch the pitches) tend to get called strikes. There should be a way to amend the official definition to better match the unofficial one.
Honestly the strike zone should just be baseball-shaped.
Yeah it’s a feature of all fuzzy rules in everywhere in bureaucracy. “Let in only the good students” is a rule that allows broad interpretations of “good.” A sharp rule like “let in only students whose GPA x SAT score is in the 90th percentile” is a sharp rule that can then be automated.
Much like baseball, the fuzzy definition in academia works fine until you suspect the ref is biased, at which point it's nice to have an objective standard you can fall back on.
Way back in Astral Codex Ten Open Comment 314, I had a lively interaction with a good bunch of folks about free will denial. I was posting under my pen name at the time. Now, this is the real me.
Last week, I had an essay published on free will denial in The Missouri Review, July 2026 issue. For those who are interested, I must indulge your forgiveness for taking the liberty of pasting in the beginning paragraphs (but it’s “open comments,” right?). Would love to discuss. It is not academic, but a literary, creative nonfiction, personal essay. Hopefully engaging and easy to read.
SLEEPWALKING TOWARDS BETHLEHEM
On What We Know but Do Not Believe
David J. Frost
The Missouri Review, Volume 49, Number 2, Summer 2026
1. Two years ago, my fiancée told me that I’d been sleepwalking and carrying on full-fledged conversations with her each night. When Kelly first relayed this, her brow twitched slightly with what I took to be suppressed irritation. These mornings, I never remembered anything about whatever it was I’d done overnight. After all, somnambulism is a classic example of automatic behavior. Yet the archetypal image of the sleepwalker—nightcap with drooping tassel, eyes closed, arms outstretched in front, wrists bent down—is wrong. Sleepwalkers engage in complex behavior: they might cook a meal, drive a car, carry on a conversation, murder a relative—all without the accompaniment of consciousness. It raises the question: if I’m not conscious, how am I seemingly able to adapt to changing circumstances? Who was talking when I was conversing? As Kelly and I sat up in bed that first morning, her pent-up prickliness made me acutely aware of the score between us. What was the regrettable con- duct I unfreely enacted each night but for which I must nevertheless accept responsibility? Evidently, I had been peeing in the closet.
2. During the first night, Kelly says, my rustling woke her up, and she asked me what I was doing. “Nothing,” I said. “Going to the bathroom.” But she could see that I was standing in front of the open closet and said, “That’s not the toilet.” “Yes, it is,” I said, sounding offended, while gravely continuing to conduct my business. We would learn later why it’s said not to wake up a sleepwalker. Although it’s like a preprogrammed routine is being run, the person’s capability to react in complex ways can remain, but often only in limited form. I could interact with Kelly, for instance, but likely only to counteract her attempts to thwart me. If you attempt to impede a sleepwalker, they might instinctively find it aversive and—one article suggested—react impulsively, yet they’d be without the benefit of consciousness to rein in their anger. That was the theory, anyway.
3. Luckily for Kelly and me, my nonconscious self was more likely to be cavalier and arrogant than angered by impediment. While follow- ing the going-back-to-bed algorithm, I responded to Kelly’s protestations, she reports, by uttering “an aloof grunt of bemused recognition.” During the night of my second episode, when she again informed me that I was standing at the closet, not the commode, I let go another self- satisfied, haughty grunt. But I apparently persevered in my task, because she could hear the unmistakable evidence—the damp applause—of my misguided micturition. A faux-fancy language affords me distance from disgusting deeds.
4. The third night Kelly screamed at me, “ARE YOU AWAKE?” “Yes,” I said. In that moment, she struggled with what to conclude about my inner mental world. For I was passing the Turing test, so to speak, yet she didn’t know if I was actually conscious. As we found out in the morning, I hadn’t been.
5. When I asked her to recount the events for me once more (in furtherance of writing this essay), I noticed that her brow involuntarily furrowed again. The biggest irritation, she said, was having to clean up the soggy mess after I’d already roboted back to bed.
6. “That sucks,” I replied dutifully. Yet, moments later, I caught my- self feeling uninvolved again, like I hadn’t done it. I knew I had done it; I took what she said for the truth. But I did not feel that belief in my body. Usually when I’ve done something wrong, my self-criticism is more despairing. The knowledge of the injured party’s disapprobation is more—I don’t know—piercing. I get a nauseating knot in my stomach. But this time, my understanding that I was responsible for Kelly’s aggravation was entirely intellectual and bloodless.
7. Because I have so tenuous a connection to what I’ve done while sleep- walking, I find it easy to be glad of my inopportune nocturnal emissions. If I had not sleep-peed in our closet, then I would not possess such a good example of an action for which I take responsibility but which I did not do of my own free will. It’s been a fascinating experience. It’s like when, at the dentist, I cannot keep my tongue from pressing against the hygienist’s elongated mirror.
8. Above all, sleepwalking gave me a concrete path into wondering what it’d be like as an experience to deny free will, as a small number of pub- lic intellectuals have recently done. It was weird enough to think that I had carried on a conversation without free will one, two, or three times. What would it be like to think every action you undertake lacks free will? Not just outbursts but even conscious decisions made after careful deliberation. What would it be like to believe that each thought in one’s stream of consciousness was an effect of previous causes?
9. For Stanford biology professor Robert Sapolsky, for instance, what is the experience of denying his own free will like? Sapolsky is convinced we do not have free will, so from his perspective, he knows it. But he’s admitted he catches himself reverting to the default setting of free will belief every time he nonchalantly accepts responsibility. “Someone says, ‘You’re dressed nice today,’ and I say, ‘Thanks,’ as if I had anything to do with it,” Sapolsky joked during a guest appearance on NPR’s Radiolab.
10. When Sapolsky manages to believe what he knows—when he is actively remembering that he doesn’t have free will—how does his day-to- day self-conception change? Is denying free will bad for you? Or does it increase well-being? I’m talking about psychological effects. I became much less interested in the classic philosophical arguments for and against free will. Released from responsibility, does a free-will denier run amok, his face frozen in a gleeful rictus? Or does he lose all motivation and succumb to fatalism, thinking it doesn’t matter what he does? Would denial, as Sartre worried, serve as an excuse for bad behavior? Perhaps, as Sapolsky imagines, pervasive denial would bring about a fairer, more just, less cruel society, “a giant leap forward in our moral progress,” as he’s said. How would our thinking change if we cleansed our vocabulary of the term “free will”?
11. When Sapolsky and others deny free will, they don’t mean that everything is destined, that what will be will be, no matter what we do. It’s not fatalism necessarily. The precise logical implications of denial are, however, not a matter of what Sapolsky might want them to be; they are a matter of science and logic and will depend on arguments and experimental evidence. And, as it turns out, there is already an empirically informed debate going on.
12. My research into the psychological implications of free-will denial led me to novel places on the intellectual landscape: to a confrontation with bodies of knowledge called “informational existential threats,” to the conceptual affinity between progressive political opinion and free-will skepticism, to the idea of an Area 51 for anti-free-will research, and finally, to the existence of brain-scanning technology that can read the inner contents of your mind and know what you are going to do before you know.
13. When I sat down in my office to start my research, the first thing I noticed about online pictures of Sapolsky was the compassionate twinkle in his eyes. Framed by long, curly gray hair and a gray Garibaldi beard, Sapolsky’s eyes sparkled like quartz in the Tolstoian rocky crag of his sun-wrinkled face. A specialist in neuroendocrinology, he’s spent thirty sunny summers in Africa studying simians. (Sorry! I couldn’t control myself.) Returning to the same troop of baboons each year, he’d ob- serve their behavior, tranquilize them, and take blood samples. Sapolsky found, among many other discoveries, that baboons ranking low in the troop’s dominance hierarchy had chronically elevated levels of stress hormones (glucocorticoids) with concomitant negative effects on health, like hypertension, and on behavior, like an increase in isolated sulking and displacement aggression. Sapolsky estimates that nearly 50 percent of all violence among his troop of baboons is displacement aggression. Shit, as we know, flows downhill. “A high-ranking male loses a fight and so chases a subadult male, who promptly bites a female, who then lunges at an infant,” Sapolsky has written. The more a baboon displaces aggression after losing a fight for rank, Sapolsky says, the lower his glucocorticoid levels are when the good doctor, sans sun hat, tests his blood.
14. That explanation—a diminishment in rank causes stress hormones, which cause the loser to punch down to gain relief—is …..
If folks want the whole essay, it’s behind a paywall at the website of The Missouri Review.
But a PDF is available to free subscribers of my Substack.
Thanks!
I'm a hard determinist, in fact I think that is such an obvious conclusion I reached it in high school and all the philosophy I studied in the meantime didn't change anything. But I don't find this very relevant to daily life, because the epistemic limitations on any individual human or even a large information system are still so tight that the world ends up feeling not fully determined. My metaphysics and my existential world just don't need to intersect that often. It's still a load-bearing belief, it supports my entire worldview about science and reason, and if I thought the world contained genuine random-ness and uncaused effects I'd be extremely bothered. But it nevertheless doesn't play much of a role in mundane daily life or the way I view people, because both me and those people lack the information to predict the next state of the universe and have to rely instead on the sort of behavior-prediction heuristics that everyone does no matter what their own metaphysics are.
Hi Cjw, Thank you for your reply. What you said is very interesting. I can relate to a degree. I am pretty confidently Mr. Determinism. Even when people bring up quantum mechanics I have stuff to say that does not move from basic determinism. Either I say randomness does not get your free will. Or I say QM is indeterminant in precisely those places that physicists say "we don't understand." What I mean is "understanding" is synonymous with determinism because you cannot understand something if when you do P sometimes it does X and other times it does Y. Feynman will say that if anyone tells you they understand QM then they do not understand QM. So QM is a theory even if highly confirmed which we cannot understand. When people ask, What is going on that something can be indeterminately in a superposition of both X and not X? physicists have to say "It just is. About why, we can feign no hypothesis." Where they are quoting Newton. If we wanted to feign a hypothesis an example would be the hidden variable approach (I think).
Anyway, if the universe is fundamentally indeterminant, then there's a problem for free will there too. In fact free will skeptics sometimes call themselves hard incompatibilists rather than hard determinists. Because whether the universe is deterministic or indeterministic it'd bad news for free will (intuitively speaking, meaning one should define free will or be clear that free will has two definitions the LIB and the COMP).
As far as hard determinism not making a difference to real life,... I am somewhat sympathetic to line that of thinking too. To paraphrase you in my own way of thinking, metaphysically or ontologically everything is determined. Rewind the universe, press play, if you do not change any antecedent conditions then after "play" you get the same events again. BUT epistemologically we cannot do that rewinding. And yes everything being determined means its predetermined if only we were a Laplacean demon. But we are not Laplacean demons. The laws we do now are not exceptionless. And we rarely have sufficient detailed data about initial conditions to plug into laws, exceptionless or not.
However, we are getting closer and closer. There are some things we are learning which are still not like a Laplacean demon's level of detailed knowledge or exceptionless prediction.. but we know that chronically elevated levels of glucocorticoids retard brain development in kids. We know how serotonin and the other neurotransmitters work. And we are getting closer to being able to predict what someone will do based on this knowledge that we have. But it's not there yet.
However I can imagine we are getting closer. But then we are back to not real world stuff. Let me try this. Dennett says the criminal is responsible because he intended to thieve. The hard determinist says (as Sapolsky often does and Sam Harris does too in his way) BUT WHAT CAUSED THAT INTENTION? Then a progression can be imagined. If Dennett can characterize the cause of the intention in a way that suggests the criminal is responsible, then Sapolsky will uncover the always available cause of that cause, let’s call it N-1, (with apologies to the lit journal N+1). Then Dennett can either characterize that cause in a way that suggests responsibility or else he can’t and the cause will characterize itself as responsibility-undermining.
Let me add to this N-1 progression, that over time as scientific knowledge progresses it gets harder and harder for Dennett-types to cite a cause that suggests responsibility and more and more likely to make it all seem unfree.
So while we don't know now, we are knowing more and more and what we know I think can be used to predict. I should give an example but I've gone on long enough. I will work on saying all this more clearly by having a concrete case.
RE: the moral responsibility for crime. Whether or not the criminal was "free" to choose otherwise is irrelevant for the logic of punishment. I'll stipulate that a criminal is a deterministic optimizer that has no choice but to respond self-interestedly to the world around it. Punishment is then justified on the basis that it shifts the optimizers' incentives by altering the payoff matrix of the options available to it. Calling the punishment "morally deserved" vs "game theoretically optimal" is just semantics.
Sapolsky's argument is actually self-defeating because he's arguing that deterministic brains can't be influenced by rational control, even though the fact that he's making an argument at all presumes that he's influencing the behavior of the people who hear the argument!
I like your point. I am not sure Sapolsky says what you say he says.
What I think you are talking about is what popped up in the reading I did as "deterrence". The folks I read (like Sam Harris, Gregg Caruso) tended to say that society could continue with policies of punishment aimed at deterrence, even if criminals would not be considered fundamentally morally responsible. It's a sound philosophical position but one which may not make sense in the real world. Or so it is beginning to seem to me. The phil position is no one is ultimately morally responsible (but they may be proximately morally responsible) (which needs clarification I know.) And the position additionally says deterrence can still make sense (Harris says this for sure, Sapolsky maybe doesn't, I cannot recall Caruso on this) precisely because Harris's hard determinism countenances that people are still responding to incentives. Harris says for critics to say he thinks people are not responsive to incentives is like the critics who say his hard determinism is basically fatalism, which he says it's not. On good grounds I believe. Although I am not reproducing those grounds here.
I guess what I need to work on is an ability to explain the diff between ultimate and non-ultimate/proximate moral responsibility. And how it is that Harris thinks people still act "voluntarily" within his hard determinism. This would have to do with his "co-opting" of COMP's definition of free will and using it for his definition of "voluntary" which I alluded to at the end of my essay in the Missouri Review but cut for space and also because it was too hard to explain to non philosophers. I mean I tried and I could feel I was losing my audience. Most people won't dig through or wade through writing like this here in ACT. I have been trying to write--not here but in published essays--for the general reader.
I keep saying "Harris" instead of "Sapolsky" because Sapolsky did not really engage with the philosophical literature. Or not like a philosopher would. Which makes sense as he is not trained that way. Harris is better suited to my quest to find someone saying something non-ambiguous on these points. But even Harris is lacking in detail for all the stuff on free will he's produced.
I will say this. Usually I see people say that free will deniers-- be they Sapolsky or Harris -- have a self defeating position or argument because why are they bothering to write a polemic and convince us of their view if we are not free to choose to believe it or not believe it. And to this I say--and Sam Harris has also said--that is not how believing works. (You have not made this mistake, to be clear.) Believing is not a matter of choosing the proposition and now you believe it. As Nagel said "All belief should be involuntary." My undergrads balk at that but I think Nagel is right. Reasons compel the mind. Evidence compels the mind. The "should" in there is important too of course. Lots of people do seem to "choose" their beliefs but that is not a practice that is likely to arrive at true beliefs.
So that is the nature of philosophical discussion or really maybe any attempt at communication. You accuse Sapolsky (or let me say hard determinism like Sam Harris') of being self-defeating because HD seems to hold that people are not reasons responsive. And another critic says it's self defeating but for the opposite reason. This other critic says HD is self defeating because it conflicts with his stated purpose of persuasion which is kinda what you said but how persuasion works this other critic has wrong. You have it right it seems to me at least relative to this other critic. My issue with your position is I do not think Harris nor Sapolsky nor hard determinists in general are committed to the idea that people are not reasons responsive, or not driven by incentives. Merely because people lack libertarian free will does not mean their behavior cannot be influenced by incentives.
I feel like I may have stated that without a lot of argument in its favor. But that is not entirely worthless. ha!
Harris has parried the fatalism claim and I think the same moves of his there could be used here, to show people are still responsive to incentives.
Also bear in mind - and I don’t think I’ve seen this argument before - that if you’re a hard determinist then the people doing the punishing are also unable to do anything else. They’re just responding to incentives and bear no more moral responsibility for e.g. capital punishment than the murderer they’re executing does. Whether you call it self-reactive deterministic social dynamics or justified moral opprobrium is immaterial. The behavior is isomorphic and so the distinction is meaningless. This is why I find the dialogue on this little more than empty semantics (as, of course, most philosophy is).
I was a prosecuting attorney for 15 years and that never bothered me either because I didn't really care about the intentionality at a metaphysical level, only the legal one of mens rea. You put a thief in prison to incapacitate him and end the thefts, and because if you don't have a system that does that you'll get people taking matters into their own hands more sloppily. You also do it to add costs to crime and to mark a behavior as beneath a certain sort of person, so the criminal class is disincentivized and the typical pro-social normie views it as outside the scope of their options.
I grant that this sets up a situation where people who "aren't responsible" may be punished, but that particular type of responsibility isn't really needed. If I had been punishing innocent people (actually pretty rare if you look into it, certainly in this century) that sort of error would undermine confidence in the system and would harm its purpose. That's why you cannot extrapolate from the utility functions I mentioned earlier to a system where you instrumentally incarcerated people just to send good signals.
The law actually provides a sort of dividing line here, via the principles of both mens rea and actus reus. Not only do you require the necessary criminal intent, the actus reus distinguishes being involuntary and voluntary actions but does so along the lines of whether the act is the sort of thing we associate with deliberation. A reflexive jerking of the arm from a neurological disorder strikes a man and knocks him over the wall into a river, no actus reus. From my point of view, one might say that's not so different than a situation where one had some bizarre impulsive reaction to the look on that man's face and pushed him over, due to a nearly unfathomable chain of behavioral stimuli accumulated over the pushers' lifetime. Or even further, perhaps he takes a moment to deliberate first, perhaps he takes 10 minutes and watches and plans out knocking that man over right as he walks past the perfect spot because he looks so irresistably pushable for some reason, we're climbing the mens rea ladder into higher levels of criminal culpability. Perhaps as some say, the deliberation narrative is just a story we tell ourselves about something that already happened a few fired neurons ago. No matter to me, people seem almost universally to understand and draw a line between reflexive movements they deem "involuntary" and those they think "voluntary" and everyone seems to understand the distinction and the system works, and the sort of person who pushes a man off the bridge under what seems to us to be a "voluntary" decision-making process ends up being the type of person who needs to be incapacitated and removed from society no matter your stance on free will. So it all works out.
All that adds up to compatibilist free will.
Yes it does and? Yes it does and that's all to the good if you ask me. Well done Cjw. Thanks!
It did, and you can't accurately describe yourself as a free will sceptic if you effectively believe in compatibilidm.
I really like this comment and appreciate the admirable clarity of your writing. I have not been able to be as clear here in these comments. And in my Missouri Review essay (which was clear but limited in scope) I was more or less prohibited from dealing with certain issues or ideas because my exposure to "the general public" suggests they would not sit through either my attempts to explain these ideas or because people's eyes would glaze over or whatever.
In any case I need to continue to try to write clearly, more clearly, even in internet comments. Because apparently it can be done! (Please let me know immediately, if you have any commenters here who in their response to you demonstrate that they failed to understand what you meant. :-))
That being said I take again another opportunity to think out loud.
You wrote: "Perhaps as some say, the deliberation narrative is just a story we tell ourselves about something that already happened a few fired neurons ago. No matter to me, people seem almost universally to understand and draw a line between reflexive movements they deem "involuntary" and those they think "voluntary" and everyone seems to understand the distinction and the system works, and the sort of person who pushes a man off the bridge under what seems to us to be a "voluntary" decision-making process ends up being the type of person who needs to be incapacitated and removed from society no matter your stance on free will. "
Yes. That and probably almost everything if not everything you said before gets a yes from me. In the terms of the literature I read, this is what Caruso calls quarantine and crucially he lays out his philosophy of crime as a public health issue requiring quarantine of very specific sorts ALL within the context of denying (ultimate) moral responsibility (usually b/c of notions about determinism and libertarian free will.)
So. Caruso says what you said more or less. He calls it quarantine and he says it will be justified in some cases even when there is no ultimate moral responsibility.
I don't know if staccato definitions makes for clarity but...
Denial of libertarian (LIB) free will implies no ultimate moral responsibility, but it does not imply no proximate responsibility. Proximate MR is compatible with denial of LIB, compatible with the truth of determinism.
LIB free will says you could have done otherwise than you did where the otherwise is in this absolute sense of... if we rewind the universe and let it go a LIB free person could choose soup this time when they had salad before rewinding... and, crucially, they could choose soup without changing any antecedent conditions, so, while still wanting salad for example. That is a very rarified notion of "otherwise" which always chapped Dennett's hide or rubbed him the wrong way. The more common uses of otherwise countenances that we will be changing some antecedent conditions before we press play. But the rarified use of "otherwise" is appropriate here because we are defining the absolutist philosophical position. Maybe no one will end up holding that position. But we need a name for it. And turns out lots of people will aver to having such an ability perhaps because they think the human mind is magic or their thoughts around here are not cogent or they think something logically impossible but apparently not psychologically impossible.
What is the nature of proximate responsibility? I am working on that. Whatever it is, there is an isomorphism. Deny LIB means no ultimate MR. Deny COMP means no proximate MR. But now I have to get in my car so I will have to stop typing. Hmm. I'll just add that this business is relevant to Sam Harris's sometimes misunderstood claim that he wishes to "co-opt" the COMP definition of FW that his friend Dan Dennett came up with, and use it not as a definition of the word "free will" but instead as the definition of the word and related cognates for "voluntary action" or "action for which we will be held proximately morally responsible."
Harris thinks the phrase "free will" (FW) means to most people LIB. Be that as it may, it means to some people at least the LIB notion. Harris sees Dennett therefore as offering a new definition of FW. Harris basically says although he does not use these terms, "No let's let FW continue to. mean LIB, let's just get more and more people to realize we do not have LIB FW. Let's do this conceptual revision like happened for the concept "witch" and unlike what happened for "whale." See the end of my Missouri Review essay (available for free even though paywalled in some places) for how that goes. Whale was defined as a fish at one time. Witches have the same definition they did in the 1600s, magical satanic women or whatever. It's just now we know there are no real entities that meet that definition.
How did you come to know that free the universe is deterministic?
Why do you reject compatibilist free will?
So you don’t believe in random numbers, as opposed to pseudo-random numbers? There seem to be a lot of stochastic phenomena from which we can derive various useful probability distributions.
I might agree, depending on what you mean by "hard determinism". Let's say you're watching a particle, and say to yourself, "if it decays in the next T seconds, then I will have a ham sandwich for lunch, otherwise a burger". Your choice of lunch meat that day will be determined by a physical event (or lack thereof), but that event is essentially random, and thus your decision is as well.
I don't believe that is random either, the Copenhagen interpretation and similar nondeterministic interpretations of quantum mechanics fail the basic rules of logic and therefore cannot be science either. You're only unable to determine it due to an epistemic limitation. (Of course for some events the epistemic limitation may be insurmountable, or require nearly every single piece of data in the universe to be known and processed such that the prediction-engine is functionally equivalent to the actual universe.) If it decays in the next T seconds, it was always going to decay in the next T seconds and I was always going to have decided to tie my lunch choice to that outcome and always going to eat a ham sandwich.
Thankfully for me, my choice of what measurements to take concerning quantum level events was also determined.
> the Copenhagen interpretation and similar nondeterministic interpretations of quantum mechanics fail the basic rules of logic and therefore cannot be science either.
I am no quantum physicist so you might be right; however, your reasoning is IMO flawed. Logic isn't some divinely revealed immutable rule; it's just a model we've built to explain the world. It's a pretty good model, but like all models it's incomplete. The notions of "cause and effect" are no qualitatively different from the notions of "up and down": useful in most cases, incorrect in some others.
I think where I'd disagree is that scientific exploration depends on them. If you run an experiment the result of which is that you conclude the rules of logic and causality and noncontradiction don't operate, you've undermined the premise by which you conducted the investigation.
If you don't think cause and effect work you can't really do any science.
I won't pretend to understand the math of quantum physics, I had the typical 2 semesters of normal calc-based physics that only barely touches on it at all and couldn't remember a shred, but I do remember the various alternatives to the Copenhagen interpretation. One of them is called superdeterminism, and is often rejected because it fully denies free will and supposes that e.g. Bob measuring the spin of a particle instead of the energy level at this point in time is fully determined. But since I already believed all of those things that were described as the hurdles to accepting it, I had no problem supposing that was the most likely explanation.
> If you run an experiment the result of which is that you conclude the rules of logic and causality and noncontradiction don't operate, you've undermined the premise by which you conducted the investigation.
But IMO this has already happened, and more than once. Radioactive decay, virtual particles, arguably black holes (and the primordial singularity) -- all those things can be modeled quite well in ways that violate the classical rules of "cause and effect". I understand that you disagree, but the very fact that you and I are *able* to disagree on this undermines your view on logic as being immutable.
>If you don't think cause and effect work you can't really do any science.
Cause and effect isn't a synonym for determinism.
Much science is statistical. See the endless discussions of p hacking and p values here, for instance.
Fundamental science, like particle physics, is statistical, too.
(Why would a sizeable proportion of physicists believe in CI if it's straightforwardly illogical?)
So science doesn't rely on determinism.
That doesn't mean science had abandoned any notion of causation. Necessary and probabilistic causation are still available.
Even if there were a case for determinism, it would only refute libertarian, not compatiblist free will.
"Cause and effect" isn't part of standard logic , anyway.
May I ask, "And? What follows from that?" I mean, I think I understand the claim that logic is XYZ. But can you expand on what you think that implies for the larger discussion here? Like is it a general comment of a nature like "who's to say? we'll never be able to know" etc Or did you offer it with respect to a more specific point? If my question does not make sense feel free to disregard. Thanks
I am merely attempting to refute the argument that goes something like this:
1). [Propositional] logic could not function at all if the rules of causality did not work.
2). But logic does function at all times in all situations.
3). Therefore the rules of causality also function at all times in all situations.
Causation isn't built into logic. In fact, modal logic, the logic of possibilities , is a thing.l, as usual probabilist logic.
And, in fact, there are multiple human made logics , nit a single gif given one.
Cool. I enjoyed these essays by Gerard 't Hooft https://arxiv.org/pdf/quant-ph/0701097 and https://arxiv.org/pdf/1709.02874 in which if I understood... one thing he says is that some interpretations of quantum indeterminacy are based on an erroneous tacit assumption of libertarian free will. Like the business about how consciousness affects the way a quantum event happens (or that it happens or not) assumed LIB and that is where we get the idea that it cannot be known. He says if we could know what the human was going to do then we could know how the universe would react. Obviously it is more complicated than that. And I know I do not totally understand. But it was super interesting. Both articles.
You wrote: "... the Copenhagen interpretation and similar nondeterministic interpretations of quantum mechanics fail the basic rules of logic and therefore cannot be science either. " I agree with that. Some philosopher think we don't have LIB, others think LIB is a paradoxical or nonsensical idea. The causa sui is LIB (speaking roughly and loosely now you nitpickers out there) and the causa sui is as Nietzsche said "a perverse kind of logical rape" and "yet mankind has got its pride all entangled up with the causa sui" (paraphrasing).
Feynman said, I believe, that if someone says they understand QM then they do not understand QM. Because QM is not understandable. Where understandable means make sense etc. I take this to mean that those moments in which QM says something is non-deterministic it is not really understandable. Why might that be? It seems obvious to me but maybe I need to work on laying out why.... If some sequence of events is non deterministic and you try to rely on it in an explanation, you will not feel as if the explanandum has been explained by the explanans precisely because you need things to remain fixed so to speak so that they can play their function in explanation. If you say, "XYZ happened because P happened" are you not saying in some sense that every time P happens XYZ happens? (within a certain context of causal talk--whether INUS or whatever)...
In a probabilistic context (not QM probabilistic but just social science level statistical probabilistic) if you explain something by appeal to a generalization, that generalization has to have certain features or it won't help explain.
What you "say" to yourself has a causal chain behind it, don't forget. So the attitude you decide to take to this random event generator takes us back into deterministic causal chain, if you follow me. Aside from that I think I follow you that the decay is random. And if we could insert somehow a random event into the chain of events leading to my action then my action would be random. I make that point in my Missouri Review essay. So we are (or might be) in agreement at least that far. Thanks for commenting within my thread.
Did you read Ayn Rand, by any chance?
> o the conceptual affinity between progressive political opinion and free-will skepticism,
I’ve never fully understood that. The right should be more inclined to free will skepticism. Criminals being born, and not made etc.
Hello Peter Defeel, Thank you so much for reading and replying. That business about progressives and free will was one of the parts in the essay that I added in more to start a conversation than to take a side per se. As far as I was asserting anything...I would clarify this way: An earlier draft of the essay had a whole section where I explored personality types and political personality types (ie what score on the OCEAN big five personality test correlates with conservativism and what with left liberalism). I also did some research on what empirical data there might be as to a correlation between personality and free will credence or skepticism. I never wrote up a summary or my take on the empirical work because its methods suggested to me that its results were inconclusive and would stymie discussion more than help it. Be that as it may. I could be wrong there.
But I did write up a section where I laid out the affinity between progressive political opinion and modes of thinking as compared to free will denial. Did the essay mention obesity? Or did I cut that out ? Did I leave in the Ronald Reagan quote? No, I recall now I had to cut that out. But Reagan said something to the effect of “Every time a crime is committed we need to stop saying society has failed and instead say the criminal has failed and is to blame.” And then he went on about personal responsibility, which USED to be a conservative’s favorite word or one of many. I think that it is a pretty standard leftist view that when a crime is committed IT IS THE FAULT of society and not the individual. Thus, the affinity with free will skepticism. Provided we take “structural and systemic” causes to be deterministic or at least playing the deterministic role in the explanation relieving the individual of responsibility. Indeed the left would say that the factors of criminogenesis are trauma and abuse, unequal opportunities, untreated mental health conditions, toxic environment, etc. These are the factors that philosophy professor Gregg Caruso lays out, conspicuous in its absence is anything that would suggest the criminal is responsible. I think that looks left liberal/progressive. And Caruso is a hard determinist. I was prepared however for left liberals to read that stuff and reject it. It was a sort of trap I was laying. Because I had personal experience with a left liberal woman …. Well, this woman knew Sapolsky. She was a grad student in psychology and she said “oh Sapolsky is so nice, you are going to love interviewing him.” But then I told her about my passage in my essay where I talked about Sapolsky (and in the published version I add in Eagleman)—about how crime like theft might be caused not by penury (taking what’s not your because you lack what you want or need) but would instead be caused by the stuff about how stress-caused glucocorticoids cause changes to the developing brain such that kids stressed out by poverty have measurable deficits in impulse control… Well, to that she said, “That’s a pretty convenient explanation for a white male: blame the dark skinned person’s brain rather than structural racism. And it sounds too deterministic.”
I replied that I don’t know what a convenient explanation is—either some reference to a cause explains or it doesn’t. Whether it is convenient does not touch whether it explains or doesn’t. And as far as deterministic. All causes are deterministic. All explanations are deterministic, otherwise they don’t explain. (But that is a point I feel myself short of being able to explain let alone defend right now). Upon reflection, her response seems to me like a very typical response and one that I would say counts more as a symptom of something than as real thinking. After all, Sapolksy was saying not that it was the poor person’s brain full stop, but that SOCIETAL or structural inequalities were to blame for the differences in the poor person’s blame.
Just what was triggering her progressive intuitions I can only speculate. But that won’t stop me. It could be a knee jerk reaction to socio-economic explanations which lack any reference to race. There has been a debate on the left in which those two kinds of things are pitted against each other. And upon reflection I should be taken aback. She had just finished saying how nice Sapolsky was and then she felt no compunction against attributing his scientific theory to his race and gender with the implication that it was therefore devoid of merit. So that was my thinking. Put it out there and see if anyone in the public sphere wanted to agree or disagree that free will skepticism has a greater affinity with the left or the right politically speaking. As for what you said specifically… You said to paraphrase in my own words: the right should like free will denial because they think criminals are born not made. Hmm. I don’t know if they think criminals are “born” as much as they want to say that criminals need to be held accountable because they are responsible, they had a chance to choose otherwise. So rather than “born” they would say that a criminal has made himself into a criminal with his choices. And those choices are choices he is accountable for — all of which assumes free will.
>All causes are deterministic.
that's not true empirically , since physics hasn't concluded that the universe is deterministic
I also isn't true semantically -- you can describe smoking as a cause of cancer while.knowing that it doesn't invariably lead to cancer.
>I replied that I don’t know what a convenient explanation is—either some reference to a cause explains or it doesn’t
In the world I live in, there are partial explanations, competing explanations, correlations masquerading as causation , etc.
>I replied that I don’t know what a convenient explanation is
One that's chosen , not forced on you by its undeniable correctness.
>might be caused by penury ... but would instead be caused
Might ..or might not. You are dealing with rival explanations neither of which is not I known for sure to be.true or false. Under those circumstances, a choice to back one or the other could be motivated.
>All causes are deterministic was something I said in conversation to someone who tried to say that my or Sapolsky's explanation of poverty-caused-stress-causing-glucocorticoids-which-stunts-brain-development was too deterministic
But it isn't true, if you are using the term "determinist" strictly, and doesn't answer her concern. If you had said, that it's only probabilistic, that would have answered her concern and said something true. The claim that everything is too extreme, it means social psychology wouldn't work at all, so what is needed is a happy medium ...influence without strict determinism...is something you could have said
>I suppose I mean an explanation has to be "rigid" or whatever she was worried about or it does not explain. Or maybe I meant just the opposite.
Don't you know which?
>" I suppose I mean an explanation has to be "rigid" or whatever she was worried about or it does not explain.
It has to be something. It's not clear that the something has to deterministic. In psychology and sociology , nothing is 100% certain, and researchers only look for some level of correlation.
Psychologists look for p values which are continuous variables , not a binary "it either explains or it doesn't". Even the conceptual aspect of explanation isnt a binary.
. If you stick to the high bar , that only determinism qualifies as explanation , you have to bite the bullet that psychology and sociology don't contain any explanations. Or you could adopt the low bar, that explanations only probabilistically constrain expectations. Then you can maintain that sociology and psychology do contain explanations , but not deterministic causes.
>Behind these less than perfect generalizations there are real deterministic events.
How do you know? Physics is my subject. Physics does not assert determinism. The question of determinism versus indeterminism is open.
It's true that macro level causation has to bubble up from micro level causation ...that getting cancer from smoking is based on millions of molecular events. But that's reductionism , not determinism. The molecular events don't have to be determined.
.
Some anti-free-willers acknowledge this by arguing the incompatibility of free will with both determinism and indeterminism .
>But back to free will. Sapolsky's explanation is not the kind of explanation that makes he and I doubt libertarian free will
So why does he make them? He wrote a long book full of many hungry judge type examples, where you have all these statistical correlations, none of which is individually an example of strict determinism. Presumably he thinks he can make his case by summing a lot of imperfect examples....since that's what he does. And a lot of his readers agree with him.
(Or he's using "determinism" loosely. But that's a bad idea, because loose determinism is compatible with loose free will).
>When I said "I don't know what a convenient explanation is" and you reply with a definition, what are you getting at
I do know what a convenient explanation is. You have the necessary ingredients to figure it out yourself, as well.
If you accept that explanations aren't indubitable , or certain,then it's possible for biases to affect a choice of explanation. So a convenient explanation is a not-necessarily-true explanation that' you prefer to others because it's in line with your biases.
I don't know or care whether that's what is going on the individual case about glucocorticoids.
But you can't get off the hook by saying that some psychological explanation is 100% certain, with no room for personal bias to creep in, because none is.
The thing is that your insistence on determinism is causing problems. It's forcing you to conclude there are no explanations in psychology. It's preventing you from understanding what a convenient explanation is.
> Both of those will have lots of exceptions but those exceptions should not really be considered counter examples to the claim
They should not be considered to as rendering the claim entirely false, and they should not be ignored or swept under the carpet.
Fortunately , there is a happy compromise called probabilistic reasoning. Under probabilistic reasoning, the successful examples raise the probability, and the counterexamples lower it. There's even a formula for it.
The whole Enlightenment project - from Spinoza on - was inspired by fresh Cartesian rationalism and revived atomic materialism. The big point is that Nature laws rule, not some kings. I recommend this thick book:
https://www.amazon.com/Enlightenment-that-Failed-Revolution-Democratic/dp/0198738404
Re "11. When Sapolsky and others deny free will, they don’t mean that everything is destined, that what will be will be, no matter what we do": My understanding of Sapolsky's position is that he denies libertarian free will, which is what almost everyone means by the term "free will". His hard determinist position largely means that he believes that "what will be will be" and that "what we do" is also part of "what will be".
You seem to be using a compatibilist notion of free will that means something like "conscious reasoning". I think the real problem is that a lot of people who subscribe to a materialist worldview seem to be uncomfortable with the idea that libertarian free will does not fit well with that worldview, and so they take something that is not free will and use the term "free will" to refer to it so that they can pretend that they have free will.
Instead of using the phrase "free will" to refer to conscious reasoning, use the phrase "conscious reasoning". When you're sleepwalking, you might not have conscious reasoning; when you're awake, you might have conscious reasoning; and in neither case, at least from a hard determinist perspective, do you have free will.
OK. I gotta ask. What is libertarian free will?
AI's summary: Libertarian free will is the philosophical belief that human choices are not predetermined by prior causes or physical laws (determinism).
In my essay I said, I believe, that LIB free will is the same definition that Hard Determinists share... an action is unfree if it is caused, which by reverse engineering means a free action must be uncaused. And that is what the "causa sui" is. It is what LIB erroneously says causes my free actions. It is a thing capable of causing (like causing my arm to move) but is itself not caused, because if it were caused it would be unfree. So, uh, LIB free will says you are unfree if caused and can be free if uncaused, whereas Compatibilism comes in and says everything is caused that doesn't make everything unfree; to be free an action need not be uncaused; to be free an action needs to be caused in the right way.
> Hard Determinists share... an action is unfree if it is caused
...if it is causally *determined*, which is being caused in a maximal sense.
The word "causation" has multiple meanings. For instance, there are differences between sufficient and necessary causation; and between probabilistic and deterministic causation. Causal determinism is itself often confused with similar concepts such as fatalism and predictability. Causation has a complex relationship with explanation. (In addition to all that, "causality" has a specific meaning in physics, relating to the lack of faster-than-light signalling)
In brief: an effect *cannot* happen without a necessary cause, but *might not* not happen with a merely necessary cause. The instability i a radioactive nucleus is necessary to cause it's decay...stable nucleus dunt decay ...but not sufficient to cause it's decay at a particular.
A *sufficient* cause means that it's effect *must* happen: yet another sufficient cause could have brought about the effect... if the cause was sufficient but unnecessary. A fatal shooting or a fatal stabbing are sufficient to cause death, but neither is necessary
A *probaliistic* cause makes it's effect more likely, but not inevitable ... as in "smoking causes cancer". A probabilistic cause is an influence -- a mere influence. The existence of influences doesn't imply determinism. Some things that are described as "determinism" , such as "genetic determinism" are merely probslistic as well.
Only sufficient causes equate to determinism: probalistic and necessary causes allow "leeway" or "elbow room". An action can therefore be caused in several senses that are compatible with libertarian free will.
>which by reverse engineering means a free action must be uncaused
Undetermined. Which allows it to be caused in various senses looser than determinism. In particular ,it could still be *influenced*. Arguments from influence , such as the hungry judge argument , don't refute LFW. When you get the quantfiers right.
>It is a thing capable of causing (like causing my arm to move) but is itself not caused, because if it were caused it would be unfree
Lack of complete determinism isn't a thing. Any not-fully-determined event is not-fully-caused , yet not a persistent entity, nor a cause of itself. There are agent causal and event causal versions of LFW. You are only addressing agent causation. Since indeterministic events are scientifically respectable nowadays is, event causation can be used to spin up a naturalistic theory of LFW, without souls, downwards causation , or causa sui.
That's the best damn summary of causation and determinism I've ever read. I'm bookmarking this one! Thank you!
My problem with free-will vs no-free-will arguments is that there's no way to empirically validate either position. Neither hypothesis is falsifiable outside a framework of logical argument. We might as well be arguing about the existence/non-existence of a creator being.
Determinism is verifiable.
If you get the same result every time for a given experiment, that's evidence of determinism. It's not verified, but it's verifiable.
Sufficiently detailed models of naturalistic libertarian free will can be tested, because they predict brain mechanisms and processes you can test for. It's not verief, but it's verifiable.
"That's the best damn summary of causation and determinism I've ever read." Yes. I wonder if you like this one?
https://davidjfrost.com/causality/
:-)
Here is something I just wrote to a friend via email who asked about the definition of free will:
There are two definitions in the debate. The so called Libertarian definition is the one that makes the problem seem like a problem because that definition makes FW incompatible with determinism or with how the universe works if we think accurately about how the world works. (If quantum indeterminacy is the truth of how the universe works that introduces randomness not freedom and so is also incompatible with an intuitive sense of free will.)
If the LIB definition is what gets the problem going, the so-called compatibilist definition is offered by philosophers trying to solve the problem.
What does each hold? LIB says that to be free an action must be uncaused, because any action that is caused is unfree. People can be shown to believe we actually have this ability via philosophers questioning them like Socrates and putting situations to them and asking them if the person has free will in that case. For instance, assume you chose soup today for lunch. Imagine we rewind the universe and press play…without changing any antecedent conditions could you choose otherwise than soup? If the person says yes they are using the LIB definition. The LIB definition of FW is “an ability to have done otherwise then we actually did do.” And without changing any antecedent conditions means, everything is the same. Determinism holds that if everything is the same after rewinding and we don’t change any conditions, then you will have soup again. But the LIB person says we could choose salad this replay even without changing anything ie even without changing things such that you want salad. LIB says you could choose salad even while continuing to want soup, or even while all the causes that are germane would normally cause you to choose soup. That shows that LIB FW means an ability to cause things but this cause is itself uncaused. That is the causa Sui from my essay. It is an self caused or uncaused causer,
Okay so that’s LIB. Whether it makes sense or not, it is what lots of people aver when pressed.
LIB, again says, to be free an action must be uncaused. OOMP says, no every action is caused, that does not make every action unfree. To be free an action need not be uncaused, it needs to be caused in the right way or by the right kind of cause. So if an action is caused by a pistol to my head that’d be unfree. But if the action is caused by a desire of mine that I identify with and do not lament, then that’d be a free action, according to this COMP kind of definition. COMP-ists believe in determinism, and they also deny LIB is true. So they say you could not have done otherwise than you did (not in the ultimate sense) but what matters is not that you could have done otherwise but what was the source of your action.
And philosophers have shown pretty well I think that all the various ways that nonphilosophers try to solve or understand the free will problem fall into this schema. Either they are offering or using the LIB definition or the COMP one.
> (If quantum indeterminacy is the truth of how the universe works that introduces randomness not freedom and so is also incompatible with an intuitive sense of free will.)
Randomness is indeterminism, indeterminism. is freedom from determinism , freedom in the libertarian sense.
The objection to randomness based LFW is.much weaker than usually assumed.
Randomness is often held up a something obviously incompatible with will. But indeterminism is not an objection to FW in itself: it needs to unpacked into a series of objections to specific features of a kind of free will "worth wanting" -- purposiveness, rationality, control and ownership. These objections can be answered individually.
(Explaining naturalist libertarian free will in terms of "randomness" is creats a communication problem, because the word has connotations of purposelessness , meaninglessness, and so on. But these are only connotations, not strict implications. "Not deteminism" doesn't imply lack of reason , purpose , or control. It doesn't have to separate your actions from your beliefs and values. Therefore,I prefer the term "indeterminism" over the term "randomness".)
So, how to explain that indeterminism does not undermine other features of a kind free will "worth wanting": there are concerns about ones decisions being rational, being connected with ones desires, and ones actions being appropriate to ones decisions.
Part of the answer is to note that mixtures of indeterminism and determinism are possible, so that libertarian free will is not just pure randomness, where any action is equally likely.
Another part is proposing a mechanism , with indeterminism occurring at different places and times, rather than being slathered evenly over neural activity. In two-stage theories, such as those of James and Doyle, the option-generating stage is relatively indeteministic, and the option-executing stage is relatively deteministic.
Another part is noting that control doesn't have to
mean predetermination -- it can also mean post-selection of gatekeeping.
Another part is that notice that an undetermined choice between things you wish to do cannot leave you doing something you do not wish to do, something unconnected to your desires and beliefs.
>And philosophers have shown pretty well I think that all the various ways that nonphilosophers try to solve or understand the free will problem fall into this schema. Either they are offering or using the LIB definition or the COMP one.
Libertarianism importantly subdivides into a demand for omnipotence, or complete freedom from influence, and a more modest kind. It also subdivides between supernatural and natural mechanisms.
> And philosophers have shown pretty well I think that all the various ways that nonphilosophers try to solve or understand the free will problem fall into this schema. Either they are offering or using the LIB definition or the COMP one.
That's a pretty grand statement. I suspect this is a false dichotomy arising from premises imposed by the logic you're using. Certainly, Mahayana philosophy doesn't see a simple dichotomy between free will and determinism. But then again, they use a paraconsistent logic with four possible states. And I suspect — suspect — that makes their interpretation more nuanced than Western either-or structures. I need to think about that some more...
I am working in--and my essay announces itself as existing in--the framework of western analytic philosophy's take on what it calls the philosophical problem of free will and determinism. Mahayana is probably wiser. But bringing the two together is not something I'm signing up for at the moment.
> "If quantum indeterminacy is the truth of how the universe works that introduces randomness not freedom and so is also incompatible with an intuitive sense of free will."
But the human brain may have selective power which quantum fluctuations fit the intention or will (as material basis for possible actions), and clutch onto the "anticipated" fluctuations and let them rule the dynamics. Would that not be a compatibilist version of free will, where you got to enjoy your "anticipatory" neural biases just as much as your organic endowments?
To hard determinists I have this question: In what (causal?) ways do physical processes "determine" or "dictate" the biology and our thinking? Say, an adaptation happens - which implies a novel physical basis for your behaviour. But given that physical basis, who can "determine" from the first principles what is the macro-scale effect? This problem (pattern recognition?) could be undecidable per Godel or Turing. Emergent complexity is organized by intricate macro-scale processes and emergent centers of control, but all the credit must go to atoms moving in the void?!
And by the way, even classical thermodynamics is not "determined" by atoms, since statistical postulates are employed as additional assumptions. Hence we have Boltzmann's Stosszahlansatz and Loschmidt's paradox. The reduction logic is not in logical deduction, but statistical approximation or asymptotics.
Hi David, Thanks for your reply. I’m not sure I followed as you will see below. But here’s my reply for what it’s worth.
“What will be will be.” Let’s call that fatalism. Or call it shmatalism. I say that to keep myself honest as it were. It shouldn’t become demerited in virtue of what we name it. So, let’s ask, what does it *hold*? It holds or means that what we will do is fixed no matter what we decide to do. Is that an implication of causal determinism? Sam Harris does not think so. But back for a moment to what fatalism holds. Some ancient rebuttal went like this. If determinism is true then should we not call the doctor, since we will get well or we won’t no matter what we do? No. You will get well if you call for the doctor and you will not get well if you don’t. So by all means you should call for the doctor. And we can make a choice (or what seems like a choice) to call the doctor or not, **even if** whether we call the doctor is already decided (but of course only already decided in the eyes of a Laplacean demon). It can seem contradictory that you should call the doctor even though whether you call the doctor is already determined. But if we introduce the difference between metaphysics/ontology and mere epistemology it might make it easier (for some) to see how it’s not contradictory. Metaphysically it is decided, but at the level of our understanding and knowledge, our epistemology, it is not known to be decided and therefore can feel like a choice. And without harping on that language of “it can still feel like a choice” let us instead put it this way: you cannot wait around to see if you will get well. You must act.
Let me move on to your next point. You said I seemed to be using a COMP notion of free will meaning something like conscious reasoning. I am not sure what you are attributing to me. Because I do not simply use one definition of FW. I’ve remarked that there are two definitions. And what free will means, which of the two definitions it is becomes an interesting question. Many philosophers used to say that the folk had the LIB definition in their heart of hearts or that when pressed with Socratic questioning the folk reveal they’ve been using “free will” according to the LIB definition.
And philosophers used to be pretty certain they could figure this out from the armchair. Then along came experimental philosophy. They took thought experiment formerly considered from the armchair and put them to the masses using the survey technique. This paradigm shift occasioned a reconsideration of the role of folk intuitions in the philosophy’s methodology of conceptual analysis.
But I should not go down that path even though what I’ve said so far about X-phi won’t make sense to someone who already didn’t know what I was talking about.
So let me return to “conscious reasoning”… Yes materialists do not think LIB makes sense. “So they take something that is not free will and use that so that they can pretend they have free will…” etc. But now you seem to be talking about compatibilist. But If the implication is that I am that compatibilist, I do not think you have any textual evidence to pin that on me.
And as far as sleepwalking, I did not to use it to illustrate lack of free will. Only that it got me to thinking about free will.
You wrote: sleepwalking or wakeful walking “in neither case, at least from a hard determinist perspective, do you have free will.” Yes for a hard determinist no matter what event you bring up it will lack free will in the LIB sense because every event has a cause and the LIB definition of FW requires that the “free” event be without a cause. (And the Hard Determinist shares the LIB definition with LIB)
> Metaphysically it is decided, but at the level of our understanding and knowledge, our epistemology, it is not known to be decided and therefore can feel like a choice. And without harping on that language of “it can still feel like a choice”
Sure, but "feels like a choice" and "is a choice" are two different things, and that distinction is the thing that I wanted to harp on. My gripe is that the term "free will" should be reserved for a thing that really "is a choice", and not one that just "feels like a choice". That gripe I have isn't with you personally, and if I misread what you posted above, and you're not actually conflating those two things under the same heading "free will", then it's not at all with you. I just enjoy complaining about it.
> And as far as sleepwalking, I did not to use it to illustrate lack of free will. Only that it got me to thinking about free will.
> But If the implication is that I am that compatibilist, I do not think you have any textual evidence to pin that on me.
Okay, it seems I misunderstood, apologies.
Hi David, Thanks for the apology. But it's all good. I am finding this posting difficult. My writing takes thousands of hours to become clear. Like my Missouri Review essay reads much much different as compared to what I am producing here. The essay is way better. I was hoping posting here would be a chance to ... I guess I am not sure what of many things I had in mind. A chance to discuss, listen, think out loud, test out ideas. I will try to remember my purposes. And I'm just not going to be able to get a reply to each reply to my original thread. I am saying this for myself and my other readers besides you. You are all good. I appreciate you.
For what it is worth, I will say the difference between "feels like a choice" and "is a choice" is also a difference I am interested in.
>materialists do not think LIB makes sense.
Materialism.doesn't imply determinism, and libertarianism doesn't have to be based on supernatural spirits.
> to the existence of brain-scanning technology that can read the inner contents of your mind and know what you are going to do before you know.
If you're referring to the studies were readiness potential (RP) is believed to be the brain making a decision before you're conscious of it, Aaron Schurger argues that the RP is largely an artifact of averaging many trials together, and it reflects ongoing, fluctuating, and random neural activity in motor areas. IIRC, when this background activity happens to cross a threshold, a movement is initiated. But aligning many trials on the movement makes this random buildup appear to be a smooth, purposeful signal.
Curt Jaimungal interviewed him...
https://www.youtube.com/watch?v=yDDgDSmfS6Q
The Libet stuff--Libet 1.0 (readiness potential), 2.0 and Haynes's stuff which could be Libet 3.0--is interesting. Maybe. I think my take on it (which may not be clear in the essay, even in the full essay--had to cut lots for space, before and after acceptance) is different than most people’s take and unassailable by the arguments Libet debunkers usually offer. (I’m not trying to be grand and call myself unassailable. I am just writing quickly.) One thing I think is that Libet et al is an empirical case against LIB not COMP. So the COMPs who say “Yeah my conscious thought bubbles up from my nonconscious mind and is detectable before I’m aware, that’s fine, it does not make me unfree. It’s still me . Etc.” The COMP who says that is on good footing as far as I am concerned.
The rest of my position on Libet is that the most important thing to take away (or that I take away) from the Libet and Haynes stuff is something that could have been argued a priori without all the empirical science. And if fact was argued by Nietzsche and more recently by Sam Harris.
What do I think is the most interesting thing to take away from Libet? It is that conscious thought does not appear uncaused or out of nowhere nor is conscious thought like a decision one long unbroken chain of conscious preceding thoughts back to the conscious reception of the stimuli. Instead nonconsciousness plays a huge role in supplying the content of thought to consciousness. So decisions involving conscious careful deliberation ultimately can be traced back to some thought that bubbles up from nonconsciousness—bubbles up as much as a decision bubbles up when you ask yourself “What am I hungry for today?” And the answer comes. Happy to discuss more but will post this incomplete response now since I am falling behind answering folks comments.
Appreciate it!
David
Oops, so to connect the final dots... The brain scans are picking up the nonconscious processes before they provide the content to consciousness. But we need not have done experiments like that. There is an argument I think A reductio maybe. Because where else is conscious thought going to get its content. My essay quotes Nietzsche or references him anyway--he did not use the phrase "bubbles up" -- he said thoughts come "unbidden". Sam Harris for his part say in meditation you learn to pay attention to the phenomenology being alone with your thoughts, you realize they come unbidden or out of nowhere. And in my essay at the very end I describe a meditation session and how sure enough when you meditate you realize how wild the mind is. Thoughts coming in out of nowhere while you are trying to focus on the breath. We are often 3 seconds or more already going down a tangent like plans for dinner having turned from the breath to the thought about dinner (which bubbled up, unbidden) WITHOUT REALIZING it. Indeed, when you realize you are not focusing on the breath, it is often ALSO the case that you can recall that you have not been doing so for a relatively long time. What does that mean? That means that when the thought about dinner bubbled up you did not notice that it bubbled up, you just went along and thought about dinner--you identified with it for lack of a better term, you adopted it, probably, as part of your consciousness's domain.
I think that is pretty interesting and I imagine that some large percentage of the population, when pressed, without first reading the above, would say that their stream of consciousness or their experience is all the domain of consciousness and its a chain unbroken of conscious mental events all day long etc. But I think we know it's not. There was a journal article that I read and did not really write up for my essay called "Do conscious thoughts cause behavior?" And we gotta be careful we approach it asking the right question and so on because their answer is a qualified "yes". But when you look closely almost every one and maybe indeed every one of their kinds of cases involved a chain of mental events from stimulus to decision and action in which nonconscious processes played a role. Do nonconscious processes cause behavior? Yes, they are implicated in every causal chain from stimulus to action. (There is some subtlety here but I've written enough again for now)
Thanks!
The whole question of free will always seemed to me to be an argument about semantics.
"Did the murderer choose to do it?" Well, his brain shuffled some electrons and puffed out some neurotransmitters, and he doesn't seem to meet the diagnostic criteria for us to feel sorry for him and send him to forensic mental health. He was probably also not dissociating or sleepwalking at the time.
"But did he, like, choose to do it?" Does it matter? What real impact will the answer have? If we want to choose whether to let him go, or to hate him, those questions can be answered without finding an answer to whether he had free will.
The whole way down it's just semantics.
"What about uncaused effects?" Could be baked into quantum mechanics, but does free will really hinge on some weird fact about observing subatomic particles for anyone? Is your will any freer, or less free, because an electron went left instead of right?
"Am I the one choosing?" Well, how do you define "I"? Do you think you're the brain? Then you're choosing! Do you think you're the identity or the consciousness or some other facet of intelligence? Then you're sometimes choosing! Do you think you're an immortal soul separate from physical reality? Then I dunno, I'm not a theologian, but that sounds like a problem which depends on how you define those words.
I bet actual philosophers have gone over this time and time again, but I'm confused as to why it doesn't get acknowledged more.
>The whole question of free will always seemed to me to be an argument about semantics
Its.*partly* about semantics, and so are many other arguments.
>"But did he, like, choose to do it?" Does it matter?
Yes. It matters if we punish someone for a crime committed under duress, and it matters if we punish someone who is not a mentally capable adult. Both of these relate to free will under compatibilist semantics.
>"What about uncaused effects?" Could be baked into quantum mechanics, but does free will really hinge on some weird fact about observing subatomic particles for anyone? Is your will any freer, or less free, because an electron went left instead of right?
Yes, under libertarian semantics. Libertarian free will require a that you could have done differently under the same circumstances, and that's only possible if the universe works a certain way, and the brain works a certain way. It's partly a matter of semantics -- the libertarian definition has to be the one the fuss is.about -- and partly a matter of reality.
>The whole way down it's just semantics
No. See above.
>"Am I the one choosing?" Well, how do you define "I"? Do you think you're the brain? Then you're choosing!
Even if the choice was a foregone conclusion a million years ago? There isn't a single obvious to that question, either.
> No. See above.
I see the factual question about whether the universe is deterministic, but don't understand how that's connected to free will. It seems like, if I kept asking questions here, I would eventually arrive at a choice between "neither determinism and randomness is really free will" or "God did it"-style metaphysics about how the randomness is supernatural, and we might as well just short circuit the bit about free will and go straight to arguing the metaphysics instead.
> Even if the choice was a foregone conclusion a million years ago? There isn't a single obvious to that question, either.
Again, I think this is a semantic argument, because if we drilled down into "what does it mean for a brain to choose something?" then we'd eventually have to define it in a way that gives us an answer. Either you're asking something like whether the brain is capable of responding to stimuli with a bunch of caveats about time's arrow and what a response means (it is), or whether the brain is capable of spontaneously doing its own thing (maybe, but then it just decomposes into a different argument that's not about free will). The wrangling is either semantic or it turns into a different philosophical argument with free will just plastered awkwardly on the top.
I feel a bit lied to by all the people who suggested this was an important question with real-world implications, because it seems like a problem created by religious people to help them think about sin. I may just not be cut out for philosophy.
>I see the factual question about whether the universe is deterministic, but don't understand how that's connected to free will
Determinism would refute free will, under one definition of free will.
>It seems like, if I kept asking questions here, I would eventually arrive at a choice between "neither determinism and randomness is really free will
Not from me. I'm a naturalistic libertarian.
Naturalistic libertarianism doesn't require any kind of soul, or fundamental third alternative to determinism and indeterminism. It assumes indeterminism, and there is no need to override determinism if the universe is not fully.deterministic in the first place. Partial determinism means there are some things I cannot do, but also some options I can choose between.
According to science, the human brain/body is a complex mechanism made up of organs and tissues which are themselves made of cells which are themselves made of proteins, and so on.
Science does not tell you that you are a ghost in a deterministic machine, trapped inside it and unable to control its operation. Or that you are an immaterial soul trapped inside an indeterministic machine. Science tells you that you are, for better or worse, the machine itself.
Although I have used the term "machine", I do not intend to imply that a, machine is necessarily deterministic. It is not known whether physics is deterministic, so "you are a deterministic machine" does not follow from "you are entirely physical". The correct conclusion is "you are no more undetermined than physics allows you to be".
So the scientific question of free will becomes the question of how the machine behaves, whether it has the combination of unpredictability, self direction, self modification and so on, that might characterise free will... depending on how you define free will.
There is a whole science of self-controlling machines: cybernetics. Airplane autopilots and , more recently, self driving cars are examples. Self control, without indeterminism is not sufficient for libertarian free will, but indeterminism without self control is not either
All of those things can be ascertained by looking at a person (or an animal or a machine) from the outside. They don't require a subjective inner self... unless you define free will that way. If you define free will as dependent on a ghostly inner self, then you are not going to have a scientific model of free will.
>Again, I think this is a semantic argument, because if we drilled down into "what does it mean for a brain to choose something?" then we'd eventually have to define it in a way that gives us an answer
We can define it in two ways that give us two potential answers.
Which still doesn't make it entirely semantic.
>Either you're asking something like whether the brain is capable of responding to stimuli with a bunch of caveats about time's arrow and what a response means (it is), or whether the brain is capable of spontaneously doing its own thing
A response doesn't have to be deterministic, so that's a false dichotomy.
>but then it just decomposes into a different argument that's not about free will).
I dont see why that would be the case.
>because it seems like a problem created by religious people
I'm a *naturalistic* determinist.
> It matters if we punish someone for a crime committed under duress, and it matters if we punish someone who is not a mentally capable adult.
Yes, of course it's wrong to "punish" someone for a crime they didn't "choose" to commit.
But that doesn't mean it's wrong to put them in prison. When people have a tendency to commit crimes, we put them in prison to stop them from committing crimes.
This is prison-as-incapacitation, not prison-as-deterrent.
If they don't want to be in prison, it's still punishment from their POV.
Let's suppose there's a person who commits lots of murders all the time; let's suppose this person does not have free will by some definition. I think this person should be in prison. It sounds like you're arguing they should go free?
>. It was weird enough to think that I had carried on a conversation without free will one, two, or three times.
You were carrying on a conversation in some unusual state of mind. It's not clear that free will was lacking, or that it's present the rest of the time, since some debt it wholesale -- and there is also semantic confusion about what the term even means.
Any chance that sleepwalking may involve some consciousness, but we would not remember consciously? After all, dreams are conscious experiences even if we hardly remember them. For comparison, Daniel Kahneman was puzzled that people would "prefer" enduring lengthier colonoscopies if only residual memories are less gross (due to better ending protocol).
Hi dasmonde,
Thanks for your comment. I guess there is a chance of that. I don't know the empirical science of sleepwalking. And, in any case, in my essay I only use the sleepwalking episodes as a path into thinking about free will.
But as far as the significance of sleepwalking as more than just an anecdote... sleepwalking foregrounds how a lot of behaviors we think require consciousness can be performed by nonconsciousness, to coin a term. This thing I just said of course assumes we are not conscious while sleepwalking instead of your contention that we might be conscious and not forming memories or will later forget what memories are formed.
But, if we are at an impasse there, then I would just come up with another example to demonstrate how much of the behaviors which many of us seem to think only consciousness can control, can be and actually are controlled by nonconsciousness. Timothy Wilson’s book Strangers to Ourselves comes to mind.
That is one change my essay speculates must needs come eventually: presently people give way too much to consciousness not enough to non-conscious parts of the mind. Plus I am not sure if regular people or the folk think of nonconsciousness as a bunch of little selves or if they think of them as blind instinct. It might be that soon we will have to think about nonconscious processes as little selves that can make decisions (by which I mean we won’t conceive of them as merely instinctual).
Thanks!
Presently people give way to non-conscious parts of the mind incomparably more than at Freud's time - whether it is in marketing, or management, or politics, or popular culture, or science. Perhaps we are neglecting consciousness actually, as public discourse became more rhetorical and emotive rather than rational. Trying to influence anyone by factual argumentation rather than Cialdini's prescriptions (or social "mechanisms") is almost a bad social taste nowadays. The biology and neurology are primed to delegate some causality to YOU (as a decision making center), is it not a shame to make wholesale excuse that we are so determined?
https://link.springer.com/article/10.1007/s10516-018-9377-3
Sometimes people ask me why I'm a determinist and honestly I just don't think I have a choice.
Funny - I believe in free will because, if I’m wrong, it’s not my fault!
Heads up! It looks like the comments section on the archive page of the old SSC blog has gotten spammed by crypto bots. (Sorry if I am the thousandth person to notice and post, saw it last night and figured I would share in the next open thread.)
I made a new blog post on how many dollars it costs to take a life, the opposite of effective altruism, calculating the cost for everything from driving and smoking, to abusing amazon prime or buying drugs. Check it out if you're interested!
https://postalpigeon.substack.com/p/effective-evil-harming-the-most-lives
Not really the point of your write up, but I expect that you could sponsor some extreme BASE jumping events. As with bull riding, I don't think there is much money in BASE jumping so a few million dollars could probably encourage enough marginal behavior to cost a number of lives. Be subtle, though, so the BASE jumpers don't catch on.
I did think of paying for promoting bad health practises as another thing, but sponsoring dangerous activities is another good (or bad) idea. It's hard to calculate but if you paid for free base jumps for people, about $500 each jump. About 1 in 1000 jumps results in a death, which sounds way too high for the number of people that still do it. So only $500,000 needed to cause one death!
Actually, this would be paying for tandem jumps, with a trained jumper leading, so it would be two deaths, so only $250,000 each!
Nectome is seeking a Research Associate with a DVM degree to join our brain preservation research program, which investigates post-mortem cerebral perfusion and tissue preservation using both small and large animal models.
We strongly value previous research experience and hands-on work with large animals. The successful candidate will plan and execute experiments involving research animals, performing anesthesia, euthanasia, surgical preparation, surgery, vascular cannulation, intravascular perfusion, tissue harvesting, and downstream analysis.
This is a full-time position in Vancouver, WA, with compensation ranging from $100,000 to $150,000 USD per year, depending on experience. Applicants should submit a CV, a brief cover letter describing their experience, and up to three references to b@nectome.com. Review of applications will continue until the position is filled.
Can you share the reason why you ask for references up front? I'm not a fit for this position, but I've always been reluctant to supply references before an interview. This is mainly for two reasons:
1. Worst case scenario; the company has no intention of hiring the applicant based on their resume, but uses the references' contact info to headhunt or bother them to find new candidates.
2. My references' time is valuable. They're busy. They don't need to know every time I apply for a job, but in cases like this I need to notify them that their contact info has been/will be shared.
Now, 1 is less likely to happen, but if you have a valuable reference that's hard to reach...
FYI: I'm not claiming that your company is doing this, just a general pet peeve.
My latest Fog o' War post. In which I muse upon whether Putin is acting rationally within the constraints of the information being given to him, or whether he's delusional. I go back and forth on why he's making such poor decisions, and a case can be made for either hypothesis. Plus I highlight some of the other what-the-f**k-were-they-thinking-! moments of the Russian authorities seen over the past week.
https://substack.com/home/post/p-206879119
Minor note- I think your post title says 13 June 2026 instead of 13 July 2026
Corrected. Thanks!
I’m calendricly challenged.
Are the Russian authorities entirely irrational though ? Or are they merely betting that they have more oil/tankers/trucks/whatever than Ukraine has drones ? If you need one tanker to get through, and you have effectively infinite tankers, you might as well send 100 of them and call it job well done.
One obvious retort is, "but what about the lives lost on all those vehicles ?", but this is Russia, they don't care. Another retort might be "ok but their domestic economy is already starting to struggle for lack of oil", and here they do care at least a little. But the Russian mentality had always been to endure hardship, and I think the leadership might be betting on it kicking in these days as well. This is why they are framing the Ukraine War as a sequel to WWII: it's us against the world, we must protect the Motherland at all costs, sacrifice for a brighter future, etc.
"Russian mentality had always been to endure hardship, and I think the leadership might be betting on it kicking in these days as well."
Quoting IC: "America has patriotism. Russia has spite." (the Russian grit and endurance is legendary, and yeah, I think Russia can go on for far longer than people think.)
[I am rooting for "whatever ends this idiotic war" at this point.]
>"For instance, his claim that Zelenskyy, who is Jewish, is a Nazi."
AIUI, the post-Soviet connotation of "Nazi" is anti-Russian rather than antisemitic, so the claim doesn't sound as patently absurd in Russian as in English.
I'd say it's not just post-Soviet. Pro-Palestinian people call Israel a Nazi state, with the same connotations: if a country is controlled by an ethnonationalist ideology that aims to either forcefully assimilate, expel, exterminate or at least limit the rights of minorities "to preserve its essential character and its purpose as the homeland for its people", then it's a Nazi state. The minority in question doesn't have to be Jewish.
Those seem like two different ways to generalize the historical Nazis, whether you focus on the part where they:
a) fought against Russia;
b) tried to increase their Lebensraum by exterminating/expelling ethnic groups they called inferior.
Yes. The former explains Neo-nazi Synagogues, the latter does NOT.
I think you're channeling the scenes of the Jewish Suicide Squads from Life of Brian.
That would be considerably more likely had I ever seen The Life of Brian.
https://www.thenation.com/article/politics/neo-nazis-far-right-ukraine/
Seriously, this should not be controversial. We can look at an army commander (former "minor" elected official, think American Congresscritter), and see he's making neo-nazi rhetoric.
https://euromaidanpress.com/2022/04/07/what-is-azov-regiment-honest-answers-to-the-most-common-questions/
Whitewashing. Although it is true that Jews do serve in the Azov militia, that's where we get things like Neonazi synagogues (with swastikas on display), and Azov making terroristic threats against Zelensky (so if you don't want to call them Nazis, do call them terrorists, it's accurate).
https://en.wikipedia.org/wiki/Ihor_Kolomoyskyi
The oligarch behind Azov is jewish (well, the first backer of the group the new backer is apparently sunni muslim, but you didn't see neonazi synagogues under him). But if you want to know why Putin is saying "They're Nazis!" You have to look at what he was seeing... which is some very "nazi-like" people who Do Not Like Russia, and Want The Russian-speaking Ukrainians to continue being bombed, even if Zelensky doesn't want to.
https://easternbrief.substack.com/p/the-azov-divorce-how-ukraines-elite
But then we encounter situations of cognitive dissonance when one era's Nazis become another era's oppressed. For instance, we have the example of the Grand Mufti of Jerusalem, Haj Amin al-Hussein, who was appointed by the British to preside over the Arab population of Palestine (many of whom had emigrated from Jordan and Syria in the earlier part of the century). In the 1930s, he incited pogroms against the (small) native Jewish population, and the more numerous Jewish settlers. Despite the complaints of Jewish leaders, the British turned a blind eye to his atrocities — although they finally got fed up with him and issued an arrest warrant against him not long before the outbreak of the war. He fled to Germany, and publicly aligned his Arab nationalist cause with Nazi Germany. He met with Hitler in 1941 and Hitler made him an honorary Aryan and confirmed that the Arab goals of eliminating Jews from Palestine were part of the larger Nazi effort against the Jews. Al-Hussein spent the war broadcasting Nazi propaganda in Arabic, and after the war, he was interned. But with the assistance of anti-Jewish elements in French and British security orgs, he was able to flee to Cairo, where he formed a government in exile. His internal rivals eventually sidelined him from leadership, and they went on to found the PLO.
So, we had German Nazis, Arab Nazis, and now we have (reputed) Ukrainian Jewish Nazis. And of course we also have Russian Nazis, but they seldom get mentioned in Western media...
https://x.com/ChrisO_wiki/status/2069783708966502623?s=20
Yes, Palestinians and their supporters call Israel a Nazi state, because of the ethnonationalism and the violence. This has *nothing whatsoever to do* with the bit where Russians call e.g. Ukrainians Nazis. There are two completely different definitions of the word "Nazi" at work.
Well, really three. There's the Russian version, which means "Anyone who fights a war against Russia", the American Leftist version, which means "Anyone to the right of Joe Biden that I want you to hate", and the version everyone else uses, which is "Murderous ethnonationalists like those guys who used to run Germany; exact boundaries TBD", And there may be value in trying to pin down the borders of that third version, figure out whether contemporary Israel fits the definition. But any attempt to develop a Unified Theory of Nazis that tries to tie any two of these definitions into a cohesive whole is doomed to failure.
Might as well try to find a single definition of "Spam" that includes both canned meat and mass unsolicited email.
Neither of them even belong to the emperor, so it looks like this checks out.
Given that this is a cultural meaning in Both Ukraine and Russia, yeah. Hitler is seen more favorably in the Ukraine, because he fought against Russia. (the genocide? not so much). Hitler is only the antichrist in Western Civ (India for sure loves him, and that's a very different culture).
Putin is a chronic vacillator, he is incredibly biased towards not doing anything. Like Merkel. That's how Russia has squandered most of its temporary advantages over Ukraine, allowing the EU to turn Ukraine into a credible threat.
He has driven the country past multiple off-ramps, because a ceasefire would've required doing something completely different, and quickly. And he hates quick decisive actions.
Strange factoid:
All states have Safe Haven laws, under which overwhelmed parents (usually mothers) can drop off infants at locations such as hospitals, relinquishing all parental rights, no questions asked. In most cases the laws apply to infants up to one month of age. When Nebraska enacted its law in 2008 the legislators forgot to add an age limit, and within a few months parents dropped off 35 children ranging in age from 10 to 17.
Those kids becoming wards of the state might be better for all involved.
From what I could find out, several of the older children were from the same large family, the mother died and the father was too overwhelmed to raise them. Other relatives soon took the children.
Yeah if the parents of a child truly believe they can't care for them, it seems better for the state to say "okay" and take care of them.
The unintended consequences on this seem pretty obvious to me.
Unless maybe you add "and the state is authorized to castrate the parents". Which carries... tradeoffs of its own.
It's already possible to do this with an infant, I wouldn't expect much change in the birth rate if it applied to all-age kids.
World #1: Bad parent decides to abandon his kids to the streets/kick them out of the house. Eventually maybe the abandoned kids get picked up by the cops and put into the foster system. (Or maybe some pimp adds them to his stable, or they sleep on the streets, or find someone to take them in informally.)
World #2: Bad parent can't or won't bother keeping his kids fed, clothed, in school, etc. Eventually maybe CPS gets called and eventually after a year or so of bureaucratic and legal procedure, the kids get taken away and put into the foster system.
World #3: Bad parent takes his kids to the fire station and says "I can't/won't care for them anymore." They get put into the foster system.
It seems to me that in almost all cases, World #3 is better than World #2 or World #1.
The potential downsides of this are:
a. Maybe some parent does this in a fit of anger or crazy and later regrets it.
b. Maybe some parent doesn't want to support his (sick, disabled, troublesome, annoying) kid and fobs him off on the state.
c. Maybe the foster system is such a meatgrinder that even the kind of parents who want to abandon their kids are much better, and maybe this program somehow encourages many of those better-than-nothing parents to abandon their kids to the state.
d. Lots of people skipping birth control (and for some reason not having much problem with birth) because they know they can always give the kid up to the state. Including parents who continue to drink, ingest substances, etc.
>d. Lots of people skipping birth control (and for some reason not having much problem with birth) because they know they can always give the kid up to the state.
Is more children being born a bad thing? Or are you concerned this will make fertility more dysgenic?
>Maybe some parent doesn't want to support his (sick, disabled, troublesome, annoying) kid and fobs him off on the state.
Is this so bad? Having the burden of supporting disabled children not land randomly on some parents would be good. In the past a lot of these kids would have been exposed to the elements.
I was reading more about the relinquished older children in Nebraska and as best I could determine, most of them had serious mental/behavioral issues that the parents couldn’t manage. Keep in mind that if you exclude the siblings with the dead mother and overwhelmed father there were only about 30 children.
I have a friend who has worked in child support for many years. Apparently there are a few men who have some mix of minimum-wage employability and like eight cases by five women. She and her coworkers used to joke that they should put a vasectomy clinic in the basement of their office, and just send the dude down there after his second separate judgement.
>She and her coworkers used to joke that they should put a vasectomy clinic in the basement of their office, and just send the dude down there after his second separate judgement.
Likely that the women would have gotten impregnated by some other loser.
…but perhaps one who could occasionally make his child support payments….
The smallest bank in the United States is gone. Last week federal regulators shut down Kentland Federal Savings & Loan of Kentland, Indiana, which had $3.73 million in assets and $3.65 million in deposits. It had been founded in 1920 and remained in family ownership ever since.
Why?
Regulators said it was severely undercapitalized.
When you want to call a friend, how do you answer "Who's free now?"
I made a little mobile app that connects you with a friend in less than 30 seconds. It’s called Beacon. You send a beacon to a group and everyone gets pinged at the same time. The first person to answer the beacon gets connected for a 1-1 audio-only call.
(The signal drops silently for everyone else, no pressure to answer or missed call badges. There are quiet hours settings and of course you can manage inbound pings with OS-level focus modes)
It works best if join with at least four people you don't speak with as much as you'd like. I have a couple dozen connections on the app now and it feels like magic to me.
Would love feedback from ACX-ish people who still like phone calls, or wish they had more of them. You can get the iOS version on TestFlight here:
trybeacon.chat
I have an Android version too but still adding people manually for now. Ping me if you want in the Android beta. Yes it works cross-platform!
Never mind about ios vs. droid…If I have some friends i communicate with via my phone’s native calling feature, and others i communicate with via WhatsApp, can your app handle only one of those? Both?
Neither, the calls are via the Beacon app itself.
It’s important recipients know it’s an incoming Beacon and not a normal phone call.
Would these calls use data on my data plan?
Beacon uses VOIP so if your device was connected to WiFi it would not use your cellular data plan. If you were not on WiFi it would use cellular data at an estimated 16MB per one hour of calling.
Got it, thanks. That's helpful info to know
Is this an actual problem that people have? Maybe I'm just old but if I want to talk to someone then I call my friends in order of how much I want to talk to them until someone picks up. How common is the problem of "I need to talk to one of these 20 people and I don't care who it is and I have no time to dial them sequentially"?
It’s a big problem for many people. Most people don’t answer calls most of the time! Calling twenty people and waiting for it to ring out (and reaching nobody) would take ten minutes of dialing and waiting. And then there would be twenty people feeling like they owed a callback.
What if you’ve only got 15 minutes to chat but spend half that time dialing?
If you’re truly old maybe you are time rich and so are many of the people you’d talk to, but for people who work and have kids and other commitments, it’s very hard to catch up regularly and spontaneously with all the people they might like to, but there are also lots of little moments that could work, and Beacon finds them.
Well then I'm definitely not in your target market. I have to say that I find it difficult to identify with the scenario as you describe it. "My time is extremely valuable" seems incompatible with "I don't care who I spend the next 15 minutes talking to." But I'm middle aged. Maybe social norms have changed.
Still, I can see how it's a clever idea even if it doesn't solve a problem that I identify with. I didn't think social media would be useful either. Best of luck.
I believe you posted this before, and I can't help but have the same thought. It seems like an excellent premise for a Seinfeld episode. Why not just start a group chat via text, and then nobody will have to feel like they got left out?
Group chats are a very different use case and none of them solve this problem! There is a reason nobody combines all their family and multiple social groups into one massive group chat.
Fair enough, but I still can't help thinking that this is somewhat like tossing the apple of Discord into a room. I get a Beacon and I have to pee, so I wait and then I call, but I got beaten to the punch. Well, did he really want to talk to me or just anyone? What are they talking about behind my back? I don't know, maybe I'm just very old-fashioned.
Love this idea
Please give it a try and lmk what you think.
If you feel friction at trying it, can you introspect on why let me know — it might be something I can address!
In order to contribute to the quality of our perennial discussion of the topic, I will henceforth refer to contemporary architecture as "Corbusmaxxing Bauhauslop." You're welcome.
Can you explain the "Corbusmaxxing" part? Is this a reference to the Swiss architect Corbusier?
It'd be at least interesting to say Bauhaus's influence ripples through all modern architecture and makes it slop; I'd disagree off the bat but be open to hearing your arguments.
I think your biggest obstacle is going to be getting me to believe that all contemporary (last 50 years? last 100 years?) architecture (residential AND commercial, all subcategories inclusive?) fit in a single bucket, stylewise.
I was referring to the High Priest of Reinforced Concrete, yes. "Contemporary architecture" should be taken to exclude prolefeed.
By "prolefeed" you mean what?
“Stuff regular people like” is my guess.
That would be counterintuitive to me. I thought a component of the complaint about modern architecture was that regular people like it (in the same way regular people go in for all kinds of ugly trends, for example the ridiculous false dormers and cascades of 6"-deep gable ends you see on the facades of production houses built in the last two decades) but elites who appreciate True Beauty see it as the ugly trend it is, yet here it would mean Brendan Richardson has excluded the stuff regular people like. What's left would then be "Corbusmaxxing" (so, lots of concrete?) Bauhaus-slop (which I associate with brick facades, rounded corners, and block glass masonry windows for some reason, though that might not be the intended image) that is beloved by the elites.
I guess I find the discourse on architecture confusing in general and this is just an example of why.
I tend to look at buildings from more of a technical standpoint--use of materials, performance, serviceability, human factors, etc. Whatever emotional or artistic vibes the buildings give off when looking at them from the street is not as important to me, so criticisms of those aspects often end up seeming like a hall of mirrors.
I think I would recommend you study some pictures of the Palace of Westminster and the Scottish Parliament Building until you can intuitively understand why most people have an intuitive preference for one over the other (and which one that is).
It’s tough not to feel like AI has unfairly changed the rules of the game of life. It’s difficult to pour myself into non AI related projects in the same way I did in 2021. Like, if we are three or four years from superintelligence, shouldn’t I just be working on that, even if I’m woefully unqualified? I even find it suddenly tough to care about politics and similar. Who cares about immigration policy if the game board is about to flip? This has caused me some minor mental distress.
I will, apparently, forever recommend C.S. Lewis's short essay "On Living in an Atomic Age" where he points out that nuclear weapons just added "one more chance of painful and premature death to a world which already bristled with such chances and in which death itself was not a chance at all, but a certainty".
> If we are all going to be destroyed by an atomic bomb, let that bomb when it comes find us doing sensible and human things [...] not huddled together like frightened sheep and thinking about bombs. They may break our bodies (a microbe can do that) but they need not dominate our minds...
And, yeah, I think there's a real threat to AI... but most of us can't do any more about it than we can about people with their fingers on Big Red Nuclear Buttons - and if you can care about immigration policy in a world with nuclear weapons, it seems like it should be possible to do it it in a world with a potential for superhuman AI.
Every few years I rediscover that my anxieties are deflected from other areas of life to the One New Thing that I've noticed.
I was stressing out about AI and wealth inequality to the point where I would stay awake until 3am sleepless. Then I started exercising and hanging out with my friends again, worked a little less late into the evenings, my stress dropped back to a sensible level of worry, and I realised that I had been taking the result of a number of different stressors and assuming it was all AI. Exactly like I did last time, except with job related stress, or the time before that, with friends, and so on back into the past.
I am not saying this is the case for you, but maybe this comment will resonate with someone and help them find perspective again.
This is the idea of free radical anxiety in a nutshell, and I think everyone suffers from it. I have certainly spent a lot of my time in life being anxious, and the more I have worked on it, the more I have come to the conclusion that it is almost entirely self-generated. There is some essential anxiety in me that freely attaches itself to things that I think about or imagine. These things then become a smoke screen that hides the real source of my discomfort. I have developed my own mantra to help me navigate this. "The definition of anxiety is knowing something to be true but refusing to believe it." (YMMV but I think it's a defensible construct.) I came to the conclusion that one of the biggest causes of my anxiety was fretting about things that I really had no control over. That's a difficult thing to believe but in many cases, it's true. (For example, I spent an inordinate amount of time keeping myself awake wrestling with all the misery that was in the world. Needless to say, that made me miserable. Now I have adopted a couple of homeless people that live on my block or in environs, and make an effort to communicate with them and give them some money when I can. It's not like I'm solving the world's homeless problem, but I'm doing what I can, given my reach.)
Your programme of more human contact and more involvement with your physical nature is a good antidote.
Settle down. AI will be significant but it won't be everything. It will propagate slowly through the economy. There will be many roadblocks, both technical and political. Businesses take many years to adapt to new technologies and workflows. Electricity took 40 years to fully penetrate the US and that was a much more valuable innovation. The real world is complicated and things take longer than armchair philosophers appreciate. Just relax. Intelligence doesn't mean omnipotence. The world will change, equilibria will shift, but life will go on.
I would agree with you on a lot of this, were it not for the fact that AI is going to have and is having a profound effect on warfare. It's interesting that AI and the birth of the Internet are rather opposite in the sense that the Internet was essentially begun as a defence-related project that also became very useful in the civilian world. I feel like AI is going in the opposite direction. Something that has been developed by private companies with a profit motive is now really too important to be left completely in the hands of civilians because of its defence implications.
I don’t think the analogy to the internet is particularly good. The internet didn’t replace general human analytical ability.
The Internet did that too. Something else will too in thirty years.
If we're actually going to get superintelligence any time soon, you *are* working on it by contributing to its training data in whatever field you're working in.
Just make sure to do excellent work, and document it well.
We're not going to get superintelligence in the next five minutes/five years (maybe not even the next five decades).
AI will make money, but that money will go where it always goes in new technologies, and we are not going to see "glut of productivity means enough tax revenue to pay everyone to sit around and play the bongos". If you have a job, and manage to keep that job, you will need to keep working that job to pay for the cost of living, just as you do now.
You'll need a raft of new skills/upskilling, as with the past drift to credentialism and 'employers want freshly minted employable staff out of high school with these, these and these sets of skills' emphasis in education, just to get/hold on to that job. Same old same old.
The game board flips, and you will still be breaking stones:
https://en.wikisource.org/wiki/New_Poems_(Yeats)/Parnell
PARNELL came down the road, he said to a cheering man:
'Ireland shall get her freedom and you still break stone.'
Maybe it helps to internalize that, despite all the risks, the modal outcome is till that we will muddle through as always.
How screwed is the world right now, in an economic sense?
Specifically, assuming the U.S.-Iran situation remains at its current impasse for at least a few more weeks or months, and the Strait of Hormuz sees either no commercial traffic or sharply reduced commercial traffic, how bad will the global economic effects be and when will they start.
My overall picture is like this:
1. The strait effectively close around the beginning of March.
2. Because it takes on the order of a few months for oil on an outbound tanker to turn into refined product available on the consumer market, the direct effects of this weren't felt immediately.
3. But by the same token, even if full traffic resumed tomorrow, it would take some time for production effects of the closure to be fully reversed.
4. The period of U.S.-Iran cooperation following the signing of the MoU saw a significant increase in traffic while it lasted, but it never came close to normal. So the total amount of oil (and other key materials) shipped during that time can only amount to a few days' worth of normal production.
5. The only reason there haven't been larger price spikes yet is because of governments putting their thumbs on the scale in ways that aren't sustainable or don't especially help the long-term picture: things like releasing reserves and lowering fuel taxes.
To me, all of this together suggests a picture where when the price stabilization measures start to fail, they'll fail hard. But I don't have high confidence in this, and I don't have a great sense of either timetables or magnitude. How long before prices spike and how high will they spike? How bad will the economic fallout be and how long will it last? I've seen this issue discussed *somewhat* in the last few months, but far less often and less thoroughly than seems appropriate given the scale and scope of the potential consequences.
I have no idea. Oil prices are going up. https://tradingeconomics.com/commodity/crude-oil So yeah things are getting worse. But I think almost everyone wants this to work out so I'm not that worried.
I admire your optimism, but I don't think "everyone wanting this to work out" (depending on what you mean by "working out") is the case. There is a real schism in Iran between those who want to go to the mat with the United States and those who want to make a deal. As of now, it looks to me like the faction that wants to double down on this is in the ascendancy. I think the same schism exists here in America, but it's not quite as rowdy. There are a lot of people in Iran who are willing to die to defend their country, and there are not nearly the same number of people in this country that are willing to die for rather fuzzy reasons. A scary but rather abstract fear of Iran getting a nuclear weapon does not compare with seeing an American or Israeli bomb kill your children. Trump has kicked a hornet's nest, and the way forward is unpleasant, no matter how you go. If the other Arab countries in the region were willing to go to the mat against Iran with the United States, that would be something, but I don't think they are. I see the danger now as turning a war of choice into a war of necessity.
I think it is rather telling, at least symbolically, that a young man who was beaten severely protesting the Iranian government when that was going on has now joined the army.
Yeah I have no idea about Iran. I meant most of the US. (I guess there are a lot of 'war hawks' in congress/ government.... we could vote those a-holes out of office.) TBH at this point I don't really care that much about Iran getting a nuke.
Thanks for the link, very interesting!
Holy cow, gas up 48% for the year! Never been happier to have a hybrid car that averages 350mpg...
OTOH we have an oil burning furnace so this winter is gonna suuuuuuck
" But I think almost everyone wants this to work out so I'm not that worried."
I don't think this is remotely true in any useful sense.
First and foremost, Netanyahu doesn't especially seem to want it to work out. The combination of Iran's commitments to Lebanon and Lebanon's proximity to Israel give him a near-veto on the peace process (though Washington can try to pressure him into not using it).
But second, I think Trump only "wants it to work out" in the most superficial sense that if you have him a button that would restore the status quo ante between the U.S. and Iran at no cost to himself, he'd push it. The war has clearly not worked out like he hoped it would, and he'd much prefer it stop being important.
But we can trivially see how little real will their is behind is "wanting it to work out," by the simple fact that it *hasn't* worked out. What was obvious to make it work out was obvious quite some time ago. Washington simply needed to sue for peace, meet Iran's terms and deal in good faith, recognizing the plain facts that a. the Americans and Israelis were inarguably the aggressors and b. the Iranians had the upper hand, and with it a large (but not unlimited) latitude to dictate terms. Sticking to the plain intent of the MoU as it was announced in June would have been sufficient, if Trump had been willing to do that.
It's hard to be sure exactly what happened instead, whether Trump intended to deal falsely from the start or whether he simply chafed at the (largely self-inflicted) humiliation and listed to bad advice about what to do about it. But either way, the ceasefire collapsed because
1. The deal was made in a way to be contingent on Israel's actions, without Israel actually agreeing to it. Israel naturally didn't consider itself bound by such a deal. [1]
2. Washington and its allies repeatedly tried to undermine Iran's negotiating position mid-negotiation, by routing ships along a route that Iran hadn't agreed to, one that would have been strategically advantageous for Washington if it became the main route. This was a pretty clear act of bad faith.
3. When Iran responded to the above with (admittedly provocative and belligerent) act of force, Washington immediately switched back to trying to intimidate and out-escalate Iran instead of making a more measured and restrained tit-for-tat response. This had the predictable consequences.
I think 3 especially reeks of poor impulse control and someone whose desire for things to "work out" all but disappears the moment their temper rises.
This, in a nutshell, is why I'm pessimistic for any actual near-term peace deal (and remained pessimistic even when the ceasefire was still holding). Succeeding at a complicated negotiation from a position of extreme weakness--a position the pretty much requires repeatedly showing restraint and eating crow--is simply not something I think Donald Trump has the mental capacity to do. I think he was scared to the negotiating table by dire predictions from somebody in his inner circle, but even when the dire predictions start turning into even-more-dire realities, he's just not temperamentally capable of the things that he needs to be able to do to fix this. This is, let us remember, a man who threatened the Iranian negotiating team with violence in front of the entire world.
I think peace is only likely once Trump is no longer able to sabotage the negotiations, either because he's no longer occupying the office of president, or because some combination of election results and mounting popular pressure force congress to grow a spine and reassert themselves as an independent branch of government.
[1] The U.S. probably has enough leverage over Israel to have made it stick anyway, but it would need to be willing to use it quite a bit more strongly than it has in the past.
Well Trump will be out of office for many years before congress grows a backbone. What's the incentive? We the voters must demand a congress with a backbone.
Anyway, by most I mean most americans, no one likes this war and we are done with the Trump lying. As I heard someone say, "All this winning is killing us, less winning please." It's hard to see how the war goes into the winter and home heating oil season. But I don't really know anything.
In an economic sense? Far, far worse than you've priced in.
200-300 million dead, worst case (this year only. If conflict continues into next year, expect worse).
Potentially a lot of fallen leadership, lots of countries changing hands, burnt governmental buildings...
Where does the death count come from? That’s a 3x multiplier on WW2!
Well, you'll want to start with the baseline: 8 billion people in the world, and you have, say, half a billion in some form of famine (this is, ahem, normal -- wars suck, and there are kids who die in India from malnutrition every year.)
https://carnegieendowment.org/emissary/2026/03/fertilizer-iran-hormuz-food-crisis
America is having drier months than the Dust Bowl, and wheat's not been planted (you can pull the year over year). Russia's also down on wheat planting, and Ukraine is at war, so there goes the Middle East's breadbasket (egypt is a net food importer nowadays) Now, the next risk is not having diesel/gasoline to harvest. And rice not getting planted.
You're looking at everyone on the edge dealing with "higher prices" -- that's Sudan, Yemen.
https://reliefweb.int/report/ethiopia/drought-east-africa-january-2026-gdo-analytical-report
And East Africa's got drought too. ENSO and the Iran War are doing a number on some of the least well-off places in the world.
Qatar and other countries on the gulf don't get to move food by boat, so they're stuck using gasoline (which is scarcer than you'd think, not all of these places have working refineries right now, Iran certainly doesn't), or flying the food in (which is expensive, even if they do have "oil revenue" buffers.)
Iran itself has 90 million people or so, and they're under "water famine" (can't grow crops if you don't have enough water to even support your capital -- and it's difficult to get any food into the country).
Again, this is worst case, but it's a pretty bad worst case. Israel is fond of saying "Israel has a right to defend itself" -- if it gets really bad, Israel may find itself pitted against Article 5. (Which is to say, Israel is allowed to "defend itself,"** so long as it's defending itself against international pariah states).
**Defined as clearing entire regions of Lebanon, and saying "the Lebanese will never be allowed to move back."
https://www.hrw.org/news/2026/03/23/israeli-officials-signal-stepped-up-atrocities-in-lebanon
Sidenote: just skimming, I mark about 40-50 million dead from famine in World War II (the differential is mainly from China, whose statistics nobody should ever trust). And yes, that's over four years.
Interesting failure of AI capabilities. I asked Fable to compose two verses of We Didn't Start the Fire for 2000-2004 and got this:
Y2K, hanging chads, dot-com crashing, Bush v. Gore,
Towers fall, anthrax scare, boots on Afghan soil for war
Enron books, iPod drops, euro cash in every store,
Wikipedia, "Shrek" is green, Ozzy's on TV at four
SARS outbreak, shuttle lost, Concorde makes her final flight,
Saddam's caught, "Rings" sweeps gold, human genome read outright
Facebook launched in Harvard dorms, Athens flame is burning bright,
Boston breaks the Bambino's curse, tsunami on a Christmas night
If you try to sing this you will notice it doesn't scan at all. My version for comparison:
Y2K, Bush v Gore, Tiger wins all four,
Richard Hatch, Concorde crash, Wikipedia
Anthrax, Dot Com, iPod, Yemen bomb
North Tower, South Tower, US in Afghanistan
DC, Shock and Awe, Euro notes, shuttle flaw
Facebook, Abu Ghraib, Red sox curse dead
Schwarzenegger, DNA, Janet Jackson, Boxing Day
Kill Bill, Spotless Mind, Ronald Reagan, goodbye
Interestingly, if you ask it in a new chat, Fable thinks its version is better than mine, precisely because it thinks its version is "more singable"! So it seems to do very poorly at understanding rhythm and melody.
This is the first time I've seen any lyrics for We Didn't Start The Fire that wasn't the We Didn't Start The Fire bit.
Feels like a link is appropriate. Although this video's quick cuts are hard on the eyes. https://www.youtube.com/watch?v=cDPnsTRAvIM
If you listen to the song and read Fable's and then yours, Fable's definitely fits better. Yours has like half the syllables of the original song.
Its not really about the syllable count. (I agree my lines are on average slightly shorter than the original, although the original does have 9 syllable lines "Moonshot, Woodstock, Watergate, punk rock" though only one and I have two here). Its about the placement of the stress. Try to sing Fable's line that starts with Wikipedia. It can't be done to the rhythm of the song; the 'a' sound at the end has nowhere to go. Notice how I put wikipedia at the end of the line, like Dimaggio in the original. I don't really know enough about music to say exactly why this is required, but it is. And even I, a non music expert, can hear this. The lines that start Facebook and Boston have the same problem. The emphasis doesn't start in the right places. Fable also got the rhyme scheme wrong. It did AAAA end rhymes instead of internal rhymes on the 1st and 3rd line of each verse, which are required to make the song sound right.
There's definitely places where the stress is gunked up, and "Wikipedia" sticks out like a sore thumb for having too many syllables to fit the meter without elision. I don't see the issue with the Facebook line, though?
Part of my job is writing, and I sometimes send people AI-written text for review without disclosing it, just to see whether it's good enough (it never is, even though it has been provided plenty of examples and explicit quality criteria).
Today, I gave two texts about the same topic to five "blind" human reviewers, one written by me, one by Fable. I also gave these texts to a separate instance of Fable as well as to Gemini 3.5 pro.
All human reviewers independently felt strongly that my text was better across all dimensions. And what made me share this anecdote was your last point: In my case, Fable and Gemini *also* both preferred the AI-written version! They explained very confidently why it flowed better, felt more natural, the examples were more relatable etc., when the human reviewers felt the opposite.
It looks like reddit has finally decided to make itself more-or-less inaccessible to those without an account.
First, the mobile version of the site started stopping you from reading it after some short amount of scrolling.
Then, they disabled old.reddit.com completely unless you're logged in, both for mobile and desktop.
You can still read new reddit on desktop to your heart's content, but they may well be planning something around this as well.
It really made me realise how much of my daily news and general mental stimulation I was actually getting from reddit, despite the fact that I hate it. I'm not sure where I'm going to go for my cheap mental stimulation in the future.
So reddit is definitely making themselves less accessible, but I was just able to reach old.reddit.com in an incognito window and go through a couple pages. What do you mean it's disabled?
Alright I guess it's still some kind of A/B test going on. It was working intermittently for me a week ago, but nowadays I just get "Log in to use old Reddit" every time I try.
Well hopefully every time you get that you are deciding to go to some other website so that the signal they get is that making that move would reduce their user numbers and have very low conversion to logged in users
I assume they have realised they are no longer a company that sells eyeballs to advertisers, they are a company that sells human-generated text to AI companies. a
I'm fond of Reddit. There's a lot of raw human truth in the answers, though of course mixed with the various forms of bullshit, And it is impossible to be the crassest person posting.
> It really made me realise how much of my daily news and general mental stimulation I was actually getting from reddit
I was in the same boat when Reddit jumped the shark by my own metrics, and I find Substack to be pretty good. It's like "smarter non toxic" twitter, but also with full length articles. And it's pretty good about clustering your feed with similar authors / people that you've already followed or liked.
You might also give hackernews and Mastadon / Fediverse a try. Mastadon is higher friction (you have to manually find nodes with channels worth following), and the content was ~10x - 100x less than Reddit or Substack back a few years ago when I was trying it, but there were some good things there.
I think Substack is benefitting from being new. It'll get corrupted like Reddit as it gains traction, because there's nothing about its business model that gives it innate resistance to brigades, echo chambers, and so on. It might not become a monoculture as long as its model encourages creators to intellectually own their spaces, but there's similarly no innate resistance to people leaning on Substack to de-platform this or that creator until most creators lean in one direction. It'll continue that way until someone pulls together enough capital to create the spiritual successor to Substack, and the cycle begins anew.
The Fediverse truly seems to have put effort into bias resistance, by contrast. AFAICT there's no Mastodon site "brand"; it really wants to sink into the background and bake decentralized influence over content into every module. Even if someone tried to enter the system and lean on people, it couldn't reach everyone.
Yeah, I think you are probably right, and it hurts, because I really like Substack and will hate to see it go.
On Fediverse, I think the main problem is just volume / content, is that still the case? It's a very "hackernews" and "old slashdot" type of crowd. Which is fine for me, those are in a sense very much my people, but that's like <1% of the population, and they just don't produce the diversity of thought, content, and viewpoint that the bigger platforms do.
I don't know how Fediverse fares on volume. This is a tricky question, because it's practically impossible to know. There's no central registry of "hey, I'm a Fediverse server, I'm run by Jan, and I host content about Croatian politics" or anything like that. By design. Some dude might put up a server and encourage other servers to advertise his, but they don't have to, and he might want to be unknown anyway. A server might be an online smoke-filled back room for planning criminal activity. An underground network might spring up, using the common protocol for convenience, but being organized as cells of no more than 3-4, constantly shifting location, and if authorities with lawyers or guns find one, the others pull up the drawbridge and re-organize on different IPs.
There are multiple crawler tools, but they're not guaranteed to find everything, because the Fediverse is not required to be completely connected. By design.
When Leftists do dynastic politics, they're embarrassed enough to pretend the dynasts have qualifications other than who they're related to. Trump sees no need to do that because he has a sleazy third-world mindset.
https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcTsZtGtlD8B9jqOXSO5uWrI8NrGYKuWQ49CI_DtzYLl4zWwB51JnGwMbeDe&s=10
Why not just link to the original tweet/Truth?
Either way the actual call here seems reasonable; it's an appointment for the next six months (this seat is up for reelection then), and Graham's sister is a plausible neutral figure to do roughly what he would've done who can be reasonably trusted not to run for a full term herself, letting Republicans run a neutral primary without the governor having to put a thumb on the scale (which he presumably doesn't want to for state politics reasons).
Slime Mold Time Mold has a post about different human (dietary and other) drives existing, but the number of them being unknown. This reminded me of a post by I think Ozzy, saying that the urge to eat ice was often a sign of iron deficiency. This sounds to me like we have a "consume iron" drive that sometimes misfires into a "consume ice" mode
Iron Deficiency leads to people eating dirt, or cornstarch.
https://scienceinsights.org/why-am-i-craving-corn-starch-pica-and-iron-deficiency/
Is there a way to rewrite a Substack page so that comments newer than time X are re-hidden behind "1 new reply" or otherwise highlighted?
for firefox there an extension called "acx-tweaks" by "Pycea", which does this kinda, i think.
for me it shows a "~new~" for all the comments that i havent seen yet.
NYTimes: "New York to Enact Nation’s First Statewide Moratorium on Data Centers" https://www.nytimes.com/2026/07/14/nyregion/new-york-data-center-moratorium-hochul.html
Remarkable that no matter how crazy and low-class the Republicans become, Democrats ALWAYS have to do something to piss off the nation's wealthiest people for no reason.
AI data centers are incredibly unpopular:
https://news.gallup.com/poll/709772/americans-oppose-data-centers-area.aspx
It's not just a Democratic phenomenon either. The Utah state senate president lost re-election over this one issue:
https://www.nytimes.com/2026/06/24/us/j-stuart-adams-utah-senate-data-center.html
All of the data center companies should come up with a national plan to solve this. My proposal would be to pass a federal bill which requires data center companies to fully take on the cost of energy (either building their own or paying utility companies enough so that rates don't increase for civilians), noise limits at a certain radius which can be addressed with sound baffles, and a formula for the allowed annual water usage (based on reserves, community usage and other relevant criteria). In exchange, they get permitting reform to streamline the approval process for building, and acquiring energy, and a federal pre-emption of state level bans.
Since data centers are not in fact the cause of higher electricity prices, those policies will not work:
https://x.com/CityJournal/status/2077092838630735935
That data ends in 2024. In any case, perception matters more than reality in politics. Currently, 70% of Americans disapprove of AI data centers, and 66% cite electricity costs as a reason. Maybe you think that is because people are irrational and stupid. And yeah they are. But you can't exactly change that, so what do you do when you're at the mercy of the politicians that they are about to put into office?
I don't know if some token legislation addressing people's stated concerns will help (when the reality is probably anxiety over job loss, unrelated to actual facts about data centers), but it wouldn't cost that much and it could even be beneficial on the net if it speeds up new energy projects. And it seems to me that the alternative is probably new data centers getting banned in most of, if not the entire country.
What is the relationship between data centers and electricity prices? Honestly asking.
It would be counterintuitive to me if there were none, having worked in both an electrical utility and data center construction, and I’d want to make sure I understood why if that were the case, but I’ve been surprised by other things before.
Depends on the local electricity supply.
If there's consistent excess generation capacity, then utilizing the local grid more intensively could reduce average cost (i.e., more variable charges to amortize fixed costs across). Similarly for any other situation where the existing grid scale is below the cost-minimizing size. Residents should welcome DCs.
If there's an average excess but intermittently binding capacity constraints, it'll depend on how the utility is allowed to charge. A possibility would be that the DCs draw a steady load and pay a flat rate that includes guarantees for prioritization, which would leave the preexisting users facing a more volatile remaining supply (e.g., lower average prices but with unpredictable spikes, and/or rolling blackouts). Mixed bag for residents, depending on their resources for absorbing volatility.
Then there are areas where the grid is roughly fully loaded. Without additional investment the increased demand from DCs would just drive prices up. Residents likely to be adversely affected, unless a competent & benevolent bureaucrat or politician (good luck with that) negotiates a positive-sum investment in the grid by the DC builder.
Thanks, that's what I wanted to know.
So, how do you know what kind of area you're in, and how do you know how much of the relevant type of effect data centers are having on your electricity prices?
I doubt that, even if the data were available, the general public would be able to evaluate it.
Which leaves only trusting experts or "experts" with the attendant difficulty of distinguish between those.
There's a kind of catastrophic cycle where people believe that a policy that only makes a problem worse will solve it, thus doubling down on it more the worse it becomes in a vicious cycle. The classic case of this is rent control (also recently doubled down on by New York politicians), but seems like energy prices are going the same way.
Pissing off the nation's wealthiest people is a reason in and of itself. It's fighting the good fight, and Democrats would be a lot less lame if they'd do it more. The U.S. seems to have this bizarre collective Stockholm syndrome where it watches small numbers of people act like idiots and children and assholes and says to itself, "obviously these people must be amazing because they have money," and contorts all it social and political life around trying to idolize and please and emulate them.
I think AI data centers specifically are an excellent fight for Democrats to be fighting[1]. AI is a suite of technologies that could be both incredibly powerful and incredibly dangerous, and the big AI companies have done very, very little to inspire confidence that they'll handle it responsibly or use it for the public good. Having the public tell them "no," occasionally and seeing how they react to not getting absolutely everything they want is pretty much the bare minimum that a functional polity ought to be doing about it.
[1] Though that being the case, I fully expect them to roll over or half ass it as they always do.
Seems to me people don't idolize all wealthy people just for having money, but instead they idolize certain wealthy people (for being the big heavy hitters on their team) and demonize others (for being the big heavy hitters on the opposing team).
Like, when it looked like Elon Musk was buddying up with Trump and doing Hitler salutes and stuff, people who liked Trump liked Elon Musk. But those same people hate George Soros.
It's all based on tribal identity, which is what I always say, because (at least as far as I can tell) it's always the best explanation.
If a wealthy person wanted a free pass to build data centers or whatever else without Democrats bothering them, they'd simply have to build a public identity as being very left-wing or otherwise Democrat-aligned, maybe with a little extra edge to demonstrate they're not just lukewarm about it. Then everyday people who vote D would support everything this person does and says, especially if it means defending them from criticisms by people who vote R.
If I could put all billionaires to work running hotdog carts on the weekends, I totally would.
https://www.shystersdogs.com/
(was this the guy in the Lime Skittles are back "apology commercial"?)
That'd be interesting. If you force the billionaires to actually be hands-on and man the cart themselves, and deny them the right to just turn it into a normal business they hire other people for and run from an office, it could go one of two ways:
1: Billionaire thinks "Well, I've got the cost of my overhead (cart, ingredients, payment system, personal security detail, etc.) and my labor (thousands? tens of thousands? of dollars per hour)" and they price that into the hot dogs, make no money because nobody is willing to pay $9,999.95 for a hot dog, then write the whole thing off on their taxes so we all end up subsidizing it
2: Billionaire thinks "This will look like a publicity stunt, like that lawyer who opened a hot dog stand...I can afford to give away hot dogs as part of my stunt, so I'll just do that" and it basically becomes a "line up to shake hands with a billionaire and get a token hot dog you can eat or throw in the trash afterward" photo op kinda thing
I'm honestly not sure which outcome is more likely, but now I'm hungry
"Seems to me people don't idolize all wealthy people just for having money, but instead they idolize certain wealthy people (for being the big heavy hitters on their team) and demonize others (for being the big heavy hitters on the opposing team)."
There are people who think like that, yes. But you seem to be mistaking "some people think like that" for "all people think like that."
"If a wealthy person wanted a free pass to build data centers or whatever else without Democrats bothering them, they'd simply have to build a public identity as being very left-wing or otherwise Democrat-aligned, maybe with a little extra edge to demonstrate they're not just lukewarm about it. Then everyday people who vote D would support everything this person does and says, especially if it means defending them from criticisms by people who vote R."
As much as I hold large swathes of the Democratic Party in utter contempt, this is just demonstrably not true. Democrats are not fantastic at holding themselves to real principles, and I wish they were better at it. But they do it more than zero. In fact, in the last two years they've done it in numerous cases of the highest profile imaginable. Now the Democratic leadership has been disgustingly much better (in at least some cases) at closing ranks and holding the party line against their own voters than they usually are against Republicans (though that wall finally seems to be crumbling). And I've definitely seen and even talked to *some* rank-and-file Democrats who do think this way, especially older ones. But in coalition-based electoral politics you need almost all of the coalition to think this way almost all of the time, and again, there's been strong evidence over the past few years that this just isn't true of Democrats as a coalition.
It does seem to be true of Republicans right now, as much as I wish it weren't. I honestly do not know what real-world actions it would take for the Republican base to turn on a public figure who consistently parroted the right tribal messaging. But we already know that neither protecting pedophiles nor blatant corruption, nor actual treason are sufficient cause, so the existence of any line seems speculative at best.
My impression is not that *all* people think like that, it's that the vast majority do, enough to where I could just write "people think like that" and it's basically true. But I know this is only my impression and I could be wrong.
>Having the public tell them "no," occasionally and seeing how they react to not getting absolutely everything they want is pretty much the bare minimum that a functional polity ought to be doing about it.
They might react by giving money to J.D. Vance to make him President of the United States. Why do Dems have this self-destruct instinct?
You don't think Republicans have the same self-destruct instinct?
"As long as it feels like it punishes my outgroup now, I'm in favor of it, and let's not think too hard about what it does for my ingroup later" seems pretty common across the board
Voter's values are not that malleable. A massive political campaign can shift the outcome on issues where public opinion is split or people don't feel strongly. But not on a hot button issue with a 73-27 split (it'll probably be even more dramatic by 2028).
Vance will either have to take a more hardline stance against AI, or lose the election.
> Pissing off the nation's wealthiest people is a reason in and of itself.
Seriously? You know ~75% of billionaires are self made? That they've directly created ~10M jobs, and that's just direct jobs, not any of the many millions more enabled by being their suppliers, contractors, customers, or affiliates?
That "tech oligarchs" are the source of essentially all US economic growth over the last TWENTY years, not just the 7 most charts show?
https://imgur.com/a/UOJI1Q7
How dare those scoundrels create tens of millions of jobs and drive nearly all the economic growth in the richest and most successful country in the history of the entire world!
Why, we'd be SOOOO much better off if they'd never existed at all, and we were tens of millions of jobs and many trillions of dollars poorer, and didn't have Amazon, Google, Netflix, or Walmart!
Oh, and all your political heroes have gotten rich off of them, too. Democrat politicians reliably get higher insider trading returns than Republicans, clocking in at 40%+ returns per year vs the R 29%: https://imgur.com/a/u5D9C9B
That might tell you something about their actual priorities and alignment, if you paid attention to their incentives.
Are you reading the Forbes list again? Most billionaires are not listed there. No, most billionaires are not self-made. Most billionaires are nobility, which tends to take exception to being listed on Forbes.
I'm all for saying "American Billionaires" tend to be better, but...
Kochs aren't "self-made", Scaife wasn't self made (and SBF was -definitely not- self-made). These are billionaires (and heavily involved in politics).
(I'm also not headhunting Musk or his ilk -- some people really do get there by hard work and honest effort).
> Are you reading the Forbes list again? Most billionaires are not listed there.
Yes, I actually wrote an article about how many billionaires there really are worldwide.
And you can do top-level economy triangulations with the power laws / incidence seen in the rest of the world as a check to see that there are some really hilarious mismatches. I ended up at a little over 2x the listed billionaires.
https://performativebafflement.substack.com/p/3k-it-is-to-laugh-just-how-many-billionaires?r=17hw9h
As you may have seen from some of the comments, Democrats include a large number of people who hold that pissing off wealthy people is a virtue, on faith. There are multiple reasons for this, and they're well-traveled reasons: failure to value management, failure to value networking, failure to understand price theory, sheer envy, etc.
Democrat party leadership includes a large number of those wealthy people, and some of them _do_ understand price theory and management and networking (not to mention how much it actually costs to run a political party), and they probably find themselves riding a line between courting the people who don't and not pissing off the ones who do. Their strategy seems to include trying to make some sort of distinction between "good wealthy" and "evil wealthy", but that's an epicycle, whether they realize it or not.
It's tempting to embrace eschatology at this point and hope for a time when they all figure out where this leads. I prefer the more boring option of continuing to explain, to whomever will listen, since it's based on the premise that most people are smart enough, and deserve the incentive to lean into that. Plus, you never know when they'll have an insight of their own that you missed.
Hate for AI data centers is bipartisan at this point, and decisively so. At least among voters, which will quickly transfer to elected officials after one or two election cycles. I do not understand why data center companies aren't treating this as a threat to their continued existence.
Anyone have recommendations for a rationalist-adjacent counselor licensed in New Mexico? Telehealth is fine, and second-degree referrals (some you know/your therapist has a rec for someone in NM) are also welcome !
If by rationalist-adjacent you mean a practitioner of authentic CBT, the place to look is National Association of Cognitive-Behavior Therapists.
Being smart is not as fun or exciting as people imagine. You are used to being smart, so your mental content is no big thrill, just the regular taste of spit in your mouth. It doesn't feel special at all, it just feels like living and thinking. Whereas you never get used to failing to fit in. It's freshly bad every goddam time. Know what I mean, honey pie?
Getting smarter, on the other hand, feels amazing
Sadly those days are behind me
Yeah, I'm in my late 50s, so I don't forsee my mental sharpness getting any better, though I do a very intellectually demanding job and I don't seem to be hitting "can't do that anymore" or "can't learn that" limits yet.
Same, and I'm in my mid-50s. There's more and more to learn and improve at, and the bottleneck I feel is finding the time. Sometimes I even get annoyed at the fact that I have to stop and eat something, or sleep.
Yeah, being dumb wouldn't be so bad if I couldn't remember being smart.
Yeah, haha, I drank 2 beers which made me dumb enuf to post what I did last night, and people are showing way more interest in what I said than they do my serious posts. Have all you guys been drinking beer too?
You've got to participate in internet discussions with one hand tied behind your back, just to keep it fair!
Sometimes dumb posts make the smart posts more convincing. Proves you're not a bot, y'know.
I hate the scenario where the smart people can't fit in, so they get dysfunctional, and when they later meet other smart people and finally have a chance to fit in, there is too much accumulated dysfunction that prevents a reasonable outcome.
Mensa is the textbook example, but to a smaller degree everything.
I wouldn't know
But then you go to a rat meet up and meet just the right kind of people and it's like taking a shower after a long hot day.
At least you can listen to this song thinking it's all about you:
https://www.youtube.com/watch?v=Me-opLaRJ5w
Your mileage may vary. In my experience, being smart can be extremely satisfying or endlessly frustrating; what makes the difference is the opportunities available to you.
Hi, If anyone is interested, I read this great article yesterday:
https://substack.com/@becomingnoble/note/p-206981864
Not mine, so I hope the plug is OK.
TDLR:
The middle class might look OK in raw economic data, but actually lost material non-financial assets, and replacing those in the Free-Market is very expansive, so in fact the middle class is Poorer today than our parent's generation, and because the lost assets were not financial, it is not reflected in Statistics.
TLDR is still true but your summary intrigued me. I read it and thought, "Oh, like how every middle-class family used to have a piano in the living room, and now maybe one in ten has an old cheap keyboard in a closet in the basement somewhere, and even getting a good Yamaha keyboard with weighted keys is probably not feasible for them because they're all drowning in credit card debt and what do they even want one for because they have Spotify subscriptions?" But I have no idea if that is a good example of what the article is talking about.
Can you provide an example?
The examples the article provides are:
1) Public Schools at most places were "Good enough", no need for private schools.
2) Kids could walk around, for example for School, or to playground. No need for big Yards or driving them around, or for so many extra-curricular activities.
The examples the article provides are:
1) Public Schools at most places were "Good enough", no need for private schools.
2) Kids could walk around, for example for School, or to playground. No need for big Yards or driving them around, or for so many extra-curricular activities.
The public / private school thing seems misplaced to me, as someone who believes private schools are superior _if_ parents are not forced to pay for public schools. In other words, a system where parents can either pay a set rate and use a public school is worse than one where parents can pay to use the private school of their choice, since they get to navigate the quality-price tradeoff themselves. Lamenting private schools strikes me as lamenting that everyone has to choose their own flavor of ice cream when vanilla used to be good enough for everybody. (Mandatory taxation for public schools confounds this.)
The complaint about kids' freedom to roam seems legit and sympathetic to me.
Overall, some people disagree that the upcoming generation has it worse off, but the people who agree seem to come from both sides of US politics, so that's something. That said, there's plenty of disagreement on a remedy.
Do you have kids in school?
A society where only children from well-off backgrounds get good educations is one I expect to descend into a gilded-age dystopia within a generation or two.
Parents are not the sole beneficiary of children's education. The children themselves benefit, but rather more important than that for this context is the society itself benefits. And most especially, it is the children of the parents who are the least concerned with getting their children a good education that we as a society should most badly want to have good educational opportunities. No altruistically, for our benefit. Selfishly, for our own.
A society where only children from well-off backgrounds get good educations is one I expect to quickly evolve into one where most children get them due to their parents seeing the obvious benefits to their own status, not to mention the benefits of having a child who goes on to a highly-valued and therefore high-paying job. And also due to a lot more people seeing the benefit in supplying good education to parents who demand it enough to be willing to pay, and hiring themselves out as teachers.
The real challenge here comes if the state somehow decides it's worth gatekeeping such educations or insisting they be universal even if they end up universally bad, and manages to persuade 51% of voters that that bad education is actually good (or, perhaps, better than the alternative).
I don't *really* know if the thesis he proposes is correct.
It does say something that sounds sort-of-true (Lost social capital), but I can't really verify, and provides an explanation for a known problem, namely that People feel poorer today, but this isn't reflected in the statistics.
I am, however, sympathetic to his arguments. If you read the article, you see he mention other things - like the breakdown in public trust drove people to more expensive neighborhoods as a way yo filter the "unwanted".
I didn't provide his entire argument here, just a short summary.
Almost any child is a service his parents provide society at large from the goodness of their hearts.
I take his point as being that the new generation is worse off in ways that don't show up as numbers. He's correct, but he's not the first person to point that out. I'm also pretty sure he's not the first to point out those specific issues (home prices, marriage rates, etc.). It's probably worth collecting some of those issues and showing them to yet another audience, but that's pretty much all this is.
It'd be interesting, extra-credit effort if he presented a workable remedy.
Maybe the new generation is worse off only inasmuch as having everything be better and easier means you don’t build as much character
Shouldn't Open Threads be ordered newest to latest?
Scott, do you set that manually for each Open Thread?
This appears to be a Substack (mis)feature, not an ACX feature, let alone a feature of OTs. When I read an OT and click on "new replies", those replies may open at the top or at the bottom, from one OT to the next.
(I find it frustrating.)
With very few exceptions, OTs on ACX have been sorted newest first. I think the Substack owners set this but don’t know for certain. Might be something that occasionally slips Scott’s mind.
I noticed that the "report comment" functionality has changed. Previously you just wrote a text... and it was reported to Scott? Now you have to choose one of the provided options, and it is reported to a Substack team.
The provided options do *not* include "created yet another account to avoid a ban".
I noticed this a few days ago as well. There also isn't an option for "violated the norms of this specific substack instance, regardless of whether it follows general Substack rules".
In theory, "ban evasion" is a general thing that would make sense to add to the general Substack rules...
With other local norms, yeah, it would be difficult for an outsider to evaluate them. Substack should keep the old functionality together with the new one, so that global rule violations would be reported to global support team, and local rule violations to the blog owner.
If Wimbli ever sets their user name to Wavy Gravy i might not block them for a while.
https://en.wikipedia.org/wiki/Wavy_Gravy
It would be good for a chuckle anyway.
Account spam is Spam.
Well this should be interesting seeing as how Trump recently piped up, apropos of his own unrestrained id, to tell our NATO allies that the 2020 election had been stolen.
“Trump is expected to make election conspiracies a focus of his national address”
https://www.startribune.com/trump-expected-to-make-election-conspiracies-a-focus-of-thursdays-national-address/601868622?utm_source=gift
(No paywall)
Not sure if this is appropriate to essentially advertise here, but I prototyped a star convex 3d model encoding scheme and shader for rendering the encoded geometry with a single quad. I wrote about it here https://joem93.substack.com/p/3d-spherical-model-encoding. I usually just don't tell anyone about anything I do probably for fear of failure, criticism, etc, but in this case my desire to see what people do with this concept is stronger. If you know someone who might be interested in this, consider passing it along. I *have* to imagine this would be useful in some context within videogames
The idea is: you encode your 3d model by recording its surface colors and distances from a center point onto two images; then render it by raymarching each pixel, transforming your ray points to spherical coordinates, and sampling the radius texture using the angular coordinates. The first time you sample the model radius to be greater than your ray point radius, you sample the color texture and set the pixel's color to that color. I've been kind-of thinking of it like an upgraded (and significantly more expensive) billboard rendering scheme
My implementation is still a little under-baked because it doesn't calculate depth values. I guess that's an exercise left to the reader (and/or future me). There's also other issues like visual discontinuities due to raymarch step sizes (I've since learned you can apply a little random noise to step sizes to get slightly blurrier but more visual appealing rendering), and poles creating pinch-points in spherical coordinates. I really like the idea of using other maps from R^3 to R^2 but I haven't investigated it yet. Theoretically, if you find such a map (any such map), it would work just fine as long as it isn't lossy wrt your model's surface (ie when multiple surface points map to the same value). In the best (and probably most impossible) case you would directly map points to arbitrary UV coordinates, but that would probably take training an ANN or something, and almost certainly wouldn't be faster than the spherical version. Perhaps a cubemap map would work
You know, it's just now occurring to me the color texture alpha channel could be used to encode radius values, instead of using a whole other texture. You would lose transparency, but I'm not sure how transparency would even work with this
A purely anecdotal report on using coding AI’s to find difficult bugs…
The bug was that a buffer that was used as shared memory between two threads was declared on the stack, so that if the thread that declared the buffer finished first, the other thread could overwrite the stack.
Well done DeepSeek for finding that subtle stack corruption attack. (I imagine the other coding AIs would also have found it’
Does anyone know if there's a way to turn off 'also replied' notifications for Substack? I like getting notified when someone replies to me, but 20 notifications saying 'the comment section has people commenting in it' feels like worthless spam after awhile.
Transformative AGI by 2043 is <1% likely
By Ari Allyn-Feuer and Ted Sanders
https://arxiv.org/pdf/2306.02519
Has this been discussed here, including by Scott? What do you think?