I only started reading ACX after the move from SSC. I have gone back and read the top posts from SSC but still felt like I was missing out on some of the old classics. I built a website to resurface old content from blogs by sending weekly emails. Let me know if you have any suggestions of other blogs / content you’d like to see, hope you find it helpful!
Some suggestions (note that I endorse the first the most):
- Overcoming Bias (by Robin Hanson), which has been around for ages and is very rationalist-relevant (it was in fact the birthplace of the Sequences!)
- Marginal Revolution (by Tyler Cowen and Alex Tabarrok)
- Wait But Why (by Tim Urban)
- Ribbonfarm (by Venkatesh Rao)
Also, I'm happy that you've made this website in the first place! I am subscribing to read the Slate Star Codex archives, since I considered reading them during the SSC-ACX gap (June 2020 - January 2021) but never really made it all the way. Also, I suggest that you make the "Success" dialog more prominent.
Thanks for the feedback, just updated it with your suggestions! Will send out first posts next week. Currently working on allowing for more granular control of which posts you want, eg only send posts in past year, or only posts of type x, or send y posts in a week, etc.
Kevin Drum commented on this, and I have to agree with Kevin's assessment...
"Let me say from the start that I think this is kind of nuts. I'd personally put the overall risk of extinction from everything combined at about 0.1% or so. But that's just me. I'm an eternal optimist, I guess."
>Apparently there's a group of AI experts who are absolutely obsessed with extinction and have spent more than 1,000 hours thinking about it. These people seem to have thought themselves into a frenzy and they're obviously driving up the average a lot.
Wow, I wouldn't have come up with that interpretation if I tried really hard. We should trust people who spend more time on things more than people who spend less time in general. Otherwise, our way of doing science and policy implodes!
Just want to flag that the report highlights a divergence between "domain experts" and "historically accurate forecasters", which goes like this:
1. The median expert predicted a 20% chance of catastrophe and a 6% chance of human extinction by 2100. Superforecasters saw the chances of both catastrophe and extinction as considerably lower than did experts. The median superforecaster predicted a 9% chance of catastrophe and a 1% chance of extinction.
2. The gap between forecasts of experts and superforecasters was not uniform across topics. Experts and superforecasters were furthest apart on AI risk—and less so on the risk of nuclear war.
Asked this question in one of the threads, but I'm also curious about what the broader board population thinks.
Who do you consider a "AI Doomer?" If one person says "AI development should be banned" and another person says, "Regulators should require AI products be subject to testing standards before deployment - and periodically thereafter - to confirm they work as intended, and should require developers to ensure that AI-generated outcomes are explicable" are they both "Doomers?" Just the ban proponent? Neither (since neither of them are stating a belief that AI is going to kill us all)?
I'd personally limit application of the "AI Doomer" to perspectives along the lines of "humanity will go extinct if we don't ban AI," but a lot of the pro-AI circles also trend pretty libertarian, so I think the actual usage may be be more expansive, in much the same way that some people in our corner of the world consider zoning to be "tyranny," "food regulations "basically fascism" and so on.
It's in the name. "If ignored, [Thing] could be our doom." Anyone who thinks unchecked climate change will wipe out humanity is a doomer. Anyone who thought an unchecked Y2K bug could lead to societal collapse was a doomer. Any Pompeiian who was worried about Mount Vesuvius erupting was a doomer.
The difference between a Doomer and whatever the alternative is is whether they think something can be disastrous, or just unpleasant. (I guess the second ones are Gloomers? I'm calling them Gloomers.)
Is it "if ignored [thing] could be our doom" or "if ignored [thing] will be our doom?"
Probably just a typo but the climate change example used a "will" and the Y2K example used a "could" so I wnated to clarify.
I'd challenge that a doomer should have to breach some threshold of belief that the doom *will* happen. Otherwise anyone who acknowledges the existence of any X risk at all would be a "doomer" about it, simply on the basis of the acknowledgement. Which would make almost all of humanity "nuclear proliferation doomers" or "large asteroid impact doomers," etc.
Probably a matter of taste. I draw the line at "could". If you think civilization is at any level of risk, you're a doomer.
And yes, almost all of humanity is a nuclear asteroid doomer. Although I guess we could add a second required layer of "we have to do something"; most people think there's nothing to be done about nuclear asteroids, and don't spend any effort on it. But I'd say they don't get called doomers just because nobody dislikes their opinion enough to do so.
For me, a Doomer is someone who puts AI x-risk within the next century at higher than one in a million, more or less. I'm not in favor of any AI regulations at this point, but I'd reserve the Doomer label for those who focus on the supposed threat of human extinction by AI.
>> For me, a Doomer is someone who puts AI x-risk within the next century at higher than one in a million, more or less.
Wouldn’t that render many AI advocates themselves *also* Doomers? “Unaligned AI is an x risk” isn’t exactly an uncommon perspective in those circles, at least from what I’ve seen.
In that case someone who (a) advocates for AI research and (b) believes “unaligned AI is an x-risk so alignment is important” would nevertheless be an “AI Doomer” if they thought the chance of an unaligned AI nevertheless slipping through was 1/500,000.
Granted, definitions can be whatever people want, but it seems like an odd usage if I can simultaneously advocate for AI and be an “AI Doomer”
Yeah, I would consider anyone who considers alignment important to prevent x-risk to be a Doomer.
Now, I'm attempting to answer your OP, but I don't think it's particularly important to define who is and who isn't one. I use the term Doomer below because it seems like the online world I read is splitting into two camps with, say, Scott and Zvi on one side and Tyler Cowen and Marc Andreeson on the other.
>it seems like an odd usage if I can simultaneously advocate for AI and be an “AI Doomer”
Maybe I'm wrong, but I get the feeling that most AI Doomers are Futurists who are or were very pro-AI but who view Alignment as the big filter to get through. Some of them have become so pessimistic that they might not be pro AI anymore. I don't think many (any?) of them are outright Luddites in general. Genuine anti-AI Luddites are more worried about losing their jobs tomorrow (For example, The Writers Guild of America). Or people worried that the AI is going to steal the election somehow. Those people are not Doomers because they aren't worried about x-risk.
Doomer is an arbitrary term—and functionally it's an ad hominem attack. And as you pointed out the goalposts for doomer are likely to move depending on the opinions of the person with their ax to grind. I think a better question would be: what percentage of people that there's a significant chance that AI will cause an extinction event for humanity?
As we've seen with COVID, the term "COVID minimizer" has migrated from the GBD folks and their ilk all the way to encompassing T Ryan Gregory—who up until recently was one of the leading COVID and infectious disease alarmists on Twitter.
Doomer is an ad hominem attack in that it appeals to feelings or prejudices rather than the intellect. It's used to negatively stereotype a person or group of people rather than a way of directly engaging with or addressing their arguments.
Doomer may not be a particularly nasty label, but it's a label all the same. Supposedly, we're all rationalists here, and we're supposed to argue things rationally rather than emotionally. I say "supposedly" because I've seen a lot of people on these threads confuse emotional arguments with rational discussion. While I admit I haven't been a saint on some of these threads—having let my sarcasm get the better of me sometimes—I've tried to focus on arguments and data rather than attaching dismissive labels to people. The use of ad hominem labels is an unsophisticated method of arguing one's point—at the risk of being ad hominemy it's sort of juvenile. I think you (and we ) can do better than fall into that trap.
But having said that, I think a lot of AI experts vastly overrate the threats of AI. And I've never heard a logically consistent scenario that would lead to the extinction of humanity. They seem to be channeling their fears rather than working through scenarios.
Some people from Metaculus have asked for clarification on some of my 2023 prediction questions. In order to say everything as publicly as possible so nobody gets an advantage, I'm putting it here:
>> 1. Will US CPI inflation for 2023 average above 4%? Are you comparing CPI at the end of 2023 vs the end of 2022, the average CPI inflation across 12 months (CPI Jan 2023 vs CPI Jan 2022, CPI Feb 2023 vs CPI Feb 2022, etc.), or something else?
Let's say https://www.bls.gov/cpi/ will be the canonical site. If they don't update in time, it will be some other measure of year-on-year inflation.
>> 2. In 2023 will there be more than 25 million confirmed COVID cases in China?
You mean happening in 2023 right? Not if the total number will reach 25 million in 2023.
Yes.
>> 3. In 2023 will Google, Meta, Amazon, or Apple release an AR headset?
Is the definition of AR that the user can see through it, i.e. that the screen is transparent, and therefore MR headsets where the user can see the environment through cameras don't count?
I will defer to Dan Schwartz from Metaculus, who wrote this question, but if he doesn't weigh in, then my interpretation is that any headset which allows AR capabilities counts, including if that happens through projecting the outside world onto an internal screen through cameras.
>> 4. In 2023 will a cultured meat product be available in at least one US store or restaurant for less than $30? Could it be a special offer that lasts for a limited amount of time? Do giveaways count? Does imported meat count? Does partial cultured meat count and if so is any percentage ok? If it's part of a menu do we divide the cost of the menu by the number of dishes or does the whole cost count?
I'm going to rule that promotions where it stops being available later won't count, but that if it's available just like any food, and some sale of the same sort that other foods sometimes get causes it to be on sale for below $30, I'll allow it. Imports are fine. Partial cultured meat is a good question; I'm going to rule that animal meat+cultured doesn't count, but plant-based-meat+cultured does if the cultured significantly changes the taste or texture or something. I'm going to rule that you have to be able to purchase it with less than $30 in your pocket, so if it's part of a package meal that costs more than $30, it doesn't count.
>> 5. In 2023, will SpaceX's Starship reach orbit? Does a transatmospheric orbit count or does it have to be at an altitude of more than 100km?
I'm not an expert on the definition of orbit. Google suggests an orbital spaceflight is one "in which a spacecraft is placed on a trajectory where it could remain in space for at least one orbit." I will consult with people who know more about orbits but try to generally use this definition.
>> 6. Will the Shanghai index of Chinese stocks go up over 2023? Are you referring to the SSE Composite Index? Because there are others, such as the SSE 50.
Yes, SSE Composite.
>> 7. On January 1, 2024, will an ordinary person be able to take a self-driving taxi from Oakland → SF during rush hour? Is the person being selected from a waitlist ok? You're saying "self-driving taxis in SF only operate at night" in your clarification, but these are waitlist only (as per a commentator).
I think the plain text of the question gives me one hour from the time I start worrying about this to make the trip, so if I have to worry about waitlists more than one hour before, no.
>> 8. In 2023 will a successful deepfake attempt causing real damage make the front page of a major news source? What's a major news source? E.g., would this count? What about TV news shows or radio programmes or news sites in general that don't exactly have "front pages"?
CBC seems major. I would count anything linked from the front page of their website, ie cbc.ca. Realistically this is a bad question because of articles like this where the deepfake itself isn't of major interest but is being used to make a wider point about deepfakes. I will probably be forced to count it anyway.
“ >> 5. In 2023, will SpaceX's Starship reach orbit? Does a transatmospheric orbit count or does it have to be at an altitude of more than 100km?
I'm not an expert on the definition of orbit. Google suggests an orbital spaceflight is one "in which a spacecraft is placed on a trajectory where it could remain in space for at least one orbit." I will consult with people who know more about orbits but try to generally use this definition.”
100km is the internationally accepted minimum altitude of space, and you more or less cannot remain in a stable orbit lower than that. In any case I doubt you’d event try to put Starship on an “orbital” trajectory there so it’s probably moot.
Obviously the Starship actually completing one or more orbits would seal the deal.
For a suborbital flight, I would say you should only count it if the flight is on an orbital trajectory at some point in the mission but intentionally deorbits via retro burn.
Maybe you could count if the rocket follows a full mission profile but (intentionally) does not go orbital.
What do I mean by this? Most often, launch vehicles going to orbit fire their engines multiple times - first a boost to apogee and then a restart and “circularization burn” to raise the perigee into a stable orbit.
Now a so called “direct insertion” where you do only one burn is possible (though less flexible) and I believe Falcon 9 uses this at least some of the time. That said, to do all the things Starship is supposed to do, it definitely needs the capability to restart its engines (probably multiple times) in space. This is not a trivial problem, so I would argue it needs to be demonstrated before you really call the rocket “orbital”.
If you really want to be lenient I guess you could count a launch where they achieve a successful full duration burn on stage 2 but it stays suborbital intentionally. But I don’t like that and I don’t think the planned profile of the test flight would have counted.
I'd also probably count it if it were on a trajectory with an apogee above 100 km and a perigee no worse than -100 km; that would demonstrate performance close enough to orbit as makes no difference, but give them a guaranteed reentry over open ocean if everything goes dead half an hour after launch.
>> >> 3. In 2023 will Google, Meta, Amazon, or Apple release an AR headset?
>> Is the definition of AR that the user can see through it, i.e. that the screen is transparent, and therefore MR headsets where the user can see the environment through cameras don't count?
>> I will defer to Dan Schwartz from Metaculus, who wrote this question, but if he doesn't weigh in, then my interpretation is that any headset which allows AR capabilities counts, including if that happens through projecting the outside world onto an internal screen through cameras.
Agree with Scott, we should interpret "user can see through it" to refer to the capability, not the literal transparency of the material. Mixed Reality (MR) devices that do this with cameras would count.
This is a bad question because if MR headsets count then it should have resolved before it was even asked thanks to Meta’s Quest Pro, which released in 2022.
Following on from the "Americans created 'Italian' food" piece, here's Max Miller of "Tasting History" re-creating the Pompeian 'pizza' from the recently discovered fresco:
I'm re-reading Albion's Seed for the first time in a few years. This passage triggered my Gell-Mann amnesia alarm bells:
"The Puritanism of eastern England was not all of a piece. Several distinct varieties of religious dissent developed there, each with its own base. A special strain of religious radicalism which put heavy stress upon the spirit (Antinomianism) flourished among Puritans in eastern Lincolnshire. The more conservative and highly rationalist variant of Calvinism (Arminianism) found many adherents in London, Middlesex, and Hertfordshire."
I'm no expert on the different strands of the Reformation, but I've always understood Arminianism as being diametrically opposed to Calvinism, not a variant of it.
Maybe I'm wrong and Fischer is right? How much should this make me distrust the book as a whole?
I think the confusion is that both Calvinism and Arminianism arise out of what were (are?) the Reformed churches of the time, and that it's debatable if Arminianism was a movement within the broader Calvinist grouping or something entirely separate.
They did eventually come to an official breaking apart and setting up as opposed camps, but within the specifically English context of the later Dissenters, probably they're best viewed as a strain of Calvinism or at least a breakaway group of Calvinists.
This is all pure ignorance on my part because I can't keep the finer points of which Reformed churches are what between France, Switzerland, the Netherlands etc.
Diarmuid MacCulloch's "Reformation" is helpful here; the Dutch split arose out of the conflict between the hardline Calvinists and those led by Arminius who had been trained in Calvinist orthodoxy but was questioning the theology. The English version were something different, at least according to MacCulloch:
"Many clergy and laity in England would have been horrified to think of themselves as Puritans, but they still regarded their Church as part of the international Reformed consensus; they were highly distrustful of the new tendency in divinity, and the new Archbishop Abbot shared their distrust. During the 1610s they gave avant-garde conformists a foreign name in allusion to the disputes about predestination that were currently tearing apart the Dutch Reformed Church – ‘Arminians’.
This was not so much because the English ‘Arminians’ were interested in the issues about salvation and predestination that preoccupied the Dutch. Although many of them did reject what they saw as a Calvinist doctrine of predestination, and there were friendly contacts between Dutch and English Arminians, the English party was more concerned with promoting ceremonial worship, use of the sacraments and a high view of clerical vocation (shockingly to many they had no hesitation in regularly calling ministers of the Church of England ‘priests’). None of these issues had any parallel in Dutch Arminianism. The Arminian label was a symbol that England was witnessing a rebellion against European Reformed orthodoxy as serious as that of the Arminians in the Netherlands. Some also nicknamed the developing Arminian clique ‘the Durham House set’, since from 1617 Richard Neile was Bishop of Durham, and Durham House, his London residence in the Strand, became an unofficial headquarters for the group, the shadow equivalent of Archbishop Abbot’s Lambeth Palace across the river Thames."
Thanks, this clarifies a lot. Glad to hear it's a sensible interpretation. Basically my whole understanding of American history is underpinned by Albion's Seed. If this was an outright mistake like I had assumed, I would have had to go through all my assumptions with a very fine toothed comb.
"The Vicar of Bray" is a satirical song from the 18th century and very good at showing the twists and turns in the Church of England from the early 17th to the early 18th century (reigns of Charles I 1625-1649 through to George I 1714-1727):
The Vicar of Bray - Lyrics
1. In good King Charles' golden time, when loyalty no harm meant,
A zealous high churchman was I, and so I gained preferment.
To teach my flock, I never missed: Kings are by God appointed
And damned are those who dare resist or touch the Lord's annointed.
(Chorus):
And this be law, that I'll maintain until my dying day, sir
That whatsoever king may reign, Still I'll be the Vicar of Bray, sir.
2. When royal James possessed the crown, and popery came in fashion,
The penal laws I hooted down, and read the Declaration.
The Church of Rome, I found, did fit full well my constitution
And I had been a Jesuit, but for the Revolution.
(Chorus)
3. When William was our King declared, to ease the nation's grievance,
With this new wind about I steered, and swore to him allegiance.
Old principles I did revoke; Set conscience at a distance,
Passive obedience was a joke, a jest was non-resistance.
(Chorus)
4. When Royal Anne became our queen, the Church of England's glory,
Another face of things was seen, and I became a Tory.
Occasional conformists base; I blamed their moderation;
And thought the Church in danger was from such prevarication.
(Chorus)
5. When George in pudding time came o'er, and moderate men looked big, sir
My principles I changed once more, and I became a Whig, sir.
And thus preferment I procured From our new Faith's Defender,
And almost every day abjured the Pope and the Pretender.
(Chorus)
6. The illustrious house of Hanover and Protestant succession
To these I do allegiance swear -- while they can hold possession.
For in my faith and loyalty I never more will falter,
And George my lawful king shall be -- until the times do alter.
(Chorus)
So you can see that Puritan, Calvinist, etc. in the Church during the immediate aftermath of Henry's Reformation, through the reigns of Edward, Mary, Elizabeth and then James on, would have been very fluid categories with some, but not total, resemblance to the Continental models 😀 (and with clergy of various sympathies in official positions as bishops, so you had 'Puritan' bishop and 'Arminian' bishop in the Church of England).
I'm an outsider here, so I do see Arminianism as a variant of Calvinism, where it grew up within Calvinism and along Calvinist principles but the major disagreement is on predestination. I think Arminians would have identified as belonging to the wider Reformed, even if the Calvinists had rejected them.
Under the rule of the Parliamentarians, there wasn't the same force holding things together, so a lot of little groupings were free to do things their own way:
"From 1649 to 1660, Puritans in the Commonwealth of England were allied to the state power held by the military regime, headed by Lord Protector Oliver Cromwell until he died in 1658. They broke into numerous sects, of which the Presbyterian group comprised most of the clergy, but was deficient in political power since Cromwell's sympathies were with the Independents.
With the abolition of the Act of Uniformity, even the pretense of religious uniformity broke down. Thus, while the Presbyterians were dominant (at least theoretically) within the established church, those who opposed Presbyterianism were in fact free to start conducting themselves in the way they wanted. Separatists, who had previously organized themselves underground, were able to worship openly. For example, as early as 1616, the first English Baptists had organized themselves in secret, under the leadership of Henry Jacob, John Lothropp, and Henry Jessey. Now, however, they were less secretive. Other ministers – who favored the congregationalist New England Way – also began setting up their own congregations outside of the established church.
Many sects were also organized during this time. It is not clear that they should be called "Puritan" sects since they placed less emphasis on the Bible than is characteristic of Puritans, instead insisting on the role of direct contact with the Holy Spirit. These groups included the Ranters, the Fifth Monarchists, the Seekers, the Muggletonians, and – most prominently and most lastingly – the Quakers.
The Puritan movement split over issues of ecclesiology in the course of the Westminster Assembly. In the course of the 1650s, the movement became further split in the course of a number of controversies. With no means to enforce uniformity in the church and with freedom of the press, these disputes were largely played out in pamphlet warfare throughout the decade."
Asking for a friend: Is there a point to fixing small mistakes in review contest submissions that are finalists, or will they be posted as originally submitted?
Is there now a culture war over AIs? I admit that I feel somewhat irrationally angry at Doomers. Why do I?
Is it because I am tribally aligned with Tyler Cowen? I don't know why that would be the case, but it's a possibility.
Is it because the AIs killing humans just seems like The Latest Hysteria -- along similar lines as Overpopulation, Climate Change, Trump, Covid and Satanism? Yes, Doomers do strike me as the drama queens of The Latest Hysteria, and that may well be why I hate them.
Is it because there is no escape from Culture War, and the more you try to avoid one the more likely you end up in another?
I don't know the answer. I don't really hate Doomers. At least not most of the time. They may even be right. They just annoy me. I need to get over it, I guess.
What is your working definition of culture war? I'm skeptical of just dropping the google result in but I don't have a working one myself, so just as a starting point - "a conflict between groups, especially liberal and conservative groups, that have different cultural ideals, beliefs, or philosophies."
Personally I find this definition unsatisfactory because it kind of renders *any* policy debate a culture war. Find me an ongoing debate, from BLM to where to put a local road, where there is disagreement, and people *aren't* dividing into groups to support or oppose particular options based on their different "beliefs."
Personally, I think there has to be some kind of threshold of scale for something to become a culture war, and I don't think AI is there yet. It's generating news coverage, and it's a surprising and new divisive issue within the kinds of niche "smart on the internet man" communities like this one that (a) were already interested in the issue, and (b) are usually aligned on many subjects such that its startling to see the break, but it feels like the average person on the street is going to have more and stronger opinions on trans rights and race issues, which is why those things seem to more to fit the "culture war" moniker than AI does (at least to me, at least for now).
[Edit] Also, who do you consider a "Doomer?" If one person says "AI development should be banned" and another person says, "Regulators should require AI products be subject to testing standards before deployment - and periodically thereafter - to confirm they work as intended, and should require developers to ensure that AI-generated outcomes are explicable" are they both "Doomers?" Just the ban proponent? Neither (since neither of them are stating a belief that AI is going to kill us all)?
There is absolutely a culture war, and doomers have a classic Eschatology. While they probably have a more rational view than Christians thinking the Rapture is soon, for instance, I think that the doomer movement needs more scrutiny.
When a group has all the hallmark traits of a doomsday cult, it makes sense to be more skeptical of their conclusions.
I think the comparison to the Rapture is quite apt, but that saying one is more rational isn't. Both stem from strong priors and not a lot of empirical facts. If we are valuing empirical facts (as implied by saying one is more rational than the other), then I feel like we need actual empirical facts, not strong priors. Like, an AI system recursively improving itself, or trying to kill someone. You know, actual things beyond a theory.
I relate to this, and have determined there’s a big fear component for me. On some level, validating the debate opens me up to considering the end of the world.
Additionally, adherents this particular hysteria feel more out of touch — partially because many of them I know are genuinely out of touch people overall. At least with overpopulation, Covid, climate change, etc, everyone can sort of see things slowly going wrong and picture what things would look like if they do go very wrong. This one, however, has a sort of nerdy ivory-towerness that feels completely removed from the lives of everyday people, so despite being a nerdy ivory-towerer myself, I feel reflexively off-put.
Lastly, the huge gap between the threats that currently exist from AI and the end of the world are so massive and poorly explained that even I as an ML engineer have trouble picturing exactly through what means AI will cause my demise. Obstacles that seem very critical to me, like the requirement for massive scaling and improvements of difficult robotics are brushed off in nearly every explanation I see. This plays into my aversion to putting resources towards it too — even if the expansion of alternative energy isn’t necessary or doesn’t save the world, at least the air smells a bit better, geopolitical dependence on oil is reduced, and I save money with my electric car. However, if AI turns out to never be an issue, what was the use of all that money, brainpower, and worry? It just feels like a waste already (especially the worry part).
I’m still totally torn on this issue, but think you really described a cultural/emotional aspect I hadn’t been able to put my finger on.
Well it seems to me that what's happening to you is *not* that you're getting sucked into a culture war and hating on the outgroup, and also is not that you've mentally joined a cult. Sounds like good old-fashioned fear to me, and I don't think it's silly to be very confused and uneasy. I am. It's just really figure how plausible the pessimistic view is.
The doomers, yes. Anyone who thinks p(doom|unaligned AI)=1.0, or anything even close to that, is too enamored of their own smartness. Thinking that this doom is imminent because their buddies made ChatGPT, and then that they can avert doom with their Extra Smart AI Alignment Research, yeah, massive overconfidence.
But, I'm just as skeptical of "highly valuable new technology". Particularly for any AI that is going to come out of anything we are talking about now, I'm not expecting it to be any more valuable than e.g. social media. Some legit use cases, but mostly noise and social strife and generally making life a little bit worse for everyone.
I'm pretty sure AI is going to get us cheap, mass-produced mediocrity *long* before it gets us Canned Super Genius, and cheap mediocrity isn't going to bootstrap itself. It's just going to generate noise, and be used as a weapon by people playing status games against one another (and against me, whether I want to play or not). If and when we do get a Canned Super Genius, I think it's very likely to be persistently misaligned, and if I think that's more likely to mean that it keeps going catatonic or self-destructing than that it tries to paperclip us all, that still doesn't add up to "highly valuable".
Highly Valuable, is not impossible but requires a lot of things to line up just right. Doom, is not impossible but requires a lot of things to line up just wrong. Mass-produced mediocrity, that looks to be surprisingly easy.
I think you have what I think of as a hate structure going on regarding people who think there'a a high probability of doom. You imagine they're all mini-Yuds, and that their sense of doom is just the delicious opposite side of the coin that, when heads, shows them being smart as fuck. I doubt that's true. Some I've seen posting on Twitter seem to be just very scared people who expect catastrophes of many kinds -- we get crispie crittered by climate change, covid wrecks our bodies and soon half the world will be bed-bound, nukes, pesticides, etc etc. Lots of very scared people around these days. As for the tech bro's who think they and the other smart tech bros will make AI safe by clamping the alignment attachment to their dicks and aligning that bitch -- eh, that's just one school of optimism. Actually, most of the people I know have heard that some are worried about AI, but believe that the scientists will figure it out like always and things will proceed as usual.
BUT I agree that mass-produced mediocrity looks pretty easy to achieve. Sort of like building a mall.
Lately I've been watching old Johnny Carson Tonight Shows from the early 70s and am fascinated by how foreign the culture seems. The past is a foreign country, they say. A striking difference is how sexual so many of the jokes are. Many of the things Johnny or the other guests say to the female guests would be considered misogynistic today, because they visibly react and comment upon the attractiveness of a women in comedic, salacious ways. It's hard to tell whether the female guests would prefer if the males behaved differently.
For instance, Johnny will often ask the guest, male or female, about their dating life. It seems to be considered natural that these famous people might want to talk about their private dating lives. Johnny might also talk and joke with male guests about good lines for picking up women. References to getting drunk and having one night stands is normal.
In one episode Johnny makes a joke about having had VD during the war, and then it hit me. I had assumed I was seeing a cultural product from the peak of the Baby Boomer era, and that explained why the show was so loose in discussing sexual matters. But then I realized Carson himself had been in WW2.
It got me thinking that maybe the WW2 vets changed the culture more than even the Boomers. After all, nobody was more worldly than the vets. Consider all the cultural leaders circa 1970 who were WW2 vets: Johnny Carson (with the biggest TV audience ever in history), Hugh Hefner, Kurt Vonnegut, Norman Mailer, J.D. Salinger, Joseph Heller -- all who pushed a countercultural take on America to various degrees. Carson himself isn't normally considered part of the "counterculture", but I dare you to watch those 1970s Tonight Shows and report back that they don't seem countercultural compared to your image of the early 1960s or the sexually conservative world that is 2023.
Maybe the Baby Boomer were overrated when it came to cultural change in the 60s and 70s. Maybe it was more about their veteran daddies.
It’s interesting to think about the Silent Generation contribution to the Sexual Revolution, but I think the GI/ prostitution angle is a red herring here. Casual attitudes to prostitution can easily coexist with quite strict sexual ethics around family formation and the behavior of “nice” women. Victorian London, for example, had huge numbers of prostitutes of all ages and sexes, with relatively destigmatized access for men who could afford it, while famously enforcing a tight code of sexual purity and chaste speech around its marriageable middle-class girls—and importantly, that ratcheting-up of domestic purity norms happened *with* the return of world-weary veterans from the Napoleonic Wars.
To take another example, Shakespeare’s plays joke openly about fun with whores in a way that would shock right across the political spectrum if written today, but there are also multiple plots that turn on the importance of female chastity in the aristocratic marriage market (note, again, as imagined by a middle-class person).
It doesn’t feel revolutionary for Carson to joke about catching the clap, but it does feel new for that to be perceived as a great topic for humor on a mainstream TV show aimed at a mixed audience. If the middle-class moms in the audience laughed excitedly, rather than finding it disgusting and offensive, then presumably it was something the moms wanted to hear and something the producers wanted to have said to them. I suspect it’s probably about legitimating a fantasy of libertine consumerist pleasure as a high-status coastal norm, to a willing audience of Middle American rubes. (Look at these Hollywood people, so rich and well-dressed, so morally relaxed and carefree! *They* understand that it’s OK to have a bit of fun now and then.) Feels very much of a piece with the mainstreaming of _Playboy_ as an openly accepted part of the suburban dad library, and that’s definitely also about class and consumption (smoking jackets, satin sheets, marble pillars...).
Didn't Scott have a piece on this at some point? At the very least, I seem to remember an extended conversation about how the cultural figures associated with a given generation are often actually members of an earlier generation (most of the famous rock stars and artists of the 60s were members of the Silent Generation or--as you pointed out--members of the Greatest Generation who actively participated in the Second World War. The Sex Pistols and the Ramones were Boomers. The bubble-gum pop stars and spiky-haired rap-rockers of my Millennial youth were mostly Gen-Xers). Which makes sense. The members of a generation generally aren't going to be old enough to be artistically influential to their own peers. Teenagers might be the prototypical consumers of youth/mass/counter culture, but they by and large aren't big producers of it. Even teen pop idols and child stars don't really count, because their artistic output was mostly designed by the adult producers and songwriters around them (although I'm curious how this is shifting with the current generation of teenagers. Current social media stars/influencers seem to be much closer in age to their audiences than the rock stars of yore, with much more control over their own messages than earlier teen stars)
I wonder about those young influencers and the material they release. It's clearly less polished, more repetative, and often drops the ball at key moments. It's the opposite of highly choreographed shows like you see on TV, especially as related to reality TV, which is heavily edited to take real (or "real") events and hype the excitement.
My kids started watching Mr. Beast recently and I've seen a few episodes. Climactic moments often fall apart. Filming seems random and out of order. Props fail in highly predictable ways. This is bad TV, bad storytelling, bad writing. It reminds me of hanging out with my friends as teenagers. I'm thinking that my generation liked the polish because we could see the improvement over the haphazard stuff we could do on our own, but the upcoming generations get too much polish and appreciate the authenticity of young influencers who don't have adults running the show and making everything perfect. It may be related to the fact that kids are spending a lot less time hanging out with their friends in random unstructured ways. That's an itch not being scratched, so they get it from digital sources instead of physical.
I was thinking specifically about the effects of the war on the psychology of these people. For instance, Carson's attitude toward sex was probably formed in those years. Casually referencing sex with a hooker was no big deal to him or much of his generation. Prudery likely doesn't survive major wars.
But taking that and adding it with your point makes me wonder if the so-called Sexual Revolution was merely a transfer of values from ex-GIs to the younger generation.
New article on my Substack - it's a review of "Innovation Breakdown: "How the FDA and Wall Street Cripple Medical Advances". This is a rather obscure book which tells the story of a skin cancer detection startup that got screwed over by the FDA, delaying the entrance of their product to market by several years and severely damaging the reputation of company in the eyes of dermatologists.
Maybe I've missed this discussion in previous threads, but, um, why Threads? To me, it seems that Twitter is a) money-losing enterprise, and b) huge magnet for unwanted regulatory attention. Why is Meta trying to replicate this seeming trainwreck?
One needing another Meta account (Instagram) to use another Meta product (Threads) is a red flag.
Another is the requirement to delete the Instagram account to delete Threads. Meta claims to be fixing this, but who knows how long that will take. Perhaps signing a suicide pact with users’ Instagram accounts contributed to the alleged 100 million registrations.
I suspect we think that with better execution we can make it genuinely positive, for the users and for the business. We have more expertise and resources in content moderation than Twitter ever did
Cambridge analytics was more about decision making around 3rd party API access than it was “content moderation” per se… but yes fake news troll farms were / are a content issue
Are there therapists who focus on helping people who get into conspiracy theories and distrust experts, which leads them to avoid doctors and medication, which worsens their condition?
In my very limited experience, the big issue isn't finding an appropriate psychologist, it's finding an appropriate caretaker. People with serious issues like this don't need a lot of complex therapy, they need to take incredibly powerful pills every day for the rest of their life that make them miserable and they hate, otherwise they will quickly become incredibly paranoid, like won't stand within 20 feet of phone lines, and in general just degenerate incredibly rapidly. The issue isn't, ya know, finding a solution, the problem is generally making sure that the person is consistently taking these pills, that again they absolutely hate, on the worst day of the year. People who are, underneath it, just people with all our failings. From the outside, their internal experience of skipping a pill seems like no more of a moral failing than buying a pizza and candy when you're depressed, just with horrifically worse consequences. This generally means there needs to be a caretaker person, who needs to ensure this person is on their meds, every day, for like 10 years, which is inevitably going to lead to a lot of screaming fighting matches between the caretaker and the paranoid. It's a brutal, thankless task that consumes decades of a persons life. Finding someone who will do that is difficult, finding someone who will diagnose the paranoid schizophrenic as a paranoid schizophrenic and get him pills and weekly meetings is easy.
I don't think that's a specialty. But most competent therapists, talking with someone like that, would make a special effort to stay out of the role of expert, to not give any advice,to try to be particularly transparent and honest, and to have a stance of "how can I be helpful?" I have sometimes been able to develop a working relationship with someone who had that stance. However, if the person is really deeply furious and mistrustful that is not going to work either. The last person of that kind that I saw said that being a psychotherapist was the ideal profession for imposters with no degree, because everything we say and do is bullshit anyhow. Expressed no interest in seeing proof of my degrees, because said it was still possible that I had stolen the name and the identify of the person who had actually gotten the degrees. I ended up suggesting he consider co-counselling, which I think is actually sort of high-risk, because this man was so utterly incapable of trusting authorities and yet was greatly in need of help.
Does anyone have any clear examples of someone skillfully avoiding Gell-Mann Amnesia? I personally just avoid news in general, but I wonder if there are other options. I will preemptively say that my news abstinence is trivially compatible with all strategies of the form "don't trust a news item unless you [take some extra, effortful step]"
Some stories are covered by sober minded experts and others are covered by complete idiots.
If the NYT has a piece about portage canoe trips that I know for a fact is not fully baked I don’t thInk I should regard every story on the site with suspicion.
Michael Chrichton came up with the idea and named it after the guy who discovered the quark. The physicist was brilliant and Crichton is a so-so sci fi writer. Now that naming is a bit of deception in itself. How much traction would the Michael Chrichton Amnesia meme have gained?
The point that GMA is making is not "all news articles are bullshit". It's "some news articles are bullshit. Remember this and be appropriately skeptical, especially when you are not an expert in the topic".
We know that _some_ articles even in well respected outlets are garbage. So if you are not capable of telling (which most people are not on most topics), then the natural response is to be wary of articles.
I agree we should remain skeptical and pay special attention to the source. Do your own homework. My quarrel is with the particular name, GMA. Crichton wasn't really coming up with anything new and naming it Gell-Mann was a bit of cheesy sleight of hand to make it seem like it was something cool and new.
It is especially cheesy because Gell-Mann was famously skeptical and very conscientious about getting his facts right. And he knew a *lot* of facts. I remember a Seth Lloyd book in which Seth said that meeting GM for the first time is intimidating because GM knows about 10 times as much as a typical person on any given topic. Another eminent physicist (it may have been Stuart Kauffman) opined that Gell-Mann knows more facts than any other human. It's impossible to adjudicate such a thing, but FWIW I can't think of anyone I'd put ahead of Gell-Mann at his peak in that regard.
Zvi recently linked to a tweet from lawyerBrian Manookian describing how he got off the hook for his student loans: https://twitter.com/BrianManookian/status/1674963884703088642, and saying anyone whose student loans have gotten passed from lender to lender can do what he did.
The approach is to demand proof of chain of title, i.e. proof that the loan company now servicing the loan now owns the loan, which I believe they prove by showing a bill of sale. He says that these companies that buy up big batches of student loans are sloppy, and do not do the proper paperwork, and probably will not be able to supply proof that they own a loan. The person who got the loan then petitions the court to declare that the person does not owe the loan company any money. Court does so, person sends judgment to loan company and they cannot pursue collection any more. (Seems to me that nobody else can either -- because the previous owners of the loan have sold it to a collection company, and really don't own the loan any more.)
This sounds plausible to me, and my daughter and her boyfriend both have substantial student debt and are interested in trying it. However, we would like to ask a lawyer for advice before trying it. Four lawyers have now turned me down when I have asked them to first advise us, then carry out the process if we are not confident we can do it correctly. We would be paying them for their time, of course. Lawyers have said it's outside their realm of competence. The steps Manookian describes sound to me like extremely simple basic stuff, and in fact Mazookian has supplied copies of the letters he used as models, so it's hard to understand why lawyers keep turning me down.
Anyhow, can anyone help with information? I am open to any possibility. You can give me general info then fade away, or you can actually do the filing required. You can give info in a post here, or you could send me an email or we could talk online. We can do it anonymously or using real names. I will happily pay your hourly fee, though I think I can ask all my questions in about half an hour. If you are law student or legal aide and understand the process Manookian is describing I am happy to work with you also.
And, by the way, I am sure there are many people on here dealing with school loans. You could help a lot of people at once. Or you could do them one-by-one and have a lot of clients you can process fast and easily. Maybe you can bring down one of the loan shell companies.
This is fascinating. I don't have student loans, but I do have a 98k mortgage that has been passed around three times, probably under similar circumstances as student loans (ie, as part of a big sloppy batch of loans).
On the one hand, you would think that even crappy mortgage companies would be more careful about bills of sale given the large amounts of money involved.
But on the other hand, mortgages come with a lot more paperwork than student loans do, and if they're buying mortgages in huge batches, maybe they actually might be sloppy enough to lose critical ownership paperwork.
During the mortgage collapse of 2008 there were a couple news stories related to this I think -- some seemed to be about the inability to address the crisis because it couldn’t be determined who owned the loans, and some about a lender/loan servicer that didn’t have its paperwork in order and so couldn’t evict people who didn’t pay.
I've been doing some reading since originally posting my comment - the TLDR is that mortgages seem to take documentation much more seriously than student loans and there are lots of additional protections for mortgage lenders, at least in Washington State (where my condo is). From what I can tell, even if they can't put their hands on the original paperwork, they can just do a affidavit swearing that they lost the paperwork, but, like, they totally do own the loan. But on the other hand, it looks like it's occasionally worked for people in foreclosure?
Either way, this seems like having such a low chance of working that I'm not going to seriously pursue the idea. From what I can see, "produce the note" is much easier to do with student loans because, unlike real estate, there is no underlying asset or public records about it.
I can tell you that these "Chain of Title" arguments are very important in patent litigation. In some jurisdictions "incorrect recording of inventorship or ownership is, in itself, fatal to the validity or enforceablity of a patent." (quote from here: https://blog.patentology.com.au/2014/02/inventorship-and-ownership-importance.html)
I don't know if the law is as clear when it comes to chain of title for loans, but maybe a lawyer with experience as a patent attorney would be more comfortable making these arguments?
I’m not sure about student loans, but I did use a similar technique to get rid of credit card debt a long time ago. The collection agency was not able to provide any proof of the original debt so they had to forgive it.
I am not a lawyer, and cannot offer legal advice. This seems like it's not going to work effectively; if the shell company doesn't have record of the paperwork, they can get it from the people they bought it from. If they don't remember that, once you take them to court they can demand you tell them who that was as discovery, and then get their paperwork. They can probably make you testify under oath that you don't believe you owe the debt, with whatever perjury penalties that incurs. It really seems like you're just banking on them not showing up and losing by default, which a lawyer isn't going to want to do.
Tennessee seems to have a $25 paywall on their court cases. Might be worth your money to look up exactly what happened with that linked case, or find others like it.
Thanks, I will look it up. Guy who gave the advice said that it's not cost-effective to these companies to defend themselves so long as claims like this are rare. They have a huge number of loans, and know they will lose many to defaults. Losing to me is just one more. Calling me into court and having their lawyer grill me sounds like it would cost them a hell of a lot of money. As for getting the paperwork from the previous company -- I think a lot of them just collapse and go out of existence, then sell the loans they are holding cheap. Also, I would not be claiming that I (actually my daughter, but let's say it's me) did not know who had made the original loan and do not believe I borrowed money. I would just be saying yeah, I borrowed it and will pay it back to the organization that can prove to me it is now the owner of the loan. But I'm not sending $30,000 or so to a company that cannot document that I owe them money. No perjury involved.
Recently convinced a relative to hire a lawyer for an eviction proceeding; it cost them $4000 in total. Not chump change, but much less than the property value. The guy's arguing it's not worth the collector's hassle, which is banking on them not showing up.
Was going to say my loan servicer has changed at least twice since they were paused for COVID (and possibly more than once before that) and I could maybe honestly say I don't know who the previous ones were since I've kinda not been paying attention until they're a more pressing problem and the repayments are starting up again.
No you would not need to lie in any way. The way this works is not that you pretend you don't know who the loan company is, or who had it before, or that you owe a loan. All you are doing is saying that yes you took out a loan, and you will repay the loan to a company that can prove via paperwork that it bought the loan and is now the owner of it -- BUT that you are not sending money to a place that can't even prove it owns your loan. It is reasonable to demand that a place trying to collect a large sum from you can document that its claim is legitimate. If it can't document it, it is reasonable not to pay it, EVEN IF you think it is extremely likely that they did buy the loan, just didn't do the paperwork. I mean, it's like if you bought a rug for $5000 online, and start getting emails from the place saying send us the money -- but they can't tell you the date you bought it, the item number, and they can't prove that they are the online rug store. It's just a guy writing from his private email saying "you bought the rug from my store a while ago, and we shipped it. No I have no records of the date you bought it and no proof that we shipped it, but I'm telling you we did."
Right, I was just saying if it was a question of being asked to provide the paper trail on the loan servicer's behalf I would be able to give an honest "I dunno".
Anywho I'm sort of concerned that Brian Manookian might be a bit of an internet clout chaser, looking at the rest of his Twitter feed. Seems worth noting that he's qrt bumping a tweet from August of last year and keeps saying he's going to set up a website detailing his process and has yet to do so.
Like I don't think he's lying or that he didn't do what he's saying he did per se, just that I sort of suspect there's probably details and circumstances that are being obfuscated in a way that prevent this from being something scalable at all.
Yeah, he's definitely got a big ego and no heart of gold, and I noticed too that he hasn't set up a website, so bad at follow-through. However, one of his Twitter threads contains what's probably the crucial document, which is the one you send first. I think there's a fairly good chance that this would not work, but it looks promising enough to me to put some effort into figuring out for sure. A couple of the lawyers, before they declined to be hired, have said things in passing about why the think it would not work:
-This is not a way of getting your loans discharged. (But I already knew that. They will still be on record in various places. On the other hand, there will be no record of ongoing failed attempts to collect. The loans will exist forever in a sort of limbo: You borrowed the money, but the original lender sold the loan, and the buyer resold it, and the last organization that bought it can never collect it because the court has declared their claim permanently invalid. And none of the earlier loan-owners can collect because they sold the loan, and really do not own it any more.)
-Manookian's posting about something he did in 2016. Loan collection companies are different and better organized now. (Have no idea whether that's true, and it's not clear that the lawyer who told me that has every tried this maneuver with somebody's student loans. How would he know that the loa collection companies are better organized now, given that he said does not handle student loan debt in this practice. His field is "debt defense," but he seems to mostly help people who have very large debts and are in big trouble.
-Most student loans are federal loans. The loan companies that take them over don't really own the loans, they are being hired by the US government to do the collection work. (He would not tell me how he knew that or how to find out more about it.)
The thing is, I have had a lot of middle-aged white men (please forgive me if you are one) in authority tell me something is wrong or impossible, and my observation is that if I argue, politely, or ask probing questions a lot of them double down and start throwing out reasons why they're *positive* their view is right, and in fact these back-up arguments turn out not to be based on solid info when I look them up on google scholar or wherever. They seem to be incapable of just saying, "based on my experience and intuition that's wrong, but I can't really back that judgment up and it's possible I'm wrong." And about several important matters it has turned out that I was in fact right.
If you are interested enough in this possibility to look into it on your own, let me know.
I have little interest in the outcome of this, but I'm curious why four lawyers have turned you down.
My first two guesses would be 1) it's too novel ("outside my realm of competence" sounds like the professional phrasing of that) and 2) it violates some professional norm against lawyers being highly lawyerly. More interesting to me is whether any of them relied on consequentialist reasoning, like this attempt to apply Econ 101 concepts:
'If this becomes a trend, then lenders will probably take notice and then take steps to ensure that each buyer of each loan can demonstrate proof of chain of title with ease. Won't these additional steps taken by lenders raise administrative costs at all lenders? Won't that increase the interest rates that these companies charge future borrowers?' (Similar to the model in which shoplifting by some increases the average prices that consumers pay.)
Are plaintiffs' attorneys known to make decisions on that sort of basis? It doesn't sound like the sort of thing that they do, but then I have never asked.
Where is the AstralCodexTen for people interested in law, anyways?
Part of it may be that lawyers have an enforceable profession duty of candor to the court. They're not allowed knowingly to put perjured testimony on the stand, present forged documents, present spurious legal arguments in briefs, etc. You don't have to volunteer evidence or arguments against your client's interests, and you can definitely present things in the most favorable light to your client's claims, but you aren't allowed to outright Saul Goodman it up.
A lawyer could probably defend an action brought by a loan servicer by demanding proof the servicer owns the loan and nit-picking every shortcoming in the documentation the servicer provides, sure, but bringing an action on the client's behalf seems like it might require more of a prima facie case. If getting such a case heard would require the lawyer to file an initial brief saying "my client doesn't think he owes this money to the respondent", then an ethical lawyer (or at least a lawyer who doesn't want to risk being sanctioned by the court) needs to reasonably believe the claims he puts in his brief.
Wait, you are misunderstanding to plan I have learned about. There is no perjury involved. There is nothing that is even sort of vaguely *like* perjury. The person with the loan does not say, "what loan? I never took out a loan." They say, "yes I took out a student loan and I will repay the loan to the original lender, or to any other organization that has bought the loan. However, the organization that asks for me to make loan repayment has to be able to prove they are actually the owners of the loan. They should have a bill of sale or similar for my particular loan." If you win, it is because the loan organization bought a bunch of loans in a big batch and did not do the required paperwork. There is actually a law about this sort of thing to protect consumers -- something along the lines of, you can't dun somebody unless you have documentation of the proper kind showing they owe you money.
Anyhow, some may think this approach is kind of sleazy, because it's getting out of a loan one really does owe on a sort of technicality. But my daughter and her boyfriend have oppressively large student loans and want to do it, and I am not about to be judgmental about that.
I know one lawyer pretty well, and he's actually a highly ethical person who worries about things like that in real life. But his view is that the job of a lawyer is to make the best case he can for his client, and that considerations like those you mention are irrelevant. It's OK to take the OJ Simpsons of this world as clients and do your best to get them off the hook -- not because they deserve to go free, but because a system where everybody gets a good defense protects the justice way better than letting each lawyer's squeamish scruples rule. I agree with that point of view. And I can tell you that none of the 4 that turned me down sounded like they gave a shit about that kind of thing. I think the reason might be that as lawyer jobs go mine is really puny. Manookian estimated that a lawyer charging typical hourly fees would make maybe $1000 learning about situation from client and then doing all the paperwork involved in this project.
If lawyers were in the habit of making decision based on the type of logic you describe, we would have dramatically fewer lawyers. The majority of the lawyer profession exists to take advantage of exactly these kinds of bureaucratic overheads and process inefficiencies.
For anyone who has quit their job to take a break to pursue personal projects, how was it? Do you regret it?
I am currently torn between burning desire to quit my job and pursue a startup idea, and ruthless pragmatism that just considers the opportunity cost of lost paychecks.
I have no kids, no rent/mortgage, and many years of living expenses in the bank, but somehow can’t rationalize to myself actually taking a risk. If you ever did quit to pursue a personal interest or purely take a break, I’d love to hear from you!
I took a three month break last year to explore career options and writing, and it was the best time of my life!
I got a job again soon after just from networking in my local area, but I do not regret it at all. I think we should normalize taking sabbaticals, it's extremely useful for mental health and breaking out of rigid mental models.
I don't think people are morally or ethically against taking sabbaticals. I think we're financially against it. I would love to if I could afford to just be jobless for three months. The privilege...
I've taken at least three year-long breaks from work, first at age 20 and most recently in 2015 when I started a profitable side business manufacturing electronic products. Now that I have children, I don't expect to have that sort of opportunity again.
I grew up in near-poverty and started work in my mid-teens. My early life left me extremely frugal and with a horror of debt, so living off savings and finding work when money ran out was a naturally self-motivated process.
I pursued hobbies, did some traveling, and had some adventures, but frankly I never accomplished as much as I expected (except in 2015). Romantic partners were generally supportive but unimpressed.
I feel that I was lucky to be stably employed during periods of economic turmoil like 9/11, 2008, and the pandemic. I would caution you that various economic indicators don't look so hot, and seem likely to decline.
Probably there were opportunity costs, but certainly opportunity gains. No apparent career impact, and no regrets. All of the trite homilies about non-fungibility of time and the advantages of youth are true. Have a plan, keep to a more-or-less heathy sleep schedule, set expectations low and curiosity high. Do it!
How hard will it be to get another job if you need one?
I've never quit to pursue a passion, but my last batch of unemployment lasted six months, during which I made no progress toward any goals. So, don't assume the time will make things happen.
I have a relative who started their own company, successfully, and they mentioned mostly working 20-hour days seven days a week for about two years. Expect pain.
On the pro side, I think of Shamus Young's autobiography; https://www.shamusyoung.com/twentysidedtale/?p=13953. "It’s been easy over the last year to just keep coding at home and punching the clock at Taco Bell. This isn’t my dream, but it’s a tolerable substitute."
I'd say, go in with the expectation that every dollar you put into the thing is completely lost.
Thanks, I’ll take a look at that. I’m an ML engineer, so I don’t expect difficulty finding a job again, though I have a pretty chill and pleasant remote gig now that I don’t want to throw away. I am definitely in the (very) tolerable substitute territory, which is part of what makes it hard to walk away.
I certainly expect any startup process to be grueling, but feel this might be my best time to give it a go. I’m in my mid/late 20s so if I’ll ever have the energy for that sort of pace, it’s now.
I’ve done it multiple times, though not always to pursue personal projects. Sometimes just to live life. It’s paid off every time with the possible exception of quitting in January 2020. I wrote more about job-quitting in my recent Substack post titled “the fifth bomb”
You might also find the recent book “die with zero” an interesting perspective on the time vs money problem. The author frames it more as an energy problem, with your time/energy being more valuable when you’re young and losing value over time. So selling your time/energy for dollars when you’re young is a bad exchange rate (from that perspective) which supports quitting jobs often.
Also recommend the YouTube short film Slo Mo (I think it’s called) about a surgeon that loses his vision and quits working.
Interesting, thanks for sharing! I’d love to take a look at your blog. I’m in my mid 20s now, so I’m certainly curious in the long-term opportunity cost equation for risky endeavors.
I do suspect growing up in a working class family one generation removed from extreme poverty has made the opportunity cost numbers seem gargantuan, though in reality, a six month attempt that returns 0 would really just mean working an extra ~3 months at the end of my career, which sounds trivial.
It’s tough stuff to grapple with so I appreciate the alternate perspectives!
Yeah I also grew up in a working class family which definitely left a chip on my shoulder too. Definitely skews perspective a bit and great that you’re aware of it
I’m late thirties and hesitate to give genetic advice, but generally think “do what you want to” seems to play out better for high agency people vs “suffer and grind”
I try to do a 24 hour fast twice a week, sometimes three, if I'm feeling ambitious. I go noon one day to noon the next, so I'm asleep for 8 hours in the middle of it. That seems to make it much more tolerable, in my experience.
Handicapping a potential Musk-Zuckerberg MMA fight. For some reason I had the impression that Zuck has been training for years, but I recently read that he only started BJJ 2 years ago. 2 years is.... not a very long time to be training, at all. At this point he probably doesn't know what he doesn't know. Reportedly Musk weighs at least 70 pounds more. I would personally not recommend that a white belt with 2 years of training and 1 competition take an extremely high-profile fight with someone 70 pounds heavier than them. For one thing, MMA fights start standing, and takedowns in no-gi take a lot of energy and skill unless you grew up wrestling. No-gi takedowns against someone much much larger is going to be very, very tough, and very, very exhausting.
Like, I don't any think of the say 170 lbs. high school or college wrestlers in my old gym could just easily take down a 240 lbs. white belt. In the gi maybe for a judoka, but not in no-gi.
Anyways, there's not a ton of fields where I know substantially more than Mark Zuckerberg, but this is 1 of them- I would personally not advise this fight. Contra Hollywood, size really matters in fights
70 pounds make a difference talking about actual athletes, where you can assume most of it is muscle. Elon looks chubby and soft. His cardio is probably terrible too.
Has Elon trained in martial arts at all? Because 2 years is a hell of a lot more than 0. I'm bigger than Elon but I've never trained in any sort of fighting so I'd hesitate to go up against the Zucc.
Respectfully disagree. You seem to argue that Zuck can’t win but that isn’t an argument for why Elon can. I can’t see him landing a KO punch so not sure his path to victory. Plus, Zuck is 13 years younger if it goes the distance.
The more interesting question to me is who will be on the undercard!
Update for the zero people who care: the combo diet of keto/intermittent fasting continues to work well. I'm down about 20 pounds in about two months. It's been remarkably easy, too. The keys to success are a) try to eat lean meats but don't be afraid to splurge now and then on some bacon or whatnot b) get some exercise but don't overdo it, cause that's when falling off the wagon is most likely, and c) don't be afraid of a little or even a lot of caffeine.
Perhaps, but I found eating lean meats lead to much better results. For Memorial Day, we had a cookout, and an entire package of ground beef patties was left untouched, so I spent the following week eating basically nothing but hamburger patties, and the weight loss slowed or stopped entirely during that time, so...ya know.
I try to do a 24 hour fast twice a week, sometimes three, if I'm feeling ambitious. I go noon one day to noon the next, so I'm asleep for 8 hours in the middle of it. That seems to make it much more tolerable, in my experience.
Do dried psilocybin mushrooms loose potency over time? Yesterday I ate some that had accidentally popped up in a Tupperware container that I was using to store damp coconut coir a couple years ago and the visuals were just not there.
Yes, the psilocybin decays over time. Just to add a little more detail over other replies, the chemistry involved is an oxidation reaction: the psilocybin reacts with oxygen from the air to give non-active blue compounds. https://www.chemistryworld.com/news/mystery-of-why-magic-mushrooms-go-blue-solved/4010870.article I.e. fresh mushrooms turn blue when you bruise them because air gets in, and this represents a slight loss of potency. (only slight - the blue colour is very strongly blue and a little goes a long way)
Things other people have suggested to stop this happening sound reasonable to me. Keep in an airtight container, under vacuum or with one of those oxygen-remover sachets (not just desiccant which only removes water). Keep moisture out because being thoroughly dry can be a physical barrier to oxygen.
Chemistry happens slower in the cold so I would normally suggest that the freezer is good. I think the counter argument (that Eremolalos is thinking of) is that when you take them out of the freezer, they are so cold that water from the air condenses onto them, which makes them damp and reduces shelf-life if you're not using them straight away. Same if your container is not completely airtight in the freezer (get some good containers with a clip-on lid with a rubber seal).
If you do thaw and you're planning to refreeze, maybe thaw completely to room temp before opening, or even better, thaw the containers by putting them in an even bigger airtight container with desiccant.
Forgot to mention the other thing from that paper - part of the oxidative loss of potency is assisted by enzymes in the mushroom.
This is analogous to green tea leaves turning into black tea - the catechins/tannins in green tea turn black in an enzyme-assisted reaction with oxygen. You manufacture tea leaves by destroying the enzyme with a quick blast of heat, once the desired level of oxidation is reached.
It may be possible to do something similar with shrooms, where a quick pasteurisation-type treatment just after harvest increases shelf-life dramatically. I'm sure people have looked/are looking into it.
You're supposed to not chop them up because that exposes more surface to the air, & store them with little envelops of desiccant stuff. Last and fussiest advice I read was something about protecting them from oxygen? Or maybe air? and recommended a vacuum storage system.
Yeah they were dried and run through a coffee grinder. Each little batch stored in a small jar with desiccant packs.
I once wrote some code to calculate the decline in potency of pharmaceuticals. The big deal in that case was ambient temperature. They held up better when stored at lower temps. Not sure if this is true of mushrooms but I could store the jars in the freezer.
Or wait for some more to accidentally pop up in some damp coco coir stored in a Tupperware container as I’m sure meant by growing some more. :)
I read somewhere convincing that you should not put them in the freezer, even after they are dried. Up until then, I dried them, put in desiccant envelopes and stuck them in jars in the freezer. I absolutely cannot remember the reason why that's a bad idea but whatever it was I read convinced me. So yeah, damp coco coir
Yeah, the number will be that precise only under a similarly precise set of conditions; temperature, amount of oxygen and moisture, mixing and so on. Often though, its not just X percent of the original potency per quarter, its X percent of the remainder which is why chemists talk in half-lives (half of the remaining potency disappears in every Y-month period).
Psilocybin is not a simple example though - from the paper I linked above:
"However, no single blue compound exists in Psilocybe mushrooms. Various oxidative pathways contribute, to different extents, to the reaction. Each of them produces various chemical (including isomeric) species. [...] The substrate to oxidant ratio seems to affect the mechanism and color."
In my experience, fresh bruised mushrooms are a "pure" blue colour, with dried mushrooms being a more greenish-blue. This is consistent with slightly different compounds being formed in fresh vs dried, and these will have different rates of formation depending on conditions. Likely more difficult to model with code than Gunflint's pharmaceutical example.
Why is there so much less generative AI music than other forms of media? Is it because music that's been created to-date is all heavily copyrighted, so it's not available as training data? Generative AI obviously can create text (ChatGPT, etc.) and images (Midjourney, Dall E etc.). I've seen/heard a very small amount of AI-generated music, but infinitesimally less compared to these other media. My understanding is that existing generative AI was trained on an enormous amount of free/open-source/maybe kinda stolen media that was found online. (I mean Midjourney will randomly throw in watermarks to some images, so obviously watermarked pictures were used in the training data). Was this just not an option for music, whose copyrights are generally protected by ruthless highly litigious large companies?
I mean, intuitively to me music seems a lot simpler than imagery- if an AI can create pictures, it seems that technologically it should be able to create music too. So I'm guessing the barriers here are legal, does that sound correct?
There’s a lot of music in the public domain, but the existing recordings are copyright, and so is the published sheet music . It’s a little weird. But an AI can learn a lot of public domain music by ear. Why not? Hook up a microphone play a song. It should have perfect pitch… after that it should be able to play with what it has. But I agree with others in this thread that I doubt it’s worth it because of the nature of music and how people want to consume it, and what people usually want out of making it. Also, I don’t know how much it gets you over a modern synthesizer.
I remember a funny experience at the Lincoln library for the performing arts years ago. I went in looking for sheet music for the song “streets of Laredo”, it one in a long line of descendants of a 16th century folk song.? The rakes lament, Saint James infirmary, a young soldier, cut down in his prime etc. anyone can use that tune, and write their own lyrics to it basically.
I found the sheet music, but they would not let me Xerox it. It was perfectly appropriate for me to copy it by hand onto another sheet of notation paper though (which they provided me with) and then do what I wanted with it. Most of the western Canaan of classical music is free from any copyright. The actual recordings are a different story though; anyone can do Don Giovanni if they want to, and they don’t have to pay Mozart a nickel. So you could get an AI to do variations on a theme by Mozart, for instance and knock yourself out. You have to stay away from Marvin Gaye, though.
This is something I've long noticed. Generative music is highly feasible: spend some time listening to random OpenAI Jukebox samples https://jukebox.openai.com/ and note that this is generating raw audio from scratch, posted in *2020* using 2019 stuff, and not as scaled-up as GPT-3 was, and there's vastly more accessible audio data. You could also get fairly reasonable samples from non-raw-audio data sources with a few tricks (https://gwern.net/gpt-2-music) which sound even better if you put a tiny bit of effort into them (https://soundcloud.com/theshawwn/sets/ai-generated-videogame-music). So, where is it all?
It seems to be a combination of: (1) the copyright situation is much scarier. Music comes with *all sorts* of bizarre complex IP laws and rights that don't apply to text or images, and this is because of (and empowers) groups like the RIAA. Every commercial startup doing music-related generative work seems to keep things very quiet, as blackbox as possible, restricting to licensed datasets when possible, and definitely not releasing stuff. (2) lack of major hobbyist interest - people may like to listen to music, but a lot of it is parasocial, there's superstar winner-take-all effects, and people don't generally want to make their own music the way they do their own images & text - think about porn/hentai as a huge driver of image synthesis. There is no porn/hentai of music generative models. (3) relatively demanding compute compared to other modalities. Audio is just very, very bulky: a piece of music is made out of millions of individual datapoints (X frequencies times Y milliseconds), while you can make a recognizable image with just 64x64=4096 pixels. Jukebox wasn't as demanding as GPT-3-175b was... but it's still a lot more compute than most people want to spend. Even sampling Jukebox is pushing it: early use of the released Jukebox model deterred hobbyists when they realized it (a) didn't fit in their consumer GPUs and (b) would take like half a day to generate a full audio file. Yikes. It'd be a lot better today, of course, but is nowhere near the instant gratification of a GAN or near-instant of the OA API or heavily-optimized diffusion model.
I'd also note that images specifically has benefited from luck: you would be seeing way less generated images today, and you would be unimpressed by image progress, if Emad hadn't done what all of the giants refused to do and bankrolled a decent FLOSS image generation model. (You would be wrong to be unimpressed, because there were, and still are, many much better image generation models than the Stability ones, and trying to judge image generation SOTA by what an SD model can or cannot do is a serious mistake - but they are all locked up behind barriers so you can't use them.)
Thank you for saying this. I had the same thought, and it is one of my exhibits that I point to when I try to make the case that AI's impact is a lot more marginal relative to its perception. The nothingburger that is AI music is proof that AI is much more incremental than people expect. AI seems to frequently arrive in a context where generation is already very cheap, potentially free, or even negative. In the case of music, the negative costs are borne by all the striving artists donating in-kind time to become hits.
The music industry has already gone through multiple waves of being concerned about cheap, fake and derivative music, and sometimes those controversies ironically become foundations, such as the initial aversion to autotune until current entrenchment.
There is a recurring ebb and flow of new innovations, such as even something simple as analytics-driven producers cracking the code as to what it takes to make a hit in the 1960s, to the modern Rotten Tomatoes-like algorithmic PCA to find the right sample banks, producers, writers, and overall secret juju to produce that true derivative—but extremely catchy—hit.
Gen Music is simply not as profound or disruptive as other techs like autotune. There is still no secret to creating the perfect hit, just as there is still no secret to creating the perfect trader. (something something arms races)
And this extrapolates to other domains. I find it interesting that Gen Art came at roughly the same time we were just getting used to free high-quality images from Unsplash. At first, it was like, "Wow, Stable Diffusion solves all my clipart problems" to "Wait, these problems were already solved by Unsplash."
Or that GPT-based coding is downstream from a world where we were already googling 90% of our issues to find a good Stack Overflow link and then blindly copying-and-pasting. The existence of GitHub Copilot initially seemed like, "Wow, this is going to end all programming", and yet here I am now spending instead of 90% of my time fishing for Stack Overflow, I'm spending 80% of my time. Or, correction, I'm sending more time, because in parallel with these innovations we have the further advancement of community-heavy frameworks and packages, which seem to lend themselves to a dependence on Stack Overflow. For example, you cannot separate the rise of React from the rise in our dependence on Stack Overflow, unlike earlier eras where PHP programmers only looked up issues occasionally.
The future is boring. Object Detection's improvement is logarithmic to current AI R&D expenditure. Moore's Law is downstream from Rock's Law. Secular stagnation puts a hard cap on demand. 1960: The Year the Singularity was Canceled (by Scott Alexander), etc.
You can drastically reduce the size of generated music (and presumably the compute and data set needed) by having your AI generate sheet music instead of audio.
That just occurred to me the other day, too. And, it would improve the sound of generated music a lot. Right now it sounds like a muddy mishmash. Rather than AI having to try to discern "what a guitar sounds like" it would make more sense to have it try to discern "what notes and rhythms are enjoyable." It could generate MIDI files and/or sheet music instead of audio files.
Thanks, yes I didn't think about answers 2 or 3 at all. The universe of paying consumers who want to make their own music at home is probably pretty small. (And most of the ones who do probably want to copy existing, copyrighted artists- 'make a collaboration of Taylor Swift and Drake and Tupac' or something).
I'm skeptical the music industry will ever want to license their IP to AI either, just to ensure that they're not cannabilizing their existing artists
I remember hearing a talk about a music-generating AI something like a decade ago (possibly a TED talk?). Some musician had creative block while working on an important commission and decided to write an AI that could sample a bunch of sheet music in a given style and then compose new songs in that style, then turned it on his own past work to fulfill the commission, then gave a talk about how he did it.
I don't think copyright is an issue; most of the training data in the current popular generative AIs is already copyrighted, but the stance of the companies is that using something as training data is not a copyright infringement (and I think they are correct).
But "there isn't a lot of sheet music posted online in places where you can download it for free" might be an issue.
Hmm, but AI isn't too bad at (very) short stories. You'd think they'd be able to produce a few bars of nice music (are they? I haven't kept up with the music side of things).
I don't think that's a difference. A piece of sheet music only "changes over time" in the sense that it has a beginning, middle, and end, and text also "changes over time" in that sense.
- How does affirmative action in college admissions actually work? Are there any concrete examples of how admissions would look with and without affirmative action?
- Do we know specifically what changed in UC admissions policy after Prop 209?
- How was Prop 209 enforced? If admissions policy is totally opaque, what's stopping the university from "considering race"?
- How was the Bakke 1978 ruling against racial quotas enforced?
- How will the SFFA v Harvard ruling be enforced?
- Do people understand what affirmative action is? In this Pew survey, a plurality said affirmative action is "a good thing", but "only 17% say college admissions should take race into account."
- "[Opponents of affirmative action] claim that making room for an applicant of one race necessarily requires a university to reject a different applicant."
Don't supporters of AA claim this as well? Isn't this the *whole point* of AA?
- How does affirmative action in college admissions actually work? Are there any concrete examples of how admissions would look with and without affirmative action?
Sort of. What we have from discovery are (1) evidence a lot of this was just open racism from admissions committees and (2) theoretical class statistics in a counterfactual world where things like personality scores were removed.
- Do we know specifically what changed in UC admissions policy after Prop 209?
Yes. There was a brief surge in white and Asian percentages (and to some extent women over men) before it returned to normal. One of the UC administrators was recently caught on tape admitting that they did this by illegally ignoring the ruling and simply avoiding keeping any records of it. So now he's probably getting removed and there's probably a lawsuit incoming.
- How was Prop 209 enforced? If admissions policy is totally opaque, what's stopping the university from "considering race"?
- How was the Bakke 1978 ruling against racial quotas enforced?
Poorly. See previous example. However, now a bunch of people have cause of action against the UC system. Remember, though, that discrimination is not an intent issue. If you can simply prove disparate impact (such as through admissions numbers) that is sufficient.
- How will the SFFA v Harvard ruling be enforced?
Probably about the same. There's already a bunch of administrators walking right up to the line of saying they're going to find work arounds. Which probably means more lawsuits.
- Do people understand what affirmative action is? In this Pew survey, a plurality said affirmative action is "a good thing", but "only 17% say college admissions should take race into account."
Affirmative action doesn't just apply to college admissions. It's perfectly consistent to argue that you want (say) a representative bureaucracy but not representative college admissions. In fact this is more normal. In places like Singapore or Lebanon it's considered necessary for social peace to make sure there are minority administrators and politicians but who gets to be a doctor or engineer is entirely race/religion blind.
- "[Opponents of affirmative action] claim that making room for an applicant of one race necessarily requires a university to reject a different applicant." Don't supporters of AA claim this as well? Isn't this the *whole point* of AA?
They deny it because if no one is harmed no one has standing to sue. Of course people are in fact losing out on slots and so do have standing. Supporters just don't like that fact because they want to be able to do racial balancing (which was illegal even before this decision) and they're going to say whatever they need to say to try and get it past the courts. Freddie DeBoer's a good source of honesty: he believes it's a reparations program and so do most of its supporters. But such racial programs are illegal and so they (broad they, not Freddie specifically) lie to get around that fact.
The long term goal here, as they tried in California, is to repeal the parts of constitutions that require race neutrality so they can do it legally.
- Does the "ban on affirmative action" mean that it would now be illegal for a politician to (as Biden did) say they will appoint a black Supreme Court justice? Would it be illegal for a corporation to deliberately appoint a black board member so their board looks more diverse (if someone could prove in court that this was their logic?)
No to the first question. Politicians have wide latitude to select people for whatever reason they wish. The second one is already illegal unless the business can show a compelling interest in why having a member of that race on the board is important. And diversity is now not an acceptable compelling reason.
The literal phrase diversity is downstream of legal decisions saying that while racial discrimination was illegal without a compelling interest that diversity was a valid compelling interest. Because these people are lawyers or lawyer adjacent they just copy-pasted (in some cases literally) the justification the Supreme Court said was allowed.
"discrimination is not an intent issue. If you can simply prove disparate impact (such as through admissions numbers) that is sufficient."
That is certainly not the case re employment discrimination. Rather, disparate impact establishes a prima facie case for a violation, and then the employer has the burden of showing a legitimate business reason for the policy that creates the disparity. Note also that the individualized use of race endorsed by the court will likely result in disparate impact; a black applicant is more likely to be able to show how race affected him than is a white applicant.
"diversity is now not an acceptable compelling reason."
Not necessarily. After all, the court left open the possibility that diversity could be a compelling interest for military academy admissions, so there was no holding that diversity is per se never a compelling interest.
1. Correct, once you have shown disparate impact you then have to find a safe harbor for why the disparate impact is allowed. If you can't then it's discrimination. The previous safe harbor was diversity. That is no longer allowed.
2. The opinion specifically pointed out this is not allowed and that the theory advanced by the liberal justices it might be allowed is wrong.
3. The court simply pointed out that it was not being asked to consider military academy admissions and that it wasn't certain that appointment to military academies, which is done through politicians much of the time, are the same thing as general university admissions.
"2. The opinion specifically pointed out this is not allowed and that the theory advanced by the liberal justices it might be allowed is wrong.'
I am not positive what "this" refers to, but the decision says: "At the same time, as all parties agree, nothing in this opinion should be construed as prohibiting universities from considering an applicant’s discussion of how race affected his or her life, be it through discrimination, inspiration, or otherwise." It seems to me that a policy implementing that suggestion will likely have a disparate impact.
"3. The court simply pointed out that it was not being asked to consider military academy admissions and that it wasn't certain that appointment to military academies, which is done through politicians much of the time, are the same thing as general university admissions."
What the court said, in footnote 4, was this: "The United States as amicus curiae contends that race-based admissions programs further compelling interests at our Nation’s military
academies. No military academy is a party to these cases, however, and none of the courts below addressed the propriety of race-based admissions systems in that context. This opinion also does not address the issue, in light of the potentially distinct interests that military academies may present" That certainly sounds like they are saying that diversity might constitute a compelling interest in the context of military academies, and hence that the court did not hold that diversity can never be a compelling interest. And there might well be other instances in which diversity might conceivably be a compelling interest, such as hiring police officers in a diverse community. I am not saying that that is likely, but only that your statement is too broad.
Does the "ban on affirmative action" mean that it would now be illegal for a politician to (as Biden did) say they will appoint a black Supreme Court justice? Would it be illegal for a corporation to deliberately appoint a black board member so their board looks more diverse (if someone could prove in court that this was their logic?)
Re the Supreme Court appointment, no. AFAIK, no court has held that the Fourteenth Amendment limits the appointment power of the President or state executives. Re corporations, board members probably are not employees for the purpose of federal anti-discrimination law, so I doubt there are any limits on such appointments.
"If admissions policy is totally opaque, what's stopping the university from "considering race"?"
Being able to prove from other evidence that racial discrimination is happening. Completely coincidentally I'm sure, there's a (successful) push to get rid of standardized testing and other "objective" measures of academic merit.
The Frankfurt Kitchen was the first modern kitchen - designed with flat, uniform countertops, an electric stove instead of a hearth, and a layout carefully calculated to minimize the number of steps between each action in cooking. The architect, Margarete Schütte-Lihotzky, took inspiration from the compact kitchens on trains or ships and assembly-line management ideas (very modernist). It was also politically motivated - she hoped that it would liberate women by allowing them to finish their household work quickly and have time for themselves.
(The podcast also discusses a related architecture-as-politics idea - designing homes with no kitchens of their own. The cooking would be done in communal kitchens, elevating it from unpaid "women's work" and reducing the amount of labor needed overall. Unlike the Frankfurt kitchen, this didn't catch on.)
And on the one hand, the political dreams didn't pan out. The labor that women saved mostly went into more household labor - now that you can cook really quickly, you're expected to make fancier meals, or spend more time on other chores like ironing. And there were also some issues with adapting to the new design - the isolated design made it harder to watch the kids, the electric appliances were a liability if you were too poor to pay the bills, etc. And this was in a public housing project, so if you had an issue with the layout you didn't much choice. (Once again, High Modernism packs everyone into evenly spaced rectangular grids!)
But on the other hand, the Frankfurt kitchen is the ancestor of all modern kitchens. The basic ergonomic design is still around, and everyone designs around electric appliances now. Architects treat the kitchen as the centerpiece of the home instead of writing it off as something for the women to deal with. And of course, women do indeed work outside the home these days, and labor-saving appliances are a part of that. So maybe give Modernism some credit for this one?
I'm having a hard time imagining that before the Frankfurt kitchen, the layout of kitchens was not designed to maximize the ease of cooking. Putting things together that are frequently used together is a basic organizational principle. What were kitchens of the past designed for, if not that?
Besides what others have brought up already, the big difference was a systematic effort to apply time-and-motion-study analysis to home kitchen design. Things had been optimized before, but through ad hoc creativity and analysis by individual householders rather than an attempt to design and popularize the One True Kitchen design. It's more or less the difference between learning to play a complex strategy game by everyone playing around with it, seeing what works, an occasionally copying someone else's strategies; versus someone making a bunch of spreadsheets based on the game mechanics and trying to work out optimal "meta" build orders and unit designs.
I suspect it's less the layout itself, and more about furnishing the room specifically for cooking - having countertops instead of tables, and building those counters with dishracks and places to hold utensils, and so on. The Frankfurt kitchen was designed for a new public housing project, so they had the opportunity to mass-produce furnishings for a new style of kitchen.
The ancestor room / area of the ‘kitchen’ served many purposes, not only cooking.
Heating of the household, and managing the movement of fuel, air, and radiant heat, was an important role of traditional hearths and early ‘stubes’, as an example.
So you can imagine the optimal arrangement for some purposes might be different from the optimal arrangement for other purposes, such as water storage and access, or food preservation, or cooking a meal to be eaten the same day.
That does make sense, it didn't occur to me that what I think of as a kitchen didn't necessarily exist prior to the Frankfurt kitchen. Thank you (and also the other commenters) for the explanation.
There are other changes, both social and technological, that happened at the same time:
-Historically people who were wealthy enough to live in a house with a separate kitchen also had servants to cook for them. So the kitchen would be "below stairs" in the servants' area of the house.
-In a house where some but not all of it has running water, the kitchen needs to be somewhere that has it. If none of the house has running water, the kitchen needs to be somewhere where water can easily be brought.
-If solid fuels like coal are being used to cook, the kitchen will be somewhere that's close to where the fuel is stored. You also have considerations of exhaust, and possibly proximity to larders.
Do libertarians or other heavily market-oriented types have a more free market solution for providing social welfare programs? Right now retirement payments (Social Security) and healthcare for the poor & elderly (Medicare & Medicaid) are the largest parts of the US budget by a huge huge margin, and I would assume the same for other countries. The simple benefit of the private sector somehow providing these goods is simply saving the federal government a staggering amount of money.
I mean, we know how to have the government provide them- the Nordic/Western European model where income tax rates are 40-60% for the middle class, plus a VAT of 20+%. I don't have a huge philosophical objection to this, but it doesn't really seem like this is the final form of government that humanity has invented- I kind of doubt 'the government has to take half of everyone's paychecks forever' is like a Fukuyama-ish end of history and we will never invent a superior governance model. If it's the 25th century and we're zipping around to Alpha Centauri in a warp drive, will the government still have to take half of everyone's paychecks in order to provide the social welfare goods that citizens demand? Seems a bit unlikely?
So- what are the more market-oriented solutions? Forced 401k savings for everyone as opposed to Social Security? (Don't some countries like Singapore already do this?) Mandated private sector insurance program covering all citizens? I think in practice you do need a government mandate for people to purchase these, otherwise free loaders will refuse and then turn to the state when they need medical care- I don't think this is a solvable political problem otherwise
The main theme in that answer is arguably that whether your market is free or not, the libertarian notes that you will have to give people an incentive to want to provide services, for health, welfare, or otherwise. You will also want to give them an incentive to improve that service, either by raising the quality or lowering the price (or both), and since it's impossible to make anyone want to work for less money in a vacuum, you need another incentive, which turns up in the form of competition.
The problem with a government-forever system like you describe is that there's almost no incentive to improve. There's one provider, it gets paid regardless of how well it does, so its remaining incentive is to exert as little expenditure as possible to avoid revolution. Elections only matter if providers themselves are elected, and it's empirically harder to improve or maintain high service than it is to simply change the election rules to make their positions immune. So the entire system calcifies. All ailments treatable today are all there will ever be. Same for all financial troubles. Assistance is impersonal and austere; see the Department of Motor Vehicles in large cities for a model of what the optimized end will be.
I like all of them, but 4 is the classic area where government regulation is needed. Providers deliberately don't publish prices now so that consumers can't compare them and price-shop. It's the same reason every B2B company doesn't list their prices on their website, you have to call & chat with their enterprise sales guy- they don't want to be commodified and have their prices driven down.
I mean- 'Price transparency makes the market more efficient - and less expensive'- you get why existing providers are not going to do that willingly, right? It requires government action
If an outbreak of Teutonic Bronyvirus-K happens, Medicare announcing it will pay for the first $200 of treatment (or everything over $5000 or the middle $1000 on alternate Wednesdays or whatever) will do bugger-all if everyone with any expertise in immunology knows that it will cost more than that to develop any sort of vaccine against the Teuts.
Generally, anyone sick from something new, is screwed, because there's no incentive for anyone knowledgeable to cure it, by government decree.
Moreover, anyone who suspects a way to generate insulin for only $3/kilo won't bother to look into it any further if they know the government rate is higher and they'd just be leaving money on the table or competing with Uberpfizer. In other words, if it costs $X to fix your health problem, it will always cost $X. If a company knows how to do it for $X/2, they'll still charge you $X. And they won't tell you except in private, so that anyone else who might know how to fix it for $X/2 won't try, because they don't know what X is.
Generally, the current state of the art in medical will be frozen in time.
Heavily market oriented doesn't necessarily preclude government action in this area. Libertarians disagree with these government actions but for philosophical reason separate from their market orientation (though there is some overlap). Milton Friedman suggested a negative income tax as a replacement to many social programs. https://en.wikipedia.org/wiki/Negative_income_tax
Charity (including churches and non-religious groups), family, and community would be the answer to how to provide these services without any government action. Groups like the Salvation Army or St. Judes do this already
I'm not so sure about that. NIT scales with income and is only available to people making below a certain level. Isn't the main point of UBI that everyone receives it and its the same amount for everyone? Those are two major factors that change a lot of dynamics of either proposal (not to mention that UBI would probably be taxed in the US while a NIT would be a tax refund, more or less, so after tax).
Combining an income tax where you pay $T(I) depending on your income and a UBI where you unconditionally receive $U is the same as a single, possibly-negative income tax of $T'(I) = $T(I) - $U
I would think the only real heavy hitters here are finding some way to reduce healthcare and housing expenses for the elderly (or ideally, for everyone). Regardless of how the payment is provided, as long as it remains expensive, you achieve only abstractly moral benefits over just taking 40-60% of everyone's paychecks. Maybe they can find some more voluntary means of getting that money, but it would still divert an enormous volume of the economy toward keeping people who are past working age healthy and off the streets.
I supposed the tech bro stereotype thing is replace all the doctors with GPT-15 and house people in Tokyo-style highrise cubbyholes with a hospital across the street.
We have great solutions to both of those problems already, we just reject them as unthinkable. For housing, we can tell people to move to cheaper places (which Libertarians may tell you is already what the market is telling people who cannot afford their current housing). Old people on Social Security will get much further in random rural places than big expensive cities.
Healthcare is even easier, but much more fraught. You run an analysis of life expectancy compared to cost of treatment and let people die if they are too expensive.
I don't advocate these options, but everyone knows they exist and they would be pretty effective. There's also no other way to lower either of those costs more than marginal amounts.
If you get rid of forced saving you need to be willing to let the people who fail at this actually fail.
And to tolerate the impacts of that. Mostly these days people are not interested in watching grandmas starve in the streets in their 60s even if the grandmas were extremely profligate and short sighted with their money for the 40 years before that.
So if we are going to be forced to take care of people then they need to have a certain level of responsibility forced on them.
To me it seems like the only libertarian solution is charity, i.e. consensual giving. A market-based approach would be to auction off those who are unable to take care of themselves as wards to the lowest bidder. Forced saving mostly just seems like a way to pretend that taxation isn't taxation.
Indentured servitude only works for people whose labour is worth more than the cost of keeping them alive. Welfare is for the people whose labour isn't. https://en.wikipedia.org/wiki/Fattigauktion describes how this worked in practice.
It's a weekly link aggregation of projects, papers, news, repos, demos and products in the "LLM space". Basically for Machine Learning Engineers or AI Engineers or Prompt Engineers however you'd like to call it.
Recently I've covered a lot of what's happening with Open Source language models that you can (mostly) run on consumer GPUs.
Why do I want more readers? The more people I reach the more I'm motivated to invest time & resources in it. I like the flywheel.
Anyone need a top-tier graduate student in either Economics or Computer Science? I have one to recommend! He's extremely smart and self-motivated. But he's unfortunately from a developing country, so he has extremely limited opportunities, so I'd like to get him into a program in the global north.
He's the kind of student of whom you can ask "I need to apply a Wald estimator to this data, but don't have time. Can you take a look and do it this week?"... and then he came back THREE HOURS later, saying, "I figured out how they work, and went ahead and wrote the code that applies it to your data files".
Why do attendants at full-service gas stations insist on topping off your gas tank? As long as I've been alive, the general recommendation has been that it is a bad thing to do, because it screws with internal portions of how the fuel system works in various ways that I don't understand. It also may be a waste of the extra gas. And yet even if I ask the gas attendant not to top it off, he usually will still do it about half the time, maybe just because they're used to doing it and it's a muscle memory thing.
My question is, why do they even do it at all? Every single official recommendation from all auto manufacturers is not to do this. So do the attendants do it because they think customers want it? Do they know that it messes with the fuel system? Are they just trying to squeeze out a few extra bucks out of each customer?
> Second, with a completely filled tank, you drive around some extra weight in gas, which will raise your consumption.
That's a very interesting take on it! However, I expect that, for myself at least, any minor cost saved by driving around with slightly less weight with fewer gallons would easily be outweighed by:
1. The amount of extra miles I'd be driving to and from the gas station more often
2. The monetary value of my time spent driving to and from the gas station
3. My own stress about having to fill up more often and procrastinating on it
I am curious, though, if you have any figures for how much gas consumption might increase per pound that the car carries around.
When I fill my tank, I wait 15-30 seconds after the pump auto-stops, then start it again. I can get another .2-.5 gallons in there. This is due to air escaping from the tank. I guess you can call it topping off but I don't think there is risk of damage to the system. I have a 14 gallon tank and if i drive it until the estimated range is below 10 mi, I usually only get ~13 gallons in before the pump stops itself.
Boy, in the US full service is pretty rare since, oh, 1980 or so. They just suddenly started making most of us pump our own gas. It was a pretty abrupt change.
So here is the last appearance of a full service gas station in a major American movie. Dan Ackroyd offers to wash the dead bugs off Twiggy’s windshield in the Blues Brothers 1980. (don’t worry it’s just a 24 second clip)
Unless something's happened since I last heard, I'm pretty sure they only relaxed it for rural areas at night, to make it easier to stay open (having struggled to find an open station driving through Clatskanie at 11pm with the needle kissing E, I'm in favor of this).
From Zip recruiter: "As of Jul 2, 2023, the average hourly pay for a Gas Station Attendant in New Jersey is $15.76 an hour. While ZipRecruiter is seeing salaries as high as $20.16 and as low as $8.85, the majority of Gas Station Attendant salaries currently range between $14.76 (25th percentile) to $18.46 (75th percentile) in New Jersey."
Why do they do it?
Because likely they are general poorly trained and selected from the bottom of the barrel.
I could be wrong, but I thought topping off your gas was only bad if you did it poorly, like extra gas leaked out from the full sensor not picking that up because the nozzle was inserted incorrectly, or something. I thought it had nothing to do with how the car runs, and I don't see how it could.
Of course, it is assumed professional full-service attendants would not have this problem, and you can go a little bit farther without needing a fill-up.
They've almost certainly pumped gas more times in the past week than most of us do in a year. If there is *any* amount of practice that helps with doing a feature of the job, they've got more of it. It's only if you think that there is *no* way that practice could improve that you wouldn't call them "professional" in the relevant sense.
All work has dignity but profession and professional have a meaning and the meaning is not synonymous with being employed even for a long time at the same job.
Surely no one working above a minimum-wage job believes that all work has dignity. The other day I watched a Target employee clean up some piss from an "emotional support" chihuahua a customer brought in. I'd love to watch a Gandhian explain the dignity in that to me.
I guess you're asking questions rhetorically, since you seem to have already rejected any answer other than your own definition. One definition of "professional" is someone with a higher degree, like a doctor, lawyer, engineer, architect, economist, etc. Another definition is simply "someone who gets paid to do that thing." If you make music to entertain yourself and your friends, you're an amateur musician. If you land enough paying gigs to support yourself, you're a professional musician. If you put tile just in your own bathroom, you're a do-it-yourselfer. If your job is going around tiling other people's houses, you're a professional. I'm not sure why you're quibbling with simply narrower or wider definitions of the same word. Tea can mean any beverage brewed with water and some plant material, or tea can specifically refer to Camilla senensis. It's not super complicated.
It does have to do with how the car runs to some degree, and the worry is not just about extra gas spilling out. I've heard several explanations, but I don't know much about cars so I don't really understand any of them, or know how much extra gas would cause the damage. But there's are concerns about the extra gas causing a cracked tank, or getting into the car's charcoal canister or carbon filter, which usually are only supposed to get gas vapors, not actual gas. Supposedly this can cause engine damage.
I'm no car person, but in order to crack the tank or something the gas should have be be under pressure, and I don't see how that could be with how the nozzle inserts into the tank. There's some kind of metal seal, but I don't know how tight that seal is, or why the gas cap would need a tight seal if it is already well-sealed, and in any case, I think the gas would equalize the pressure as the nozzle is removed.
Note I'm explaining how I don't know, not claiming expertise. I would be grateful for an explanation proving me wrong.
I'm just explaining what I've seen people say online. I can't speak to anything's veracity, because I don't understand cars. I'll bet a cracked tank is rare but does happen in certain circumstances, but then why even take the chance?
Also what about the carbon filter or charcoal canister issue?
I'm kind of a car person, and I do understand cars. At least I've restored some cars, trucks and tractors from the ground up. I have lots of friends who are fulltime mechanics currently working in repair garages. I asked them, and none of them has ever heard of a "cracked tank" or other malady from completely filling the tank.
It's like a lot of things with a claimed potential for harm, but the actual evidence for genuine damage is practically non-existent. Like your cell phone harming a jet's avionics. People top off their tanks so frequently, often just to round off the change they have coming back, or to get to the exact amount they pre-paid for, to keep from having to walk back into the convenience store and wait for their change, that if there really were significant problems from a few extra squirts of gasoline, where are all the damaged cars? If this was a real thing, there would be real cars brought into repair shops every week all over the country. There aren't.
There's got be more significant things to worry about than this hypothetical issue.
Okay, forget cracked tank, maybe that one's not real, I haven't seen it cited too much, so I could believe it's fake. But Google for "should you top off gas tank", and almost without exception, every site says no, you shouldn't. Most commonly, they seem to indicate that it can irreparably damage your car's "vapor recovery system", and possibly also cause pollution.
Affirmative action: this seems really simple to me— create a race-blind application process that records SES information. Take every decile of the SES range and make a pool of the top 10% of students from that decile. Take an equal random sampling of students from each pool. I bet you would end up with a pretty darn racially diverse student body, with the bonus of also having economic dIversity. If this isn’t “elite” enough, make your pool from the top 1% of the first SES decile, top 2% of the second SES decile, etc.
Even simpler: selective admission policies are obviously more aligned with the institution's interest than the public interest. Institutions that use such policies have self-selected into the Hillsdale College path of no federal funding. Adult education often does scale with technology, and the default policy should be admission of all qualified applicants (where 'qualified' = 'demonstrates the prerequisite skills to succeed at the program'). Or where physical constraints prevent open admission of qualified applicants, a lottery of qualified applicants is an acceptable fallback.
The most likely outcome of an SES schema like you describe is that all of the best white and Asian students dominate at every decile. Black and Hispanic students will be heavily reduced. There are lots of poor white and Asian people, many of whom are higher-achieving than black people at the same SES. We see consistently that white test scores outpace black test scores at the same SES levels (and several levels above), such that rich black people have similar or worse testing scores to poor white people who in turn far outpace poor black people.
I'm in favor of race-blind applications, but those interested in AA will never accept it because it would mean even fewer black applicants get into elite colleges.
I was under the impression a fair number of states already did this, guaranteeing admission for certain state schools to anyone who is in the top 10% (or possibly some other percent) of their own school's graduating class. Assuming here that, for the most part, people in the same school have similar SES and this divides students in roughly the same way.
Some do, and also grant guaranteed scholarships to the top x% of each school, but this may be more common in states where the state university is average or below, not for the top ones. New Mexico does this, for instance, but UNM is truly a state college, for people in the state to learn things, not very useful for signaling, or drawing many people from different states or countries.
Rather than optimizing for economic diversity, how about ideological diversity? Rather than splitting up the population by economic deciles, how about having all the candidates be classified by one of the 2D political ideology measures? Then split that 2D space into bins, and pick from the bins in proportion to the fraction of the overall population in each bin (with the same top 10% from each bin).
Because the desire the universities have expressed is to achieve racial diversity, but they cannot achieve this outcome by discriminating on the basis of race. But, since, as I understand it, family wealth/SES is heavily correlated with race, using that as a proxy could perhaps achieve similar results.
The usual meaning "acting or done in a secret or dishonest way".
In this specific case, choosing a proxy for race, so that the university can continue to racially discriminate against white and asians and in favor of blacks, SCOTUS or no SCOTUS, while concealing this discrimination under a proxy which purports to be racially neutral but is actually chosen precisely because it is racist.
Punishing children for the hard work of their parents sounds awfully unfair to me, as does rewarding children for having parents who know how to game the SES calculation, and I think a lot of people would agree with that assessment.
I mean, so does allowing children to benefit from the hard work of their parents. If you believe in equality of opportunity, then students shouldn't be able to have their parents buy tutoring services.
I disagree. I think parents supporting their children is entirely fair, though I don't think it has that much of an effect, and it's presumably in the interest of schools to have entrance exams that measure underlying ability rather than the amount of tutoring received.
There are lots of easy solutions. The problem is people are wedded to the current bad solutions so they are not interested in different ones.
Also lets say you do what you suggested and in the next class there are 0 Latinos lets say. You program gets stopped immediately because academia is 60% is racists these days who only see skin color and don't care about the people underneath.
That seems interesting to see you use the term "racist" for someone who sees the complete absence of a major group and assumes that this is prima facie evidence of a hidden unfairness. Are you really so convinced of the absence of any possible unfairness that you want to use the word "racist" for this?
If you just mean to use the word "racist" in a morally neutral way, then maybe that's fine. But it should then also be fine to use the word "racist" for the kind of structural injustice that these people want to talk about.
I am using "racist" for a group of people who judge others based on skin color uber alles and fail to consider people's individual backgrounds or any other pragmatic considerations besides fractious counter productive identity based nonsense.
Universities are the late medieval Catholic Church/monasteries of today, and the sooner the institutions can be torn down and reformed the better. They fail at both of their main goals (educating citizens and conducting research).
For the amount of money plowed into the system there is incredibly little return. Its too bad the MOOC model didn't work, as that would have been a huge boon to society.
Anyway when a group of people makes it clear they are the enemy to you, to truth, and to fairness, it is best to treat them as such.
I'm sure there is a social scientist who can explain that there is no "person underneath." After all, people only exist at intersections of oppression axes, n'est-ce pas?
Even if this worked as intended, it would still create a huge black underrepresentation (especially among descendents from slaves).
But it wouldn't work since the rich and educated would game the system. "Record SES information" is not a simple atomic task. What's the SES of a person who's parents used to be high-earning doctors but that are currently working for minimum wage at a local nonprofit? What's the SES of someone who claims to be disowned by their rich parents and that are receiving the dole living in the poorest neighborhood?
My guess is that those are pretty rare edge cases. Most parents who have money plan to use it to help their kids pay for college. I mean there could also be some poor kids whose uncles have won the lottery and plan to pay for their college. I think that the marginal impact of these scenarios is pretty small.
They are rare edge cases right now, but if doing a year of volunteer work increased your kids chance of getting into Harvard by 860%, it would quickly become very popular. There would be industries of consultants who design the best strategies for upper- and middle-class parents to minimize their SES score (just like how there's already an industry like this for our current university admission system). This is blatantly obvious to me, I'm a bit perplexed that you don't seem to get it.
This is already a thing in Italy where universities give free accommodation and/or free cafeteria access and waive tuition to the poorest segment of the student population. Anecdotally, rich parents who care enough about this manage to temporarily hide their wealth through accounting tricks. It is however generally considered that this is just a price society has to pay to support the students who are the intended target of these measures, and the fact that there are some free riders is not perceived as a big deal (we are not talking admission here).
"SES" is not a well-defined thing. Neither is "SES of the applicant’s parents". You will have to set up an algorithm calculate the "SES score" of someone. You might use parental income or capital or postal code or whatever, but you need something concrete. Whatever algorithm you set up, people will game it, and since SES is very mutable it's likely that the algorithm will be easy to game. Once again, I'm perplexed that you don't see this.
If you specify how the SES score of an applicant will be calculated, I'm very confident that I can find way for rich people to easily game it.
Once again, what's the SES of a person who's parents used to be high-earning doctors but that are currently working for minimum wage at a local nonprofit? What's the SES of someone who claims to be disowned by their rich parents and that are receiving the dole living in the poorest neighborhood? Don't you think applicants will manufacture these circumstances if it increases their acceptance rate by x10?
Where is scholarly aptitude in this? Aren't you assuming everyone is equally qualified, and only divided by SES (social-economic status, unless my Googling came up with the wrong meaning)?
Scholarly aptitude is how you create the pools of top students from each SES decile. Accept a random sample of students from the top X percentile (in scholarly aptitude and any other combination of desirable factor) of each SES decile.
Isn't it true that poorer people have less opportunity to learn? It may be that the top 1% if the lowest bracket has less aptitude for academic success than everyone in the top bracket, simply from never having been exposed to the curriculum before.
I think this is a difficult problem to solve, for many factors enter into it. No one is actively prevented from a good education, AFAIK, but still some don't get the opportunity.
If there is less opportunity to learn, that would mean that the student is less academically sound and capable, and therefore wouldn't belong in a higher education institute in the first place, *unless their test scores show otherwise.* Maybe... just maybe... we should select students based on test scores! How radical.
"It has been said that if an aborigine wrote an IQ test everyone in civilization would flunk it. "
First of all, they wouldn't. We certainly have more general knowledge than an aborigine. But even if we did, it's because the test would be designed for a world that we don't live in, and certainly not a world that secondary education would prepare you for.
"How is someone supposed to get a high test score when they haven't the opportunity to learn and practice what might be on it? "
They aren't expected to. If someone doesn't have a solid primary education, they don't belong in a secondary institution, until they develop that knowledge.
I went to the Philly Rat Fest on a whim and had a good time. I didn't know anything about critical rationality (based in Popper and Deutsch) nor that it was critical rationalists rather than a gathering for rationalists in general. Any recommendations for critical rationalist online places or sources in addition to Deitch? I like the optimistic attitude and the way human experience gets centered. I don't think Bayesians (a fair name for rationalists who aren't critical rationalists?) completely ignore human experience, but they can get very abstract. Those questions of whether humans are even a good idea....
One of the talks was about any early image breeding program. Does anyone remember the name? It offered random shadings, and people could "breed" them to make interesting images. The point of the talk was that sometimes exploration according to personal taste can pay off better than deliberate problem-solving.
I've started reading _The Fabric of Reality_ I'm just on the chapter about about explanations being of central importance and predictions being no substitute for explanations. I'm not sure that Deutsch's thought experiment of a perfect oracle compared to having explanations is fair. I think his point is that If you had a perfect oracle, you still couldn't use it alone to build a spaceship unless you understood enough to ask the right questions.
I do think that wanting (settling for?) predictions without explanations might be part of a pattern of trying to make conscious thought unnecessary.
I'm not sure whether Deutsch makes this explicit, but if all you have is a prediction, you have to guess about how much scope it covers.
I do have a real world example of what can happen if you just want predictions. If someone gave you three great stock tips or predictions of horse race winners, what would you pay for a fourth?
In this context, you might see it coming. This is a scam. The scammer sends out different sets of predictions to different marks. Some of the sets of predictions will be correct. The mark is willing to believe in a secret system.
Deutsch is heavily influenced by Popper. They both believe that scientific progress goes like this: first conjecture, then refutation. Deutsch rejects "instrumentalism", the claim that all that matters in science is the prediction of experimental outcomes. He believe, then, that what "matters" is the explanation (a conjecture), the experiment being a means to reject bad explanations (refutation).
So, yes, if an alien species gave us a perfect oracle, for example a computer that predicts the outcome of any experiment we can describe, we would not stop doing physics (or science in general) because we would still wish to understand why things work the way they work.
The instrumentalist view is surprisingly common, even from people unaware of these distinctions. It permeates the Copenhagen Interpretation of Quantum Mechanics, for example.
is an argument that an underlying problem in the replication crisis is that a lot of science doesn't have an underlying theory and therefor can't be properly tested. It's a similar to the complaint (I think?) you're making that a sufficient number of people making sufficient predictions will eventually predict the entire works of Shakespeare.
My dad has a wrist injury where he can’t put any compression on it for the next few months, but he sorely misses weightlifting and is going a bit stir crazy now that he can’t do it. Does anyone know of any devices that you can put on the forearm which can have dumbbells or barbells attached to it?
Many of the exercise machines at gyms allow you to exercise your upper arm, shoulders, upper back etc without grasping anything in your hands, or even using your forearms. I think it's probably also possible to substitute exercises with elastic bands for weight lift exercises, adapting the arms exercises by connecting the band to the area right above the elbow, rather than holding the end on one's hand.
When I shattered my forearm years ago, I shifted all of my weight lifting to core and lower body. Anything that could be done without the broken arm like machine squats, leg presses, back extensions, etc. My doctor claimed I had the most successful recovery he had ever seen for such an injury. IMO trying to workout the affected arm seems like a bad idea. Let it heal and do whatever exercise you can elsewhere.
No personal experience. My wrist weights only go up to like half a pound. But unless you can fit your forearm into a cable handle it's the best I can think of for bicep curls.
Wouldn't the solution be to find him something else to get interested in? It's a just a few months. Surely, there is muscle between his ears which he could also work on. Or the heart muscle?
I don't know about Gruffydd's dad but I find that sudden changes in the amount of exercise I get tend to be quite physically unpleasant. It's not an unreasonable request.
Really? You hurt your wrist you supposed to cool it with the weightlifting temporarily. I think trying to find a work around to continue weightlifting is pretty unreasonable.
"sudden changes in the amount of exercise I get tend to be quite physically unpleasant" Well if the change is to do a lot more exercise that could be physically unpleasant. To do less? Physically unpleasant? (Maybe mentally or psychologically unpleasant if you weirdly conditioned yourself.)
I am recalling my father in the 70s seeing an adult jogger and remarking: he obviously isn't working hard enough.
Finding alternative exercises that don't use the wrist doesn't seem like it wouldn't actually have much risk of worsening the injury if it's not even involved, so I don't see why that's unreasonable.
Yes, doing less is physically unpleasant. This is probably restless leg syndrome and I don't know how well it generalises.
He still wants to lift weights with his arms. That is unreasonable if you have a wrist injury.
RLS, not mentioned in connection with wrist injury, being idiopathic would not be generalizable to anybody including someone else with an RLS diagnosis.
Since I basically only have my own experience to go on, I assumed that most people would get at least some symptoms upon discontinuing an exercise routine even if it's not the same sort, but if that's not the case, fair enough.
Is it true that kids today have worse fine motor skills than what kids had say 15 years ago? (Presumably this is because they spend more time today with digital devices.) I've seen this claim thrown around allot but I haven't seen any studies to back it up and I can't find any when I Google.
As some others have said, fine motor skills are very task-specific.
I was in the Scouts, and could tie a knot or make an eel spear. However, the first time I ever used a computer mouse was as a teenager, when a couple of friends and I played the Macintosh game "Dark Castle" (1986). We were terrible at it, so much so that in order to make any progress, we had to split the controls so that one person used the arrow keys (like in a "normal" game at the time) and another could put their full attention on aiming with the mouse.
My kids have grown up using WASD and a mouse, so if they ever experienced something like the above, they were probably too young to remember.
I don't know about studies, but yes, it probably depends on the specifics of what they mean.
Sometimes I attend Elementary art teacher meetings, where the older teachers will talk about how the current generation of little kids don't really know how to use things like scissors, tape, glue, paintbrushes, threading, twisting, play dough, and so on. Some of them may sometimes shorthand this to "fine motor skills," but they mean a very specific subset of those fine motor skills. It's no help to them if kids are way better at operating devices with their fingers, but struggle terribly with tying a knot. Other people would care about different things -- maybe typing vs handwriting or something. I would expect typing to be better, handwriting worse, but am not certain.
I think motors skills are pretty diverse. Kids might be worse at writing, better at working a mouse or touch screen, and way worse at throwing/catching balls.
What does that say about their "fine motor skills"?
Probably depends on precisely what you measure. I think videogames on net improve fine motor skills, but 15 years is a short enough time that one is looking more at how videogame control schemes have evolved than at the shift from playing outside to playing video games; I could believe that touch screens develop motor skills less than controllers
I have been fascinated by the story of Anson's voyage around the world back in the 1740's for years. It was a British attack on Spain's Pacific coast possessions in South America during the absurdly named War of Jenkins' Ear. Eight ships were sent round the Horn into the Pacific. But the weather they faced was truly horrendous and two turned back and one was wrecked. The Spanish squadron sent to stop them was basically destroyed.
My podcast with David Grann focuses on the Wager perhaps the unluckiest ship in the squadron (it was a competitive field!). At the time scurvy was not really understood and the death rate was appalling - both going round the Horn and for the surviving ships as they crossed the Pacific. I believe one of the worst outbreaks of scurvy in naval history.
It didn't help that a shortage of manpower had led to the British making up numbers by putting on board retired and invalid military personnel. Pretty much all of these died before they got round the Horn.
As I say Grann and the podcast focuses on the wreck of the Wager - the subtitle of the book is Shipwreck, murder and mutiny which gives you an idea of what happened to them there. But the story of their various attempts to escape from the freezing, deserted island they were trapped on is the stuff of movies (Scorcese has bought the rights and has just turned another of Grann's books into a movie so may well happen).
The rest of the squadron suffered appalling losses and by the end only a single ship was made it back with some 256 crew. But astonishingly they had succeeded in capturing the Manilla galleon. The treasure was about $80m in today's money and the crew got half. The Manilla galleon itself is a fascinating thing with one voyage once a year from South America to Manilla with silver and other treasure and another ship going the other way with trade goods from China. The Manilla galleons sailed continuously for about 250 years. The British dreamed of capturing them but succeeded only (I think?) four times. Here are my two podcasts (on the Wager and on the Manilla galleons) and David's book:
It's hard to imagine how, if you discount the possibility of future nuclear war, the existence of nuclear weapons has not been an extraordinary boon to humanity. We haven't fought a World War in almost eighty years, or even a direct Great Power conflict of any sort. This in spite of half a century of Cold War between two superpowers, of the sort that would certainly have fought a Hot War in any prenuclear era of human history. The cost of establishing and maintaining the world's nuclear arsenals to date, is trivial compared to the cost of even a conventional World War.
So, yes, ignoring the only major down side of the whole thing, it's been a good thing.
Since you're eliminating the largest negative, it must be net good. Without nuclear weapons we wouldn't have nuclear medicine, I think, or other, possibly less obviously-related technologies.
But I'm not sure the question makes sense. Kind of like, "If knives couldn't be used to kill people, are there any downsides to them?"
I find their existence is net bad, since it creates the possibility of global nuclear war in the first place, which would be rather catastrophic for much of humanity.
- Any real or perceived "gap" in nuclear weapons power inherently leads to mutual buildup, reinforcing itself in a positive feedback loop. This loop can only be stopped by disarmament and inspection agreements, which are subject to revision by whoever is currently in power in any of the signatories.
- The lesson learned by every two-bit dictator around the world is that if you have even one nuke able to reach your main rival, you become basically untouchable, so there is a strong incentive for proliferation.
- The decision whether or not to use strategic nuclear weapons is heavily informed by technology and intelligence, both of which can fail and create a situation of accidental nuclear war, which is in effect indistinguishable from intentional nuclear war.
- On the plus side, they reduce the chance of war between the great powers, but industrialization is already doing that by reducing the incentive for large-scale war in general.
Nuclear weapons, or the nuclear industry in General?
Looking only at Nuclear Weapons, the only good they can do is to deter violence.
Since the only Nuclear weapons used solar were. in Japan in WW2, and caused less damage than the Tokyo regular bombing (Not a historian), and my impression is that they hurried Japan's surrender (but by how much?) You could argue that Nuclear Weapons stopped Violence by the Sheer promise of even greater violence.
The Whole MAD scenario of the Cold War seems to support this, though I am not convinced that with the Alternative history of no nuclear weapons we would have a more direct NATO-Soviet conflict.
I think averting WW3 is really obviously very good, and almost all the downside is in the risk of a future nuclear war. I do think the USSR and the USA would have ended up triggering WW3 absent MAD, too
Also on the positive side of the trade-off: Nuclear weapons made ICBMs make sense, and ICBMs were precursors to access to orbit and deep space. Yes, Goddard and Peenemünde preceded Trinity site, but, in the absence of nukes, I think it would have been a _long_ time before we got weather satellites and spy satellites and GPS (and human spaceflight, but the benefits of that thus far are much more debatable). I strongly suspect that the lives saved by better weather predictions alone probably outnumber the dead of Hiroshima and Nagasaki.
There are weaknesses with the Intelligent Design argument in biology, but those weaknesses don't exist with Fine Tuning argument in physics. For example:
1. There are no fossils in physics.
2. Physics is fundamental while biology is derived from chemistry.
3. Scientists have never observed other universes while they have observed other planets.
If you're looking for the strongest rational convincing argument for the existence of God from science, look to physics. Listen to the episode of Physics vs Biology at https://www.physicstogod.com/podcast or you can find "Physics to God" on any podcast platform.
My favorite 'meta-GUT' is that ALL possible universes exist, so God is entirely superfluous – of course we'd find ourselves (whomever 'we' might be) in a universe 'fine-tuned' for our existence!
That's Max Tegmark's Mathematical Universe Hypothesis. We discuss that in the multiverse miniseries, but it's a very flawed theory which to some degree undermines all of science.
The need for fine tuning is usually taken to show that theories are inadequate or incomplete. Theories often happen to have these issues. I wouldn’t want to rely on these characteristics of fallible human theories to prove the existence or to constrain the nature of God.
Why does it matter? There are plenty of reasons to believe human engineers may one day be able to create universes if our species makes it long enough. That would make us God by this definition, but doesn't make us worthy of worship, has zero consequences for how inhabitants of these universes should behave, doesn't mean we're imbuing them with immortal souls that will survive the death of their physical bodies, or any of the other reasons believers in God typically actually care that there might be a God. We wouldn't even be able to observe what happens in these universes and wouldn't know if life ever developed. If we had a creator, but one who was not infallible, omniscient, omnipresent, or omnipotent, and likely became causally disconnected from our universe within picoseconds of its creation, what is this supposed to mean for us?
There are many reasons why God matters, separate and aside from religious issues of worship. First and foremost, is about truth about reality. Is the ultimate fundamental reality an intelligent agent or is it an unintelligent law(s) that generates an infinite multiverse? Truth about reality has mattered to many people throughout the ages.
It's 2023 and some people still believe in Sky Daddy, or that their souls will one day live forever in a beautiful place, or that some kind of sentient, intelligent design created the universe and the Earth, where 99% of all creatures that have ever lived are extinct? Hahahaha! Sad.
"Why does it matter? There are plenty of reasons to believe human engineers may one day be able to create universes if our species makes it long enough."
I can't think of any reason to believe that, other than "Humans seem to get better at making things over time".
This does not deal with the most important piece of evidence from fine tuning.
We observe that there are a handful of physical constants that would have to be within some pretty narrow ranges for this universe to look the way it does, namely, with us in it.
There are two hypotheses to explain the apparent coincidence: a creator, or anthropics.
If the answer was anthropics (that is, for us to make the observation we must be in a universe with parameters in certain ranges, and no observations can be made in universes that don't) then we should expect the minimal amount of fine tuning necessary for us to be able to exist and make predictions.
If the answer is a creator, then we should expect omnipresent fine tuning according to precisely whatever the goals or values of the creator are. Not just a handful of parameters, but all of them, and everything else. The more the universe looks obviously tuned and designed, the more evidence for a creator.
Therefore, minimal fine tuning is evidence for anthropics, and maximal fine tuning is evidence for the creator.
We look at the universe and see that is extremely minimally fine tuned. The only way to say this is evidence for a creator is to say that the creator just so happens to want exactly the same minimal fine tuning that anthropics would end up in. And even so this gives it precisely zero explanatory power _more_ than the anthropics hypothesis. This is worse in terms of Occam's razor and is obviously motivated reasoning.
>There are two hypotheses to explain the apparent coincidence: a creator, or anthropics.
There's also a third hypothesis: deeper order. Under current established physics models (General Relativity and the Quantum Standard Model), the constants in question appear to be free variables that could be anything but happen to be fine-tuned to allow the existence of atoms, molecules, stars, planets, and complex planetary life. But we know those models are incomplete, and we don't know for sure what's really going on under the hood. Maybe there's some deeper relationship between the now-seemingly-independent physical constants that explains what now appears to be fine tuning.
Agreed, there's always the possibility the apparent coincidence is no coincidence at all. But with the info we have now, and to grapple fairly with their argument, we can exclude it.
"We look at the universe and see that is extremely minimally fine tuned."
Nice! I was aware that this is the feature one wants to test, and I was aware that in the case of the Higgs field value, it indeed looks minimally fine tuned (in fact looks likely to be metastable). Is there a url that documents the minimal fine tuning of the other Standard Model parameters? Many Thanks!
You are right up until a certain point. When we look at the universe we find that there is way more order than is necessary for an anthropic interpretation. Richard Feynman made this point and it goes back to Arthur Eddington. The minimal amount of order needed to explain our observations is a single brain with all your (false) memories surrounded by a universe of total disorder. But in fact we find that the universe is filled with order, complexity, and structure everywhere in it we look. This is called the Boltzmann Brain problem. We'll discuss that in our series on the multiverse.
This is resolved by a proper application of Occam's razor, or Salomonoff Induction. An ordered brain in a disordered universe is actually more apparent coincidence than simply a slightly ordered universe which begets a slightly ordered brain. The "total amount of disordered stuff" is higher in the disordered-universe-with-ordered-brain, but the simplest (shortest) description is of an ordered universe with a similarly ordered brain. The boltzmann brain problem is primarily used as a tool for comparing and assessing different specific cosmological theories. It certainly cannot be interpreted on its own terms to find a creator more likely than the anthropic interpretation, and the actual expert cosmologists certainly wouldnt agree with it being used that way. Theres no creator in the boltzmann universe either.
And, frankly, the universe does NOT appear to be filled with order and structure everywhere we look. The universe appears frighteningly disordered with tiny tendrils of coincidental order. I cannot stress this enough - the idea that the universe is actually largely ordered is in total contradiction with what any ACTUALLY ordered thing looks like. The universe is almost entirely empty space. Of the mass it contains, the vast majority doesn't even interact with light. At the same time that it contains a maximum speed, it also is expanding from within such that given time the universe will become filled with little pockets that slip entirely out of each others light cone and so can never interact whatsoever forever. These are totally bizarre qualities to suppose a creator would deliberately design for any purpose. They certainly do not appear ordered.
There's the apparent coincidence of the well-tuned constants, there's the still incomplete mathematical descriptions that reality seems to follow, and there is the tendency for competitive spaces to become filled by ever more competitive entities (which Scott has helpfully named Moloch), none of which especially suggest a creator over random chance. I have not seen, and you have not presented (though I admit I have not gone through ALL or your rather prolific work) a fundamentally different kind of order than these.
I think you are mistaken. The second law of thermodynamics says that all physical processes bring a system from a low state of entropy (less probable) to a higher state of entropy. Therefore, an ordered universe which begets a single brain must logically have lower entropy than the single brain itself. It is therefore more likely to get a single brain by chance alone than an ordered universe which begets a brain.
With regards to your second point that you don't think the universe has enough order, complexity, and structure - I understand what's bothering you. On the other hand, it's hard to play God and decide just what the most perfect, beautiful, ordered universe should be like. However, it seems very reasonable to say that our universe is amazing and incredible with wisdom everywhere on all orders of magnitudes - from atoms and molecules, to planets, stars, galaxies, and life. It is certainly much more interesting than a universe filled with black holes which is the most likely state of the universe (a state of maximum entropy). That point is all we need for our argument to work.
Theres a mistake in the logic. The second law states the tendency toward low entropy. But it doesnt ask about initial conditions.
For whatever reason, the universe started as a massive concentration of energy - A big ball of negative entropy (negentropy). From there the universe evolved (evolved here in the mathematical term, that is its change in state over time) into states of higher entropy (lower negentropy, spending it down). The second law absolutely does not preclude any process by which entropy is decreased within a particular control volume - it only precludes a universal decrease entropy. Entropy may decrease here as long as it increases by more there.
At the start of the universe when it was a near uniform distribution of energy in a very tight space inside a massive volume, it evolved into states such that that energy was concentrated into particular clusters of a size determined exactly by the physical constants. (Stars and Galaxies). For eons, those large balls of negentropy slowly spend themselves down fusing elements. All this is described without much requisite "order". Only the constants themselves and the tendency for exclusive spaces to be filled with competitive entities. (Stars and galaxies "sequester" energy into small pockets relative to their zone of influence. This happened because while certainly other structures were possible in an even distribution of energy, anything that didnt sequester would naturally end up sequestered)
In the shadow of a giant ball of negentropy, however, strange things are possible. In the presence of so much negentropy to spend, and enough matter and energy to play around with structures of meaningful size, random chance can temporarily spring up all sorts of incredible "orderings".
The problem of the totally disordered universe with the single brain is describing how it ended up in ONLY and EXACTLY a single brain, along its evolution from a single massive concentration of energy (and therefore negentropy) to what it must logically become. An even distribution with a volume containing a human-brains-worth of particles lined up just correctly is NOT a likely trajectory of that evolution. Not because the alignment of so many particles is so unlikely, but because it is (statistically) precluded by the things that happen first.
So no, it is absolutely consistent with the laws of thermodynamics that a universe which begets a brain can be more likely than a brain existing among disorder, simply based on the mathematics and observable initial conditions.
Finally, I do not agree on the face of it that the universe is amazing everywhere on all orders of magnitude, nor that it is reasonable to say so. It certainly seems more interesting to me than a universe of zero entropy, where all energy is exactly evenly distributed. But then, I can't exist in that universe.
The one we inhabit certainly does not seem optimal by any metric. I certainly have some complaints. Dont get me wrong Im enjoying being alive. But it certainly does not seem designed to be amazing or incredible or with wisdom everywhere or to be interesting, the criteria that you hang your hat on for your argument to work. But lets note - weve totally gone away from the physics. You are appealing to human values. Gesturing at some cosmology we havent figured out yet, saying, "look! theres room enough inside this confusion to allow you to believe in God and wouldnt that be nice since everything is so amazing and incredible and full of wisdom".
You can believe whatever happy things you like but it certainly isnt coming from the physics.
The second law of thermodynamics states the tendency toward higher entropy, not lower entropy. Perhaps you just misspoke.
Also, the second law clearly indicates that the initial conditions of our universe had very low entropy. This is because of the apparent contradiction between the observations that lead to the second law and statistical reasoning. This is something that physicists from Richard Feynman to Roger Penrose clearly say.
In episode 8 and 9 of this series we'll discuss this all in depth. We just released episode 6 so it won't be too long for those.
And lastly, it is not an appeal to human values to say that a universe filled with atoms, molecules, planets, stars, galaxies, and life is a more interesting universe filled with more wisdom (chemistry, biology, astronomy, etc. in addition to physics) than a universe with only black holes (and just fundamental physics).
But, um, what is your definition of God? It certainly does not seem to me like an argument supporting an existence of the entity described under that label in the Bible, or in the Quran...
We're not trying to prove any book or religion. Just a God who intelligently fine tuned the laws of nature. We're going to have a separate miniseries about the idea of God that emerges from the fine tuning argument in physics.
Adding a 'God' to the system doesn't remove the 'problem' of fine-tuning, it makes it much much worse - now you need to explain a super-powerful extra-universal creator! That's way way more complicated that suspiciously narrow viable ranges for a handful of free parameters
Not really. Unless you posit literally infinite regress of causes or contingencies, there must exist at least and at most one non-contingent, eternal, immanent, et cetera Thing that can cause everything else that follows: this we understand to be God.
Religion can be very complex with a lot of moving parts that have to be justified, but the metaphysical first cause is not just parsimonious, but really necessary unless you embrace endlessly elaborate and frankly—to me—incomprehensible worldviews.
The concept of cause is itself an abstraction from experience. Trying to make it precise and workable (e.g. in the context of machine learning) is far from trivial. I don’t think it is necessarily a good idea to extrapolate an ill defined notion of cause to the beginning of the universe without dealing with the philosophical intricacies of the concept first.
This can be discussed from a bunch of different angles. Classical mechanics essentially boils down to equations such as Hamilton’s, that allow us to calculate the coordinates and momenta of a system at time t given their values at time 0. It has no obviously defined notion of cause and effect. The system just does its thing; reversing its time evolution would still yield a valid solution to the equations. So it’s hard to say what causes what. Quantum mechanics is not too different in unitary evolution (basically the smooth evolution that happens when no one is looking). Cool stuff -irreversibility- happens if you take measurements, but I understand it’s an open problem still. All in all it seems that a meaningful notion of cause has to emerge statistically, rather than being fundamental; pretty much like irreversibility in statistical mechanics emerges. To think about it, this seems reasonable: you wouldn’t say that smoking causes cancer if you had access to only one person that smoked a fixed amount of cigarettes. Either that person gets cancer or he does not and in both cases there is no counterfactual. So ‘smoking causes cancer’ is a statistical statement. There has been quite a lot of debate in statistics around the status of causal statements like this. Until recently it was consensus that statistics can just detect associations. The notion that we can do more than that is fairly recent and still somewhat controversial (you can look into the work of Judea Pearl if you are interested).
"Not really. Unless you posit literally infinite regress of causes or contingencies, there must exist at least and at most one non-contingent, eternal, immanent, et cetera Thing that can cause everything else that follows: this we understand to be God."
Why not posit that it's the laws of physics? The laws of physics are, as far as we know, non-contingent, eternal, and immanent.
The laws themselves are not noncontingent; their action is contingent upon the existence, too, of space time itself. Likewise spacetime is contingent on some laws that define how it behaves.
While one may be able to posit that, the laws of physics were discovered to have fine tuned constants which indicates that they have an intelligent cause. We take up this possibility and develop this argument more fully in earlier episodes (3-5).
Do we have any reason to believe there is NOT an infinite regress of causes? Whenever we look closely enough, so far, we have found things to be more complicated. We have also not proven anything to be fundamental (which proof may be impossible, like proving God's non-existence).
>Do we have any reason to believe there is NOT an infinite regress of causes?
Short answer: No.
Long answer: The notion that the arguments for God require denying an infinite regress of causes is a misconception. The traditional First Cause argument is actually based on the notion that there cannot be an infinite regress of *essentially ordered* movers. This is because an essentially ordered sequence is defined partly by the fact that, if the first mover in the sequence is removed, all of the objects after it will cease to move. Therefore, if any object in the sequence is moving, then the sequence must have a first mover.
However, this entire line of thought is irrelevant in the modern world, because the concept of an essentially ordered sequence is based on a debunked Aristotelian notion of how motion works. In reality, any time you have a chain of moving objects and you remove one sequence in the chain, the remaining objects will still keep moving until an outside force stops them. So the argument for why you can't have an infinite regress of causes is no longer valid.
This is a good point. It seems to me that infinite regress of causes means nothing when it comes to the existence, or not, of God. We can't prove the assertion one way or another. So it comes back to simply whether or not one believes in God without proof.
The path of scientifc knowledge goes from the complex entities and events we observe, and explains them based upon simpler and simpler fundamental particles and laws.
"In the whole history of science from Greek philosophy to modern physics there have been constant attempts to reduce the apparent complexity of natural phenomena to some simple fundamental idea and relationships. This is the underlying principle of all natural philosophy." – Albert Einstein, The Evolution of Physics pg.52
Sure, but once upon a time we discovered the basic building blocks of everything were photons, protons, neutrons, and electrons. Oh, and neutrinos. Oh, and muons. Oh, and anti-matter, and quarks, and gluons, and mesons...
Good question. Depending on who you ask, they are believed to be fundamental. Or perhaps they're vibrations of strings...
Do you think it's reasonable to have an infinite chain of simpler and simpler particles? Though my intuition is that it's not, I cannot prove it. Perhaps it's simpler and simpler turtles all the way down?
We will discuss this more on our miniseries on God.
Additionally, at best it gets you to deism, an absent creator who doesn't touch the universe after it's created; not many religious implications to that kind of creator, frankly
It also doesn't get you to a conscious creator. The "creator" it gets you to could be as dead and mechanical as a rock--or, as I said in an earlier post, the laws of physics.
The argument gets you to a Creator who intelligently fine tuned the constants of nature to result in our complex universe which contains atoms, molecules, stars, planets, life, etc as opposed to consisting of a chaotic sea of fundamental particles (the result without fine tuning).
We address your question in the last couple of minutes of this episode, albeit in a short manner. We’ll elaborate more in the miniseries about God. But you’re right, that we’re not dealing with religion. Only God at this point.
Hi everyone! :) I conducted several interviews with a bunch of Berkeley AI alignment researchers in February. I distilled them into vignettes which include themes such as (perceived) social pressure, status dynamics, and self-censorship. I'm curious how many of you can relate to them! https://www.lesswrong.com/posts/yXLEcd9eixWucKGHg/the-seeker-s-game-vignettes-from-the-bay
15 % on Ukrainian victory (up from 14 % on May 22).
I define Ukrainian victory as either a) Ukrainian government gaining control of the territory it had not controlled before February 24 without losing any similarly important territory and without conceding that it will stop its attempts to join EU or NATO, b) Ukrainian government getting official ok from Russia to join EU or NATO without conceding any territory and without losing de facto control of any territory it had controlled before February 24 of 2022, or c) return to exact prewar status quo ante.
46 % on compromise solution that both sides might plausibly claim as a victory (up from 45 % on May 22).
39 % on Ukrainian defeat (down from 41 % on May 22).
I define Ukrainian defeat as Russia getting what it wants from Ukraine without giving any substantial concessions. Russia wants either a) Ukraine to stop claiming at least some of the territories that were before war claimed by Ukraine but de facto controlled by Russia or its proxies, or b) Russia or its proxies (old or new) to get more Ukrainian territory, de facto recognized by Ukraine in something resembling Minsk ceasefire(s)* or c) some form of guarantee that Ukraine will became neutral, which includes but is not limited to Ukraine not joining NATO. E.g. if Ukraine agrees to stay out of NATO without any other concessions to Russia, but gets mutual defense treaty with Poland and Turkey, that does NOT count as Ukrainian defeat.
Discussion:
Much has happened from previous update, but imho none of it fundamentally changed the situation.
What finally prompted me to slightly change the odds was Biden’s decision to supply cluster munitions to Ukraine (btw. Russia already uses cluster munitions against Ukrainians). And that is for two reasons: a) it signals slightly (not much) deeper US commitment to support Ukraine that I would expect, and b) those might be actually practically fairly important on the battlefield (for a quick explanation see e.g. here: https://www.youtube.com/watch?v=deK98IeTjfY&t=3459s).
*Minsk ceasefire or ceasefires (first agreement did not work, it was amended by second and since then it worked somewhat better) constituted, among other things, de facto recognition by Ukraine that Russia and its proxies will control some territory claimed by Ukraine for some time. In exchange Russia stopped trying to conquer more Ukrainian territory. Until February 24 of 2022, that is.
When the Ukrainian counter-offensive began I was hoping for an early decisive breakout, but it became clear in the first two weeks that wasn't happening.
My next hope was that Ukraine would make broad, slow progress across the whole front, and slowly attrit Russia forces. It's evident that isn't going to happen either. All we've seen are minor tactical gains. Russia has a seemingly endless supply of convicts to man the front lines, a compliant populace, and a military industrial base that is more than capable of replacing lost equipment. All of the talk about an impending Russian collapse has turned out to be false.
The current lines probably aren't going to change much, barring a direct NATO intervention or a 10x increase in Western assistance. I hope I'm wrong, but it's looking like Ukraine is going to have to agree to a negotiated peace on Russia's terms.
I feel like Prigozhin's Midsummer adventure warrants a mention here. How does that not affect your probability of a Russian defeat as a result of a collapse in leadership?
We understand so little about what actually happened (still happening?) that the prudent course of action is to ignore this event for now. No way to incorporate it into a forecast today without having to update it tomorrow.
I understand why that seems relevant, but now it looks like Putin has roughly the same level of control as before that exciting adventure.
Of course the very fact Prigozhin was allowed grow his private army to such an extent is sure sign of the weakness of the regime, but it is not a new information that Russian regime is weak compared to liberal democracies; that is already factored in the prediction (btw. that weakness is imho the main reason why Putin felt the need to externalize its problems by invading Ukraine, but that is a different can of worms).
Isn’t it surprising that Prigozhin did not wake up with novichok on his underpants? I think it’s quite surprising that he is still alive. That suggests either this whole mutiny is some elaborate ruse, or Putin does not have the influence to simply execute him.
I mean, second explanation is rather obviously correct; Putin needs Wagner mercenaries, or more precisely, people who are currently Wagner mercenaries, to cooperate with him both in Ukraine, and also in Africa. These guys are reportedly basically running the Central African Republic and are attempting to do the same with Mali and other African countries. It is unclear how they would react if Prigozhin would be offed. Which doesn't necesarilly mean it won't happen, of course.
Also Prigozhin has some support among Russian public, which counts for something. Putin is a dictator, but he cannot afford to completely disregard public opinion.
For people like me who only occasionally read these, it might be helpful to mention how things have changed over the long term. Maybe a graph showing the three numbers change over time?
I'd love too, but I don't think substack interface supports graphs in the comments (or am I wrong?), and probably aproximatelly no one clicks on the links, so...
Where meditation is sold as this happiness pill - just pop one and all your problems will be solved! And so one of the most common questions on meditation forums is: I’ve been meditating for years and still don’t feel happy.
This myth is harmful because people are sold a dumbed down version of meditation and never stick around with it to see its real benefits.
OK, I don't get this. I understand the maths (obviously, 100,000,000 < 1,000,000,000) and I understand the "repugnant conclusion" reference to the the happiness of populations, but I don't get why this phrasing is funny, or thought-provoking, or whatever it is meant to be...
The math is representing the repugnant conclusion, but also the arithmetic itself it counterintuitive making the equation even devoid of context "repugnant".
(If you still don't think it's funny don't worry, you're not alone)
My favorite protein bar maker is being sued for lying about nutritional values (e.g. a claim that their bars have three times as much sugar as stated). They claim they're telling the truth and the prosecutor must have tested a bad batch or something. Is there a way I can test these bars myself?
(Context: this is an Israeli company that I don't think sells internationally, so the only authorities that would check it would be local ones. Otoh the bars definitely don't taste like they're high in sugar, so I'm inclined to believe the company that they're not).
The problem probably hinges on the exact definition of "sugar". Like: Where exactly is the cutoff in chain length between sugar and non sugar carbohydrate? Do all sugars count (fructose, lactose, maltodextrin...)?
Quantifying glucose in a home lab is quite possible, though probably not practical for you if you have to ask about it. Doing a complete sugar profile would be um... a project. Finding a commercial lab to do it is certainly easier and probably cheaper.
Re exact definition: Doesn't the government (in the US the FDA, in Israel presumably some similar body) have specific definitions for how to measure things for the nutritional information on the back of the box? (to be clear we're going by that, not by some ad making vague claims).
An example of what you could do: Precipitate CuO from acidic CuSO4 solution, weigh the precipitate to quantify. If all sugars do this, how oligosaccharides react, if other contents of the protein bar would be a problem and so on, I would have to look up.
There actually turns out to be a distinction between https://en.wikipedia.org/wiki/Reducing_sugar which do include glucose, and non-reducing sugars like sucrose. "In an alkaline solution, a reducing sugar forms some aldehyde or ketone, which allows it to act as a reducing agent, for example in Benedict's reagent." https://en.wikipedia.org/wiki/Benedict%27s_reagent indeed contains CuSO4 (along with sodium carbonate and citrate), and indeed yields a copper oxide (albeit Cu2O rather than CuO) in the presence of reducing sugars (and probably other reducing agents).
I hate the big “No High Fructose Corn Syrup” on products that list fructose corn syrup as the first ingredient. But it not the high concentrated stuff you see so obviously a healthy choice.
"Sugar" covers a lot more than just "glucose" which is what most people think of when they hear "sugar". Unless you have a list of ingredients, it would be hard to tell if they mean "low in sugar", "no added sugar" (but plenty of naturally high-sugar ingredients), artificial sweeteners or the like.
They're certainly being presented as 'better for you than a candy bar or bar of chocolate' with all the fibre etc. and then I read the nutritional info and go pale.
A patient of mine saw a psychiatrist who had him do genetic testing to get additional information about what meds would help him. I quickly looked up the company that does this, and found some pretty limited research into efficacy by people not with the company. Upshot seemed to be that it helped some with choosing medications for depression, at least, but not a lot and only on certain measures. Anyone know any more about this approach and whether it's currently worth using or mostly hype?
I just read the abstract, which is pretty detailed. They didn't conclude it was useless. In fact, they seem to have sort of bailed, saying studies used different endpoints so no meta-analysis was possible. (That seems odd -- studies measures never match up perfectly in meta-analysis, there are ways to work with the differences statistically to make it possible to compare results.) But they are saying, PRS might prove to be quite useful. If there's an article saying approach is bogus I'd like to know about it. Are you sure this is the one you had in mind?
Anecdotally, a friend of mine did that, after trying a bunch first. Did not seem to make a difference for them, the "most genetically suitable" psych medication still did not work well enough. But maybe better than the dr trying randomly.
If everybody in the world suddently and permanently became more physically attractive (according to the world's current standards), setting aside any correlated benefiits such as improved physical health, would this:
A) ...simply be a positive change long-term, since we can all enjoy looking at more attractive people around us, and enjoy the benefits of being more attractive to others?
B) ...possibly be a positive change short-term, but eventually become a neutral change in the long-term steady state, because everybody's standards and preferences will simply adjust?
C) ...actually be a net-negative change, for some reason?
I think C). If everyone is more attractive, everyone has much more of an extra incentive to cheat on their significant others. It would be harder to stay professional at work with members of the opposite sex. Sexual harassers and rapists would commit even more sexual offenses, due to experiencing even stronger urges.
Want to say C*, but it's probably not actually C. With the caveat that we're literally just talking about a perception change**, have to go with B. Unless you're nailing everybody to the peak of beauty, the differences will still be maximized by society. You won't be more attractive to others, you're still just on the scale.
*(Sex is a distraction at work. If everyone looks more like Tom Cruise then seeing them will more likely remind you of how crazy Tom Cruise is. You don't see the drug use from across the street anymore, you find out when they're in stabbing distance.)
**(If we're talking about physical effects in ability, it's C. Every powerlifter in the world would suddenly lose 300 pounds of liftable weight.)
I think it must be B, since attractiveness will turn out to be relative. It could be worse, but then again, it could be better, to paraphrase Gandalf in The Hobbit. Even if you could quantify attractiveness and magically make everyone equally attractive according to your measure, people would find differences between them, no matter how small, and then a slight flaw would be amazingly ugly, and the top people would be amazingly beautiful.
B. People will get used to the higher level of physical attractiveness, and those born and growing up where everyone is not-ugly will take it as a matter of course.
So then attractiveness will shift to be even higher; now you have to be perfectly symmetrical in features and have the toned body and flawless skin even without makeup etc. to be considered beautiful.
I feel like it's A, but we must be much more attractive than past generations when everyone was malnourished and got lots of diseases, and I would have guessed that people in the past were just as horny and interested in physical beauty as we are.
On the other hand, obesity became considerably more common. I wonder if there were ever as many people seriously disfigured by diseases or malnourishment as there are today by obesity.
That's a good point! ( My personal tastes happen to run to heavier women, so I personally don't react to extra weight as detracting from beauty, but I recognize that this is atypical. )
>If everybody in the world suddently and permanently became more physically attractive (according to the world's current standards),
This question is based on a false assumption that there is such a standard. I'd argue that there is no universal standard that hold for the entire world, but rather many standards that are held by different cultures (and different groups and subcultures within these cultures, and then there are also personal preferences that can differ even more).
These standards overlap to some extent, but there are significant differences, and if you remake everyone according to just one of these standards, it will be a downgrade in the eyes of many who did not hold that standard. Also, physical attractiveness is usually only a part of attractiveness in general, and I really doubt most people focus only on physicall attractiveness rather than on general impression that combines it with other factors (behaviour, clothing style, etc).
And that's the case even now, after all the globalisation which did push Western beauty standards to other cultures worldwide, with noticable effect. The gaps used to be much bigger in the past, especially if you are willing to look some centuries back. If you look up that was considered to be a standard of attractiveness in say Russia, Japan, France and India in 12th century, you will find even less in common.
So I'm not sure if that's C (because it's hard to quantify if there will be more winners or losers), or B (because whatever effect it takes will be temporary, people will invent new standards pretty soon).
Of course, if there are people who are ugly by all standards, it would be a good thing to make them more attractive by any standard. I'm not really sure to what extent that would be possible though, and it would be a rather small effect overall, so I don't think that would change the conclusion.
C: your scenario says: health et al. stays same. But different looks - esp. a broader spectrum of less and more pretty - did signal "better"/different health/genes. Thus: loss of valuable information. (As studies seem to show, the closer to the "average" a portrait-pic was, the more "attractive" it was judged. )
I think most people do not include "the last person willing to have sex with them" in their testaments because that would set up incentives resulting in a rather undignified death. :-)
IIRC (and if my gossip sources are correct) of his four "girlfriends" that accompanied him in public/lived with him/slept with him, only one of them would have sex with him. I don't know if she got a bonus stipend compared to the other three, but if she didn't then not including her in the will makes me dislike the man in the same way I dislike lousy tippers.
I assume he paid up-front (his bunnies had it materially good, I heard. Mostly hating each other). If "my last" should not be my wife - I doubt, I would or should change my will for her.
I believe A. Generally when people to a more scenic sea shore or forest, they still consider it scenic years later and not this is boring and everyplace else sucks.
Seems like this would reduce the difference in attractiveness between real people and fictional representations optimized for attractiveness, which seems like a good thing to me. (Because I expect it would tend to make people happier with their actual partners, and also make it harder for advertisers to manipulate us.)
If you're asking purely about the direct hedonic effect of seeing more total attractiveness in daily life and whether the hedonic treadmill would eventually negate it, then I'm not sure.
You made me change my mind from (B) to (A). We already spend a lot of time looking at the most beautiful people that exist, and presumably this change would narrow the gap between those people on the one hand and ourselves and our partners on the other, which seems like it should increase happiness.
You know what, I think the Barbie movie is a great test for this! I haven't seen it and wild horses couldn't drag me to it, but the basic plot line is this example: everyone is Barbie-pretty (be they blonde Barbie, BIPOC Barbie, Ken, etc.) and they live in the Barbie-world of sun and fun and happy days and contented nights.
The story seems to be that the less than perfect dolls get exiled to the real world, but of course - what makes them less than perfect in the first place? Why are they not satisifed with their utopia?
Would we be happy with Barbie World forever, or would there always be the less than perfect among us?
About the Barbie movie. It was totally below my radar but I just read a lengthy sympathetic article about Greta Gerwig the director and the hIstory of Barbie the doll. It seems like there is more to the whole thing than I ever would have guessed.
I doubt Mrs Gunflint and I will make a dinner and a movie thing out of it - that kinda came to halt with the rise of the MCU - but I’m interested enough to watch for it on cable.
I think it’d be A. The steady state might settle lower than the initial shock, but overall it would be better.
As an example consider the sense of involuntary repulsion an outlier person can cause in you. If these people were all within the Overton window of beauty, you’d never have that experience.
Although it’s possible perception of beauty would shift towards something else, like outfits or some more nuanced physical attribute.
The themes might be of interest to some of you - if so please feel free to read, share and subscribe. As always, I would love to hear your thoughts and any differing views.
I continue hearing reports that 5-MeO-DMT is the most powerful psychedelic and I’m curious to hear more takes from a rationalist perspective. Please anonymously share anecdotes and resources.
Would you also agree that "more toxic" does not make sense, since water can be just as toxic as botulinus at the correct dosage?
I agree though that for most pharmaceuticals, the required dosage should probably not be the main concern compared to other desirable qualities such as selectivity or side effects.
If someone asked "I continue hearing reports that botulinum is the most powerful toxin and I’m curious to hear more takes from a rationalist perspective", I think the correct response is that "more powerful" doesn't make sense since dosage decides how powerful the effect is.
If the hypothetical poster is interested in LD50 values, or if OP was interested in dosages for psychedelics (LSD has way smaller doses than DMT by the way) I would presume that they look that up on Wikipedia instead of asking here. But that doesn't seem to be what OP is asking for, instead it seems to me that OP is working on some misinformed pop understanding of how psychedelics work.
I have a Substack so obscure that only one of its posts has ever generated any subscriptions. Now it occurred to me I could use the allowed once-per-year unsubtle-plug-outside-classifieds to explore that post's appeal to a wider audience than the Classifieds Thread reached.
Its point is that increased political polarization and dysfunction could simply have to do with the moment in time when the progressive side achieved victory on all its reasonable demands (think of gay marriage) and therefore moved on to unreasonable ones (think of certain trans demands).
Unrelated to that post, I also had another thought related to polarization. Are there culture-war debate formats on TV like, say, the Soho Forum Debates? If not, why not? They could be a win-win: on ratings, because people like culture-wars battles; and on political literacy, because viewers would, for once, get exposure to both sides of a contested issue.
Perhaps the relative nature of these formats would be unsatisfactory for a TV audience? (Your side will never get full support and might even win by merely going from, say, 30-70 to 40-60 behind.) Or do liberal media realize that on the identity-politics issues dearest to them they might lose once an audience is properly exposed to both sides of the argument? In the case of the US, which has a conservative TV station (Fox News), the latter cannot be a complete answer.
There's a state involvement as a ratchet aspect to this question. Over time the number of people who question if the Leviathan really is the best option to do something has grown in concert with the growth of the Leviathan. So what was progressive in the late 1800s (say the temperance movement) that relied a lot on social pressure to make a change has shifted to a demand that the pressure come directly from government itself, who cares what "society" wants. As the Leviathan gets bigger, it puts the resistors in the position of also having to rely on it to stop/retard change, which of course, just feeds the Leviathan making it easier to use it to beat up on the smaller group. Or something like that.
I've noticed this dynamic as well. My diagnosis is that the concept of progressivism as a movement is too abstract and unmoored from concrete concerns that it will inevitably go off the rails. Progress for progress's sake is unsustainable. There's no built in limiting factor that constrains what is considered a proper target of optimization. You eventually run out of sane dimensions to optimize and the movement is left to latch onto more and more unreasonable targets.
But the concept of progress for its own sake is antithetical to questioning the reasonableness of the current direction; self-reflection is verboten. Social media greases the wheels of this sort of degenerate progress because it incentivizes jumping onto these movements uncritically. The Good and the Moral will always have some intrinsic value for self-promotion. Progressivism as a concept situates itself as the arbiters of the Good and the Moral, which makes its excesses so hard to push back against. One of the biggest challenges this century is giving people the conceptual tools to say no to people who claim the mantle of the Good.
That‘s some good analysis on an obvious question that I chose to sidestep in the post, namely „why would progressives still want to progress further once they have won“. I agree particularly on the importance of the Good, the Moral, and self-promotion. A key concept here, I think, is prestige status, to be distinguished from dominance status.
In the post, the situation gets complicated by the fact that, after having moved on from the topic of trans demands, I end up not actually insisting that the progressive side has won (as I now put it in the mini-summary) „on all its reasonable demands“. There may still be implicit discrimination —- yet I argue (in the post) that a shift from explicit to implicit can already explain increased polarization.
Cynically, you can probably get all the heat (which generates the ratings and money) from a one-sided rant or bringing a strawman on to mock or talk over. Optimizing for political literacy brings the network no additional benefit, and might be a harm if they diversify their audience, reducing ad impact.
Theres a debate program somehow affiliated with npr called intelligence squared that featured "hot button" issues. Iirc one of them was "is Islam a religion of peace?" Of late they seem to do more moderated discussions or less sensitive issues, though.
I believe there is a YouTube channel that does this, but it was more along the lines of putting ordinary people on opposite ends of an issue in a room and having them discuss. I think topics included flat earth and trans rights.
I thought a boost to political literacy could be a nice side effect from a format that would be good for ratings (the way I put it in the original comment wasn‘t ideal). But I guess I understand now what the problem is, thanks to you and to Zan Tafakari, who says in a comment below that Piers Morgan indeed brings opponents on „more with an intent to mock“. Even if I’m right that contests are good for ratings, why risk a loss for your side in a genuine one, if you can instead just mock weak opponents or strawmen.
What makes gay marriage reasonable? For basically all historical cultures across the millennia the idea of two men getting married would be insane, even for those cultures that allowed the open practice of homosexuality, usually as pederasty.
The biggest difference I see is that trans demands are heavy on other people (use of preferred pronouns, access to women's changing rooms and prisons, health insurance coverage for cosmetic procedures) while gay marriage does not impact others much unless gay people deliberately go out of their way for it (wedding cake bakers)
These demands came in conflict with other influential leftist groups like british feminists resulting in some pushback.
I would say that „heavy on other people“ versus „does not impact others much“ has something to do with reasonableness (also, it‘s not only or mainly feminists that push back against trans demands).
But admittedly I never define in the post (because I can’t) what I actually mean by reasonableness. Historical standards don‘t cut it for me, but that‘s only implied, when I write that we should thank progressives for overcoming slavery and inherited absolute power.
Your posts look interesting - I've subscribed! I also write a fairly new and esoteric substack which you might enjoy: https://zantafakari.substack.com/
I mean often times the left leaning or right leaning news outlets will bring on opposing people - but I wouldn't say it's a "debate".
Sometimes there are panel shows (like Good Morning in the UK) where opposed view people are brought on - but the debate is not exactly moving the dialogue forward.
I suppose it's not entirely in media's interest to have a reasoned debate if something more polarizing keeps the eyeballs watching longer. So I think it's probably "unsatisfactory" as you say. I doubt that media are concerned that people would change their minds but have no evidence to back this up. Perhaps its an intrinsic idea I have that it takes A LOT to change an opinion, particularly a political one, and TV debates aren't likely to cut it
Thank you! And I'm looking forward to exploring your Substack (which is much less obscure than mine). Starting with post number 12, for the philosophy and evolution of language. As for the debates in my original comment, they could still be bad-tempered (like the US presidential debates were last time). And if they are good for ratings, the media don't have to have a higher purpose such as promoting political literacy. But you say that already "often times the left leaning or right leaning news outlets will bring on opposing people": I have to admit I didn't even know that they do at least that (I don't watch TV these days).
That‘s indeed a good counterargument, a worldwide phenomenon is exactly what the social-media thesis (which I frame as the competing thesis in the post) would predict. I thought there was a specific Western phenomenon, though, exemplified by Trump, Brexit, Le Pen, etc.
Anyone have recommendations (ideally ones that helped them personally) for books/things to read about how to come to terms with things one can't change about oneself? This is basically stoicism, but I'm curious if there's anything that people have found highly effective in actually creating these mindset shifts.
I have a friend dealing with something (going to change it entirely for anonymity), let's say they aren't an amazing basketball player and wish they were. What are techniques people have found effective for coming to peace with qualities about themselves they wish were better? (Since obviously the basketball example is made up, I'm _not_ looking for responses of the form "convince yourself that you never really wanted to be good at basketball anyways").
If the thing your friend can't change is really something of about the size and specificity of playing basketball, there's something a bit funky going on. There are probably dozens of things of about that size he's no good at, right? -- like, say, water skiing, chess, portrait painting, decorating one's home attractively, tennis, etc etc. Why is he so distressed about lack of awesome talent for this particular skill? He must have pinned other hopes on it. I know someone is convinced he can only find a woman if he becomes a millionaire. He's an extremely smart guy getting a degree in a respected profession that is guaranteed to make him upper middle class. His is kind and funny when he's comfortable with someone. His looks are ok. I'd say they're average, but could be better than average is he lost weight and/or worked out some. But he's convinced he has to be millionaire to attract a woman's interest. Is it something like that with your friend -- he thinks being an awesome basketball player is the only thing that might make him worth knowing? If so, that's just not true. If the story is that he's devoted years to becoming excellent at basketball, and is forced to admit he'll never get better then Above Average, that's a different story, and needs to be examined a different way. Now if the skill is something most people would consider of more significance than basketball -- say, being a good conversational partner, or a good writer -- that's still a different story. Is basketball really a good stand in for whatever he's bad at?
Oh yeah, I get that. I meant is it something on the order of basketball, or it he using a basketball as a stand-in for something most people would consider it much more important to be good at.
If it's something about the person's physical appearance it is possible they have a condition called Body Dysmorphic Disorder, which is generally considered a form of OCD.
I'd suggest checking out Core Transformation (Connierae and Tamara Andreas).
IDK about ACT, but I'm concerned it may not engage directly with the underlying structure generating the suffering that stems from "I must be X, but I'm not".
There's *usually* an implicit continuation to that sentence, ", therefore A B and C core needs will not be met/drives will not be fulfilled and expressed, and that's excruciating.", which IMO is a completely coherent stance to have *if that inference and the chain behind IT is valid*.
But it's almost never valid.
And accepting that "I'm not X" can and often is conflated with accepting a bunch of dubious (to say the least!, empirically speaking) implicit inferences that are a lot more excruciating; which is why there's so much resistance to accepting that reality in the first place!
A) Clarifying the 'outcome chain', and B) grounding that in a core state (for this, please read the book/refer to more detailed descriptions) leads naturally to the unwinding of the suffering-causing resistance to reality, and crucially, *to a re-orientation to the system's actual desires and goals and needs and values*. The surface-level thing is not so much 'accepted' in some resigned way as the undergirding delusory meaning structures are dispelled and rendered irrelevant in the clear light of direct insight and pure desire.
(In the relatively rare cases where there is in fact something to grieve, well, this allows for that to happen organically, along with the subsequent re-orientation that the denial was blocking.)
Good luck to your friend. May this situation resolve for him, and may he be fulfilled and happy!
Almost everything genuinely effective is going to need at least some details, so I expect you'll struggle to find a good response here. Like, advice with respect to sexual orientation is going to look very different from advice with respect to disability, is going to look very different from advice with respect to talent or g-factor.
The only genuinely useful and *universal* trick I can think of is meditation, starting in the mindfulness/samatha end of things and progressing to insight meditation after a few months or years when you feel ready for it. I'm particularly fond of "The Mind Illuminated," but there's no shortage of eager introductions really.
Yeah I figured it was a long shot, although Scott's answer looks promising. I can't really share more though. Unfortunately I think they're averse to a lot of meditation since most meditation practitioners are also selling philosophy and possibly even religion alongside it.
Yes I was also going to say Acceptance Commitment Therapy. I've actually written about ACT and it's relationship to Expanding Awareness (a branch of Alexander Technique) - which might also be helpful.
ACT is in many ways similar to CBT (in fact I think it is described as a third wave CBT), except that the emphasis is more on changing one's relationship to one's thoughts instead of changing the content of one's thoughts. I think ACT is meant to be for all the things CBT is for. So if a typical CBT exercise is to write down reasons to challenge a negative thought, ACT focuses more on seeing thoughts as mental events that occur but don't control you (you learn not to be "fused" to them). That is the Acceptance part. The Commitment part is about clarifying one's values and acting on them (the behavioral part). There is a self help book called Get Out of Your Mind and Into Your Life by Steven Hayes, who developed ACT. I thought it was pretty good.
Are medical professionals in training subject to hazing and abuse, and does this lead to worse patient outcomes in hospitals?
This claim of medical interns being abused by requiring them to work extreme hours and crazy shifts is pretty common. Based on a few conversations I've had with medical professionals, it appears to be accurate.
One would expect that medical professionals would be most likely to be familiarzied with literature on the importance of being well rested in order to achieve optimal performance. And yet medical interns are claimed to often be denied this "luxury".
If so many people know about this toxic work environment, and it's bad for both patients and workers, why does it persist and what would it take for the system to change?
Medical residents are worked hard for a few reasons:
-- Hospitals pay a lot of money to train them and want to get useful work out of them.
-- You learn medicine faster/better the more you practice, and a lot of doctors believe you can't really learn medicine in a reasonable amount of time working forty hour weeks.
-- "Handoffs" - ie when a doctor goes home and transfers their patients to another - are notorious for going badly. The leaving doctor forgets to convey important information and then the new doctor does something wrong and causes a catastrophe. Hospitals have tried hard to formalize/ritualize these and make them ironclad, but the best defense is still to handoff patients as little as possible, ie make doctors and nurses work very long shifts.
-- Yeah, I agree there's a sense of "I did this in my day and let's not coddle the new generation by giving them a pass".
The ACGME, the relevant body, has passed various rules - the most important is that interns can't work more than 80 hours/week. These are both overenforced (in the sense that everyone constantly has to sign a lot of paperwork saying they definitely didn't work more than 80 hours/week, waste time listening to the Don't Work More Than 80 Hours/Week Campaign, and so on) and underenforced (in the sense that in real life, hospitals can just assign you X number of patients and say this is the amount a doctor should be able to treat in 80 hours, and then it takes you 120 hours to treat them, and if you complain they'll just say 'git gud')
I assume the “handoff” studies account for the obvious confounded that “more handoffs” usually means “more time in the hospital” ie probably a more serious condition?
The 80 hours/week thing is pretty confusing to me, coming from the IT world. Here we have to log our hours every day, and if we hit above 80 the system won't even let us put in more until our manager gives approval and allows overtime. Plus if we lie and just don't put in our hours that's illegal (in the contracting world anyway. probably it's illegal in normal private sector work as well).
How does it work in hospitals? If you hit 80 hours, but you still have patients to take care of, do you just keep working and then put in your hours and the hospital shrugs and says "ok we'll pay you overtime i guess"? That seems like it would outright violate the ACGME rule. Or are hours just not tracked at all?
"Hours"? I'm guessing doctors, even as residents, are not paid by the hour. Fixed salary, and if you don't complete the work assigned they will eventually convert that to zero work assigned and zero salary.
Ah, yeah that would make sense. I guess in that case the ACGME just made up these rules without having any actual way to enforce or check them. No mandatory time keeping or anything.
The "handoffs" issue seems like it would be equally solved by having people work longer shifts but fewer total shifts, such that the total hours/week was the same?
The handoffs thing seems like a fake reason to me (for example patients are handed off anyway when they go from the ER to the ward), but so does the whole “you can’t learn medicine without working 80 h a week”. I mean, sure, up to a certain point the more you work, the more you learn, but there’s nothing specific about medicine that makes it more true than in other fields.
As far as the hand offs are concerned, you could just as well argue that they are useful because the fresh doctor might come up with something the last one forgot.
The handoffs thing is absolutely fake. Pretty much everyone else who has to deal with continuous critical workload, e.g. the military, or in my field space flight operations, knows how to do handoffs properly. None of them demand 16-hour shifts or 80-hour weeks in anything but the most serious emergencies.
But if you *want* to make your residents work 24 hours straight or whatever, because that's how it was done in your day, these kids today, no respect, etc, etc, then sure, say something about "handoffs" and just make sure you're never officially seen studying proper shift handoff technique.
"I mean, sure, up to a certain point the more you work, the more you learn, but there’s nothing specific about medicine that makes it more true than in other fields." Also, doesn't sleep deprivation impede memory and learning?
Yes. It is also funny that the profession where supposedly people have to work such long hours practicing because it's so complicated to learn is the same profession where "see one, do one, teach one" is an established concept.
Handoffs have been pretty extensively studied, long shifts (eg. 12 hours instead of 8) improve patient outcomes overall, but Dweomite's point is important too, that long shifts *should* mean less shifts
A lot of the problem boil down to a "doctor shortage", but a lot of that scarcity is artificial, both in the forms of restrictive quotas on number of doctors trained each year, and in restrictions on what nurse can do and what legally requires an MD. Keeping doctors scarce means higher wages for them, standard guild behaviour.
I think another major factor is that the primary bottleneck creating the doctor shortage is the cap on the number of residency slots, combined with the residency match system. There aren't enough residency slots to go around for all the med school graduates, and the places that hire residents coordinate their hiring so that each prospective resident has either zero or one job offer available to them.
So as an aspiring doctor, once you've been matched to a residency slot, your choices are 1) do the work they want you to do under their terms and conditions, or 2) give up on being a doctor and find some other way to pay off your med school loans. This gives you very little negotiating power, so the employer has very wide latitude to dictate the terms of employment, and you have very little recourse if mid-residency you don't like your working conditions. You have a little leverage during the match process if you look like a desirable candidate to multiple hospitals (since then your preferences between positions make more of a difference in the match process), but I think you're stuck with a residency once you're matched.
Without the binding constraint on the number of residency slots, hospitals would need to compete more to get med students to choose to do their residencies with them, and developing a reputation for having vaguely sane schedules would be one possible axis of competition. And if there are also changes to the match system to make it less collusive (e.g. allowing job offers outside the match process for senior residents looking to transfer), that would further improve the negotiating position of residents.
Would it be a case of "paying your dues" given that understaffing is common in the training hospitals? Thus someone needs to cover the shifts etc, and that in turn falls to the new staff?
And for change, I do hear a lot of chatter on the matter, but I believe the bottleneck at the training level, where there is artificial restriction on the education needed means there is a perennial understaffing
Does anyone have good resources for someone looking to get across the evidence on optimising for healthy pregnancies (both prior to and after conception)? Is there an established rationalist-friendly person or set of posts? Preferably something that says here's what's consensus, here's what isn't, here's how we know X works etc. There's obviously lots on Doctor Google but I don't trust myself to identify who's a credible expert and who is well aware of e.g. publication bias and replication crisis, and has already taken these into account. Thanks in advance.
I'm sure there are good resources, but my impression from having seen a lot of people's take on being pregnant is that most are pretty resistant to advice. People you would think would be excited to see some actual data, and hear some actual new and well-supportedf ideas, often start making clear early on in their exchange with you that they have already made up their mind on what pregnancy and birth approach is best for them: They want a home birth. Or they want a super-conventional hospital birth, because they are convinced that hospitals give you lots of drugs or just knock you out so that's they way to get the best pain relief. Or they've found a nice book about pregnancy that made them feel better about the whole thing, and now they have bonded with it and do not want the relationship interfered with. They do not want to hear anything from you that undercuts their certainty about what ever take on the situation they have bonded with.
I think pregnancy and the prospect of parenthood is so stressful that people kind of regress to clinging tightly to ideas that make them feel less anxious, and do not want more input. I don't know what to do about that, except to be persistent and kind.
In my experience, of the masses of pregnancy and parenting advice available, not 10% will apply to any given combination of parent and child. Everyone thinks they've got a data-driven approach, but biology and behavior are idiosyncratic and highly variable in every aspect; edge cases comprise a long and meaty tail.
An expectant parent may as well plan the most appealing course, while staying informed and flexible enough to quickly change their approach as circumstances merit. Anyone who eagerly promotes universal advice is best kept at arm's length.
Yes, agree that anyone who eagerly promotes universal advice is best kept at arms length. You got any special reason to think I'm an eager promoter of my favorite pregnancy ideas, which I obnoxiously assume are universal truths? Cuz if you don't, consider that a person who uses somebody's post to build and attack an ugly little straw man is also best kept at arm's length.
Listen, when I write a diatribe you will know it: I will openly angry and critical. Somewhere on this thread I wrote a diatribe about GPT4, which begins "I hate GPT4." Here I am just saying people seem much more unwilling to listen to info and ideas than they are about other matters, and it's kind of striking and surprising.. At the end I say that's probably due to the stress of the immense changes in their lives. It might help you to know that I am a psychotherapist, and many of the people this comes up with are psychotherapy patients. The most recent time I got a reminder of the difficulty making recommendations to the pregnant was about a year ago -- so as you can see, this is not a topic I'm spewing advice about 24/7. The situation a year ago was the my patient had a long history of taking terrible care of herself. In her early 20's she had ignored symptoms of appendicitis and continued to be active until her appendix burst. She then became dangerously ill, was hospitalized, and had a bunch of her bowel removed. Ten years later, pregnant, she was again not taking decent care of herself. It wasn't as bad as the appendicitis situation, but really was worrisome. Other times when I have asked whether I can make a suggestion it has been with close friends, people with whom there has been a lot of discussion of private matters on both sides -- how to handle relationship problems, how big a bid to make for a house, etc. So the "people" I am talking about are people who have sometimes asked me for advice, and have been willing to listen to ideas even if they haven't asked for them, and have given me advice and ideas. What I am saying is that when it comes to pregnancy they seem different and it's kind of startling.
Seconding this: Emily Oster is head and shoulders above the other baby books my partner and I have read (we've got one toddler and one on the way) due to things like assessing the quality of studies when citing them and being clear about the level of certainty and absolute risk involved.
Pretty sure Scott has a very long post summarizing the research on this topic. Every thing from air filters to supplements. It was a while ago but I think it was one of the “way more than you needed to know” themed posts.
I recently read Stephen Omohundro’s oft-referenced papers, The Basic AI Drives https://citeseerx.ist.psu.edu/doc/10.1.1.393.8356 and The Nature of Self Improving Artificial Intelligence https://selfawaresystems.files.wordpress.com/2008/01/nature_of_self_improving_ai.pdf, and don’t understand why they are so influential as they are 100% argument by assertion. The big assertion is that AIs will necessarily, through self-improvement, become perfectly Rational and will therefore create utility functions with which they will operate entirely in accordance. Why will they become perfectly Rational? Well, because. As far as I can tell, almost every AI x-risk argument follows from that moverless-mover assertion. Why believe it?
A one sentence summary of those papers could be: “In the beginning was the Word, and the Word was Rationality.”
If you think I am wrong and Omohundro does explain in a non-circular way why the AIs will become Rational, will you explain it to me?
This feels kind of like arguing that dictatorships can't happen, because a dictator is someone who tries to maximize their political power and politicians are not perfectly rational power-maximizers. They aren't, but you don't have to be perfectly rational to decide that being a dictator is a good strategy, you just need to be capable enough to seize power and amoral enough that you don't care about democracy. So people who theorize about government will often say "imagine that a ruthless power-maximizer wanted to gain power in this system - how do we stop him from overthrowing the government?" even though such people don't literally exist.
Similarly, to pursue Omohundro's basic drives, you don't need to be perfectly rational, you just need to be rational enough to pursue a long-term goal, and be amoral enough that you aren't satisfied at regular human levels of power that we know how to deal with. And it can be worthwhile to ask yourself "Suppose this AI was a ruthless goal-maximizer, how do we give it a goal that doesn't drive it to conquer the world?" even if such an AI doesn't exist yet.
Why should we assume an AI has *any* level of rationality? Omohundro does, but I do not understand why he does. He assumes some modicum of rationality which will increase as the AIs self-improve.
Related question: Do you believe GPT-4 is partially rational, in the sense that it is partially interested in avoiding vulnerabilities?
If an AI has *literally no* rationality, it's not capable of doing anything logical, and there's no point in worrying about it. I'm not sure you could even call such an entity an "AI" rather than "a random noise generator." But any AI worth talking about will have at least some ability for rational thinking.
I'm not sure what you mean by "avoiding vulnerabilities," but the example of AutoGPT shows that GPT-4 is partially rational - it can take a goal, come up with an action that pursues that goal, and execute the action (with appropriate wiring). The example of ChaosGPT shows that it can even be prompted with long-term goals like "try to destroy the world." It's just very bad at it, because GPT isn't that smart and taking over the world is hard.
I'm using the definition of "rational" which Omohundro does. Under that definition, rationality isn't merely logical thinking but includes, fundamentally, "avoiding vulnerabilities", e.g., protecting one's own resources of time, space and free energy. Arguments that an advanced AI won't allow itself to be turned off stem from that notion of rationality.
(I can't help but notice a suspicious similarity between Omohundro's use of "rational" and the Rationalists use of the term.)
Yes, perfect rationality doesn't exist for practical purposes. Rationality is effective use of cognitive resources ,and always an approximartion. "Avoiding vulnerabilities" has a cost.
Why should we assume an AI has *any* level of rationality? Omohundro does, but I do not understand why he does. He assumes some modicum of rationality which will increase as the AIs self-improve.
Related question: Do you believe GPT-4 is partially rational, in the sense that it is partially interested in avoiding vulnerabilities?
An AI doesn't have to be a rationalist in some technical sense, such as von Neumann rationality. Confusion between technical and loose senses of rationality is part of the problem.
Heh. I just wrote a coupla paragraphs about why pregnant women do not want to hear any new and interesting data about optimal approaches to pregnancy and birth. Ended by saying "They do not want to hear anything from you that undercuts their certainty about what ever take on the situation they have bonded with. I think pregnancy and the prospect of parenthood is so stressful that people kind of regress to clinging tightly to ideas that make them feel less anxious, and do not want more input. I don't know what to do about that."
Seems like the same explanation applies to people's take on AI and it's possible dangers. I think quite a lot about AI and the future, and find that I'm constantly somewhat anxious about it, and crave to find others who are unsure who AI development will unfold but worry a lot about certain aspects of the process. It is very hard to find a peer group. So many people are wedded to the idea that AI's gonna kill us or else to the view that anyone who thinks AI is gonna kill us is a stoopit asshole.
If you think of them as our descendants and their (current) strategy as making our lives so engaging and fun that birthrates for half the human population have fallen below replacement levels the prospect of, ah, an evolutionary step might hurt less.
Naw, Scott, but thanks for trying. Even when I'm thinking of the AI era as a big change with no doom in it, it still hurts like hell. I'm not old, but I am considerably older than you, and also happened to have parents who were in their 40's when I was born, and that gave me an unusually strong tie to life several generations ago. Then in college the literature I loved was 15th-19th century. And of course I became book-oriented, because there was no online in those days. Once read an interview with an elderly person who said "your world dies before you do." And I thought, no, that won't happen to me, I'll keep up with the changes. And up til now I have. I'm computer fluent, etc. But the shift to an AI-oriented world just looks ugly, bleak and weird to me, like a life lived inside a big bright windowless mall, and I can't really tell how much of that is an accurate perception that something vital is being lost and how much is just being unable to connect with such a different future. There is a lot about the internet era that I'm fairly sure truly is bad for people: Twitter. Facebook. And motherfucking Tinder is just *brutal.* If it goes one step further into hideousness, the poor people on there trying to find fun and love are going to have star ratings and reviews next to their names, in addition to being subject to the old left swipe. I always met my friends and honeys through actual activities. That is *so* different, and I'm pretty sure it's much closer to what we are wired for, and brutalizes far fewer people. And then there are all the shy avoidant young males who are gaming addicts. And hardly anyone reads books any more. My book consumption is way down too, because I read so much online, and I can sense that that's not nourishing in the same way books are. There is much less unspoiled natural beauty in the world than there was when I was in my 20's and traveled around Europe. People sneer at the humanities. I *loved* the college professors I had who'd spent their lives studying Renaissance literature. I don't know whether charging people money to take classes from them when they could be learning some great STEM stuff is defensible, but those people were important to me. I thought there would always be scholars like that. There won't be.
All that is bearable, though. But the further step, to a world where AI is important, is just too big for me. I loathe GPT4. I hate its blandness, it's prissiness, its wiki-summary world view, it's heartless helpfulness, and the vast emptiness inside it which is so easy to sense. I heartily dislike most of the AI honchos whose thoughts I've read or heard interviewed. They seem to me so deep into tech bro land that they are not playing with a full deck, and they're singing "Tomorrow belongs to me."
I'm like the older generation in Childhood's End. I can't even fully wish the AI-bonded generation well, because when I imagine them they seem so different from me that I can neither like them nor hate them.
"I'm like the older generation in Childhood's End." Interesting reference! I do tend to see (the admittedly currently purely hypothetical) ASI as like the Overmind. I tend to be sympathetic to Karellen, but tastes vary.
I'm also not happy with chatGPT, as it stands today, but I suspect a large chunk of that is from the wokescolding behavior that's been layered on top of the LLM itself, probably from politically correct RLHF - which competitors may tune differently.
I sympathize, honest; I'm 55. I do try to combat my innate risk/change aversion by deliberately looking for positives, but I do not mesh well with, ah, the current mainstream.
> The big assertion is that AIs will necessarily, through self-improvement, become perfectly Rational and will therefore create utility functions with which they will operate entirely in accordance.
This is also trivially false.
> If you think I am wrong and Omohundro does explain in a non-circular way why the AIs will become Rational, will you explain it to me?
There's no circularity in the original papers, so I don't know why I'd expect to succeed here.
1) Rational AI's are somewhat special fixed points, in that for many of the ways AI can be irrational, it will try to fix itself.
2) It is possible to make an AI that is self consistently irrational. These AI's will do worse at whatever.
3) If the AI is very irrational, ignore it, it's not able to do much useful or dangerous stuff.
4) Perfect rationality is mathematically simple, most designs of self consistent irrationality are more complex.
5) Perfect rationality is easier to predict, and works as a good "average case" when we are unable to predict how the irrationality may influence things.
You might be right but as a neutral observer I’m not seeing an argument. Writing “this is trivially false” isn’t all that convincing, without an explanation as to why.
Sorry, are you calling me a stupid asshole, calling Hank a stupid asshole, or some third secret thing?
Needless to say I disagree with the first two possibilities. Pointing out that someone's claims about a source text don't stand up to a 5-second examination doesn't make one an asshole, and neither does making claims about a source text which don't stand up to a 5-second examination. (That doesn't even necessarily make one stupid.)
I am not calling anyone a stupid asshole, as should be evident from an attentive reading of what I said. Believe me, if I think you are a stupid asshole you will know I think so. What I am doing is implying that you are coming close to calling Hank a stupid asshole for expressing doubts about your thesis.
Seems to me that people have become rigid, irritable and intolerant around the issue of what future AI will be like and in what ways, if any, it will be dangerous. Want evidence of that? Your tone with Hank is quite irritable. Gist of what you say to him, in a fuming-with-annoyance tone, is good grief buddy, the answer to your objections is right there in the paper, why the hell don't you get it? People are like that about covid, too, of course, but in the case of covid there really is research to back up some ideas and to prove that others were nonsense.
I don't think that's the case with future AI. There are too many unknowns, so while a few ideas are pretty clearly nonsense, lots of the other ones are still in play, in my opinion.
And I still think you are sneaking in AI's being rational, and the having of private goals, inside the Trojan horse of *really smart.*. "Yes, the argument assumes some minimum level of rationality (as would be necessary for any goal-driven system capable of achieving its goals at all). " So when and how does AI become goal driven? Do you thnk GPT4 is goal-driven? If so, what's the evidence? If not, can you give a convincing argument why the having of goals will be an emergent property in systems trained the way GPT was, only with larger data sets, more determined and persistent reinforcement training, etc.? Will preferences in music also be emergent? Will emotions? Will a sense of humor? Hey, how about a taste for fine wines?
"I don't think that's the case with future AI. There are too many unknowns, so while a few ideas are pretty clearly nonsense, lots of the other ones are still in play, in my opinion."
I agree with you. I think the largest single immediate unknown is how hard the hallucination problem is. AFAIK, it is anyone's guess whether this is
a) "easy", with a decent solution already coded somewhere in OpenAI or Google, awaiting just further debugging, testing, and deployment.
b) "possible but hard" with several candidate solutions, all requiring significant enhancements to data capture during brand new training sessions, and associated high multi-million dollar gambles needed.
c) "excruciatingly hard" with LLMs being fundamentally the wrong approach, and something much closer to human learning with much closer coupling to robotics and much less text needed.
( One admission: Yeah, I'd like to see the end game happen soon enough that I live to see it, so I'm "rooting" for (a) - but of course I have zero control on this, so my preference doesn't matter. )
BTW, I have a wild guess about the "emergent" capabilities from scaling. What if it is from the expansion of the _training sets_ stumbling on "expert" text in the "emergent" fields (or particularly "digestible" text in those areas)? Handwaving argument:
If I understand correctly, LLMs are purely feedforward nets. They have frozen weights (at prompting time), and their only version of short term memory is what they have in the current prompt plus (for those with a large enough context area) as much of the current discussion as fits. It makes perfect sense that they would have the limitation that you saw in your experiments, with an inability to plan.
My handwaving guess is that when their training set is expanded to include transcripts of working sessions with e.g. someone describing their train of logical reasoning, the LLM's predict-the-next-token training can then _ape_ the appearance of the reasoning. In other words, the crystallized intelligence of the expert, reflected in the transcript, then turns into _pattern prediction_ in the LLM. The expert's "Type II" thinking turns into the LLM's "Type I" reasoning - though it _looks_ like "Type II".
If something like this is happening, then the LLM still needs to be augmented with true "Type II" reasoning - but this is essentially what classical symbolic AI was all aimed at. (Vigorous handwaving) It should be possible to take the existing body of work in automating "Type II" thinking (forward and backward chaining, theorem proving, Cyc, etc.) and wrap it around LLMs, to get both modes of thought. Does this sound plausible?
“ BTW, I have a wild guess about the "emergent" capabilities from scaling. What if it is from the expansion of the _training sets_ stumbling on "expert" text in the "emergent" fields (or particularly "digestible" text in those areas)?”
You know that’s very plausible! And I have never heard it suggested. I’m sure the tech bros working in AI are much fonder of the “emergent” view than on the “stumbled on some good info” one, because emergence is kind of gloriously spooky and exciting. The thing I’m not sure of is how much AI could benefit from stumbling on some particularly digestible text. Because it doesn’t digest text, really — not in the way we digest books, or with food — extracting the info and looking for ways to connect it with previously-learned info and filing the whole mass away. It learns the word sequences, and connects them with previously-learned sequences and files *that* mass away. I’m not sure it’s possible to directly teach it anything. (And yet, to contradict myself, if I give it complicated instructions it does seem to grasp them and follow them. So within its encounter with me it does extract the actual info in my instructions, and use the info (not the word sequences) to guide its behavior. So I dunno.)
About its hallucinating. I have a theory about that. I once inadvertently caused it to hallucinate by presenting it with a multiple choice problem with no right answer. Here’s the problem:
Here are 4 generalizations:
-All dogs are fat.
-No pets are thin.
-No thin dogs are pets.
-All thin pets are cats.
Which of the following, if it existed, would contradict all 4 generalizations? A fat cat, a thin dog, a fat pet or a thin pet?
The right answer is, I believe, a thin pet dog. But when I typed the question into GPT I left out “pet” and just typed “thin dog”, and the result of my mistake was that none of the items in the list were the right answer. None of them contradicted all 4 generalizations. So GPT picked one of the 4 answer possibilities and wrote complete illogical nonsense about how that one contradicted all 4 generalizations. That was the hallucination. (Then later I gave it the problem with the correct answer set, and it gave the correct answer and valid explanations for why that answer contradicted all 4 statements.).
But about the episode of hallucinating: It’s like GPT doesn’t have enough “sense of self” to say, “wtf? none of these answers is right.” You can think of its response as eagerness to please, or a failure to trust its own judgment after I have told it that one of the 4 choices will contradict all 4 generalizations. So I was thinking about this and the phrase “As-of Disorder” came to mind. Am I remembering right that you work or worked in psychiatry or psychology? If so, I expect that phrase is familiar. I haven’t heard ir used in a long time, but 25 or 30 years ago it was very stylish to diagnose people with As-if Disorder. And it seems to me that is a good term for the structure of GPT4. It has no opinions and preferences — it learns what to *say,* how to *sound*. Was thinking that GPT rambling on nonsensically was sort of like the moment when the therapist makes an awful discovery about the patient: They do not really grasp the idea of a dialogue, where you tell them what you think and feel and they do the same thing, and an exchange of thoughts and feelings develops. Their conception of the situation is that they are to figure out who you want them to be, and then act as is they are that person. And hallucinating is something they do when they, as usual, have no ideas of their owh, and also can’t see a convincing way to appear to be who you want them to be. So they just spout nonsense that sort of sounds coherent and hope that is good enough to get them through the awkward situation.
Anyhow, if that theory of hallucinations is right, I’d say it’s a very. big task to free GPT of that tendency. Seems like it would have to have a whole different structure.
About your last paragraph: I know very little about Type II thinking, and the attempts to train AI to do it. Can you point me to a good source of info?
> good grief buddy, the answer to your objections is right there in the paper, why the hell don't you get it?
I'm not asking him to "get it", I'm telling readers that they should spend 5 seconds checking the papers to confirm that the claims being made about it are, in fact, trivially false. I agree that I was irritated when I wrote that; I should have omitted the last sentence. Major epistemic sins don't justify minor epistemic sins in return.
> So when and how does AI become goal driven?
Who knows?
> Do you thnk GPT4 is goal-driven?
Not in the sense being discussed.
> If not, can you give a convincing argument why the having of goals will be an emergent property in systems trained the way GPT was, only with larger data sets, more determined and persistent reinforcement training, etc.?
This is not the claim under dispute, nor one that I myself am highly confident in. I _am_ highly confident that people are actively working to make AGI, which explicitly requires goal-driven behavior in that sense (since otherwise it would be unable to perform the desired functions), and have no strong opinions on what technical architectures will lead us there, or how many novel breakthroughs are required. It is obviously possible to build (humans as an existence proof), and a bunch of very smart people are trying very hard to build it. This doesn't guarantee success under any particular timeframe, but we sure do seem to be running out of unsolved capabilities.
"But we sure do seem to be running out of unsolved capabilities." Not if you include having emotions, (and emotion & motivation/goals and are so connected in our species that they are pretty much 2 sides of the same coin), being able to recognize subtle isomorphisms, inventing genuinely novel solutions, performing whatever the hell important process our brains are up to when we sleep and dream, having and learning from role models, & forming attachments. And my guess is that all of these capabilities are required to function at the level that we do.
The first paragraph of "Avoiding vulnerabilities leads to rational economic behavior" (2.3) includes:
"But it isn’t clear a priori why self-improving agents should necessarily follow any particular set of axioms. The argument is more compelling if we can identify explicit negative consequences for a system if it fails to follow the axioms." The rest of the section then explains why avoiding vulnerabilities leads to rational economic behavior. The argument is structurally similar to https://en.wikipedia.org/wiki/Von_Neumann%E2%80%93Morgenstern_utility_theorem, which is hardly an "argument by assertion".
>The argument is more compelling if we can identify explicit negative consequences for a system if it fails to follow the axioms.
This is assuming rationality. Negative consequences are neither here nor there for an AI that is not rational. "Negative consequences" has no meaning for an AI that is not rational.
Von Neumann-Morgenstern models rational behavior. It assumes rational behavior; it doesn't argue for it. Omohundro argues for it, in AIs.
Negative consequences don't exist for an AI that has no definition of utility at all, but it doesn't follow that every AI is a perfect vNM rationalist...That's a false dichotomy.. You can be rational-ish, and in many ways, you can only be.
I have no idea what structure of argument you're looking for. Yes, the argument assumes some minimum level of rationality (as would be necessary for any goal-driven system capable of achieving its goals at all). This is not circular: the claim is that imperfectly rational agents have incentives to make themselves more rational.
But nobody is really clear whether advanced AI will be goal-driven. Seems like producing one that is what require a paradigm change. Of course we can make it goal-driven by giving it goals. Right now, I give GPT4 goals like "write a limerick." Somebody could give it larger goals, and additionally train it to generate subgoals that are appropriate, and walk itself through them step by step. And of course it is possible that in the future we could decide to give it huge goals, like "cure cancer" or " design and implement a plan for optimizing human contentment." So if you did that AI would indeed have goals and subgoals. But they would still come from us.
You don't get to *assume * that because future AI is going to be really fucking smart it will therefore be rational, in the sense of being able to read a situation and see how compatible it is with one's goals, and then design a plan for maximizing the chance one of one's goals being realized. AI now is a lot smarter than we thought it could be, and it seems to be goal-free. You don't get to sneak in "goals" and "rationality" under the umbrella of *really smart.* Seems entirely possible to me we are about to have another AI winter, as people discover the limits of what can be achieved via the present deep learning approach, even with larger data sets and more training. Could be we end up stuck with something that is 10x as good as predicting the next word, but still can't reason it's way out of a paper bag and has no goal except to sit there and respond to prompts.
Assuming AI is gonna get really smart and therefore will be able t do anything leads to paradox and nonsense, as does the idea that God can do anything: "so can he make a rock so big that even he can't pick it up?"
> Yes, the argument assumes some minimum level of rationality (as would be necessary for any goal-driven system capable of achieving its goals at all).
If by "rational" all Omohundro meant was "logical" I would see no problem here. But that's clearly not what he means. By rational he means a self-interested agent which protects vulnerabilities. He assumes that quality is present a priori.
Fun project. I created a "reverse prompt engineer-er". You can paste in any writing and it will reverse engineer an LLM prompt that could supposedly output. It's free to use. https://home.pickaxeproject.com/prompt-reverse-engineer
Does this just ask a LLM to do the prompt engineering thing? I had assumed from your description it did some sort of search or running the network backwards or something (I don't know enough about LLMs' architecture to know whether that's plausible). The response I got looked more like the sort of thing what would be the prompt used by Prompt Reverse Engineer itself than an actual response. I assume this is a bug. Do you mean it's meant to produce a piece of text which, if you prompted some LLM with it, it would likely produce the output you said?
If you're interested in finding more humanely raised meat products that are independently certified by ASPCA-recommended certifiers, check out the free Find Humane app at https://findhumane.com/
I looked into this relatively recently and found two things at work:
1. The decline under debate is very recent, lined up chronologically with the Brexit process and the Covid recession. Your linked graph stops in 2018; instead, you probably want to look at 2015 (the year before the Brexit referendum happened) through 2022 (most recent full year).
2. There are two different ways of calculating real GDP per capita in an international context, both of which involve converting local currency to dollars and then inflation-adjusting the dollars based on the US's GDP deflator. One uses market exchange rates, while the other uses "Purchasing Power Parity", attempting to optimize the GDP figure as a proxy for standard-of-living by using local prices for equivalent goods and services as the currency conversion factor. For Brexit-related and Covid-related reasons, the relationship between PPP and market exchange rates for Britain have been highlyd unstable in the period in question: exchange rates have been very volatile (ranging from US$1.08 to US$1.40 per £1.00), and tariffs and covid disruption have both radically changed the availability and prices of imported goods.
Looking at either the PPP or market exchange rate numbers, everyone took a big hit in 2020, while Britain appears to have taken a deeper hit than France and the overall OECD average (the two control groups I picked off the top of my head). The big difference is that in market exchange rate terms, the recovery looks proportionate to the decline (i.e. Britain fell more, but also recovered proportionately faster so as to bounce back to approximately 2019 levels in 2022 the same as France and OECD). But in PPP terms, the UK has recovered at the same rate as France and OECD and thus appears to have permanently (so far) lost ground in standard of living relative to other countries. UK was also growing more slowly in PPP terms between 2015 and 2019 than France, but about the same as the OECD average.
"Unprecedented economic decline" and "path to a Third World country" are ridiculous exaggerations (as the use of the phrase "Third World" should indicate, since that term no longer has any real meaning). But when the best you can say is "at least we're not as bad as Japan's lost decade", that's not really a good sign. As others note, the past few years have been harsher on the UK than the period in this chart. Even in this chart, the only countries as bad as the UK post-financial-crisis are Spain, France, and Italy.
I don't think that "being on the path to a third world country" applies to the economy as a whole. But economy isn't everything, and the decline of the NHS is sudden and shocking. An example where they no longer achieved first-world performance is the ambulance response time:
The problem is specific to UK, and is probably due to political decisions like Brexit of the last 10-20 years. That said, the extreme failures were only temporarily, and the system has stabilized a bit, though on a level that would have been unacceptable before 2020.
The belief in Britain's economic decline arises from an understanding of its long-standing, extractive, rentier economy. Post-WWII industrial decline and the subsequent financialisation of the economy have led to a dearth of technical jobs and over-reliance on property ownership, fuelled by planning restrictions that artificially limit supply (and in recent years, mass migration).
Most people's savings are locked in housing, a non-productive sector. This situation, exacerbated by significant migration and pension funds also being invested in this bubble, leads to chronic underinvestment elsewhere. The government is reluctant to rectify this, as any action deflating housing prices would upset voters.
This results in a precarious cycle. Profits divert to housing and legacy industries, inflating property values and concentrating spending on rent, while investment in productive sectors dwindles. The economy hinges on importing cheap labour to sustain GDP and attract foreign capital, which often funnels into real estate or financial services.
The concern, therefore, is that Britain's economy rests on unsustainable pillars: a housing bubble, foreign labour, and capital inflow, rather than robust domestic investment and innovation. There is a strong argument to be made that this situation is the result of an inevitable downward spiral since ~1900, when first the aristocracy, and then the government post-nationalisation, decided to sit on industries rather than refine, grow, invest. And then the big sell off and shut down of industry in the 1980s was chosen rather than trying to protect and modernise what we had left. I suspect the collapse of human capital after WW1 exacerbated this issue.
And not to be too doomerist about the whole situation, but the feel on the ground is certainly one of general malaise. Even in London there is a general decline- property used to be expensive, it is now exorbitant, rents were kinda bad, they are now catching up to property prices. The job market is ok, but is entirely focused on professional services, with a small tech sector and a huge swathe of service jobs. Leave London and the feeling is one of awful decline since the mid 2000s boom, with Manchester perhaps an exception.
Something I think illustrates the situation in Modern Britain well- over the last 5-10 years, many UK cities skylines, previously remarkably low, with the odd office building or council tower block, have filled with swathes of garish student housing blocks. The remarkable growth of the Leeds skyline, now filled with cheap student accommodation, sums up the problems with our current capital allocation- billions spent on building non-permanent, suitable only for 18-21 year olds, modern tenement blocks, with the aim of growing our 'university sector'- dozens of, putting it politely, 'second-rate' institutions, which rely on farming out expensive masters degrees to international students and pile-em-high style courses for British students.
This comment is persuasive and well-written. If you're verified to be right I hope this goes in the "best of" follow-up post. Of course, you could be completely wrong about everything -- I have no way to tell. Sounds spot-on to me, as you Brits might say.
I appreciate the compliment, but yes, I have provided no data or evidence for the claims above, so certainly wouldn't expect to be 'best of-ed'. The core elements- Britain underinvests in R&D/Capital, choosing instead to plough it into a housing bubble which stifles young mobility and talent, is well trodden ground. I hope people who are familiar first-hand with the territory nod as they read it and say 'that's the vibe'.
1) Some of this is anti Brexit propaganda. I think Brexit is a folly myself, but it’s clear there is a clear attempt to blame it for everything and even to exaggerate it’s affects.
2) that said GDP per capita isn’t everything. Wages gave stagnated in real terms, especially since 2008. They were bouncing back a bit until the recent inflationary period.
3) Britain seems to handled inflation reduction very badly. I might contradict myself here by saying Brexit could have a roll there, but if so why now?
I was thinking of replying until I read this. I agree, except the part about it being a folly. My understanding is it was really a question of sovereignty, not economics, such as that they didn't want to be forced to take immigrants from the rest of Europe they were finding disruptive. https://en.wikipedia.org/wiki/Causes_of_the_vote_in_favour_of_Brexit
As a sovereign nation, economics ought indeed to take a back-seat to governing your own people. Long-term, their economy will recover.
The Welsh can tell you how happy they are about that. I just learned the other day that within living memory they were burning down English vacation homes in Wales in reprisal.
As a Brit I’m not sure I would say we are becoming a third world country, but the last 10/15 years have been pretty bad. You can see this in the gdp per capita stats (link at the bottom, 2000s we were similar to Germany, now we are similar to France), public services (which have gone from being fairly decent to completely dysfunctional, try getting a non life threatening illness dealt with or any response from the police) and house prices now 9x average earnings, leaving many young people to be paying a very high share of their income of rent.
I think particularly for the smart/driven things have changed from a sense of you can achieve your dreams here (in the same way you can in US or maybe Germany and parts of east Asia) to the sense it’s more like much of Europe where you have to leave to make something of yourself (becoming a France or Italy, full of old people on massive pensions with little growth or dynamism). In the 2000s you could make ridiculous amounts of cash in finance if you worked hard, now it’s a pretty bad deal. Most other industries you would be much better off in the US either as an employee or setting up your own company (e.g. McKinsey associate in London earns £95k, in New York earns $185k with bigger bonus). Hard to think of many successful UK companies that have been created in recent years (Deliveroo, Sky, Ocado?). When I try to explain to american friends that £35k is considered an enviable salary when graduating university they all assume I’m joking. I think the average salary of an Oxford undergrad five years after graduating is around £50k.
Personally I’m somewhat optimistic, partly because the last few years we’ve been paying the costs of brexit (now shouldn’t be as much of a drag, and might even be some benefits from being outside EU regulatory nonsense), a lot of the UK industries have some prospects for growth (tech, pharma) and others unlikely to massively decline (law, finance, insurance). Ultimately think British political elite will fix it as they’ve done in the past, can take them a while but typically they ultimately take the difficult decisions.
That only goes up to 2018, so before Brexit, Covid, and the recent inflation and interest rate hikes. Obviously the last two affected other countries too, but all three have interacted to cause a lot of increases in food prices (as many foods are now more expensive to import) and mortgage repayments (which I think has thrown a lot of people, especially younger borrowers who had never previously seen rates above about 3%).
The NHS is noticeably under strain (I think all my lifetime you could get a same-day doctor appointment, and now it's usually a couple of weeks; I've heard (from friends, not just newspapers) that hospitals are similarly struggling). This may be partly because of European doctors leaving after Brexit.
There are a lot of strikes, which we haven't had significantly since the 70s: NHS strikes, school strikes, rail strikes, postal strikes, etc. This adds to the subjective impression that things are starting to fall apart.
"That only goes up to 2018, so before Brexit, Covid, and the recent inflation and interest rate hikes. Obviously the last two affected other countries too."
For Covid yes, but the UK was hit harder by the 2022 energy crisis than other Western European countries, because they were relying more heavily on gas as energy source.
Your direct link shows yearly gdp per capita growth, not total. But you can edit it.
But you are right. The per capita gdp seems to have maxed out 15 years ago at around $50k. To put that in perspective the previous decade saw it grow from $30k or so.
If you mean the IMF page, I was referring to the actual GDP there which, as I said, you can see by toggling the "Select an Indicator" button and selecting "GDP, current prices (Billions of U.S. dollars)."
It shows GDP rising until 2007 when it reached 3.09 T and then bouncing around until the present when it is at the nearly identical 3.16 T.
Yea, i think that per capita gdp is more useful as the size of population has increased. Even more useful would be wages. For instance GDP might increase this year (and definitely did last) but wages have not.
On most issues to do with the global economy, it would usually be reasonable to trust media institutions like the Economist and the Financial Times; and if they have a settled view about something, it makes sense that that view would come to be seen as representing reality.
The problem is that Brexit was personally, agonisingly painful for them, and the wounds haven't healed after seven years. I listened to an interview with the long-term editor of the Economist shortly after Brexit, and it felt like she was discussing the death of a member of her family. That depth of feeling doesn't just go away; and it hasn't.
So you have otherwise sensible, reliable British journalists credulously spreading nonsense that in other contexts they would see through instantly - one widely used model that purports to demonstrate British decline post-2016 compares the UK to a basket of other countries, 45% of which by weight is made up of the US and New Zealand.
The truth is that Western Europe is in long term relative decline. And that (as you've argued before) you can trust the mainstream media on nearly everything - but you have to know when you absolutely shouldn't.
(1) That graph ends in 2018, which makes it almost entirely useless. (All three articles you posted below are from 2022-2023, ie after COVID and during inflation.) (Better data doesn’t disagree, to be clear.)
(2) Inter-country comparisons aren’t too useful here, since AFAICT, pretty much every developed country currently is having economic issues around periods of declining GDP and inflation. At best, they can establish that other countries are also having similar difficulties.
(3) The US has had among the best developed-country economies during and after COVID; Europe in general was the worst affected by the Russian invasion, and so is worse off than the rest of the world; the UK speaks English, and so is most exposed to the US media, even if it’s not worse off than the rest of Europe.
Britain leaving the EU has somewhat slowed economic growth. While some British nationalists deny this even the conservative UK government admits this and has sought free trade agreements with several East Asian countries to offset this. However, pro-EU types have a political incentive to exaggerate the damage because they want Britain to rejoin the EU. Which probably would genuinely have a positive economic effect but not to the degree either Brexiteers or Europhiles want to pretend it does.
Further complicating this narrative is the fact that Europe in general is experiencing economic stagnation. This is not just a relative loss like the US where the US is roughly standing in place in terms of relative economic strength but competitors like China are catching up. It's declining in importance globally as its economy doesn't grow significantly while East Asia grows a huge amount. The "smart" Brexiteer case is that independence from the stagnant EU is allowing them to pursue trade deals with the more dynamic east and the US which are ultimately better.
As I see it, Britain's bet is that being a junior partner of the United States will both give them more success than chaining themselves to Europe and give them more freedom (since the US has no interest in British domestic politics in the way the EU does). To this end, it's aligning itself with other "individual" US allies like Japan or Thailand and attempting to take a leading role in things like Ukraine or Pacific defense.
I have no idea whether it will work (if I did I would be quite rich). But they are right that the EU has not really dealt with the degree to which they are experiencing economic decline. They really ought to be taking relatively radical action. Though whether Brexit as a specifically radical action is a good idea... well, it's too early to tell.
The idea the EU doesn't conduct or affect foreign policy is laughably wrong.
Likewise the fact the UK has gone out ahead of the US in agreeing to help defend its own East Asian allies or Ukraine points out that Britain has plenty of independence. Of course, the British left is not immune to the siren song of blaming foreigners (in this case Americans) for domestic disagreement.
> (since the US has no interest in British domestic politics in the way the EU does)
The obvious counter example to this is Northern Ireland, where President Biden has shown incredible willingness to intervene. Many people in the UK consider this a significant overreach.
That's because we nefarious Paddies are getting the Yanks who like to claim the Ould Sod heritage to intervene, since they're the 800lb gorilla in the 'special relationship' no matter what the Brits like to think, because we don't trust you Brits.
Now fair enough, I might feel like burning a picture of Leo Varadkar myself, but not in the context of "hey let's go back to the good old days of bombs and bullets":
Firstly, the US mediation in the UK-Irish disputes is international (and was requested by Britain and Ireland originally). The EU tried to somewhat take over but then used it as a Brexit chit so now it's back to the US. Secondly, the idea that that's anywhere near comparable to the amount of domestic influence the EU had is ridiculous.
Razib Khan interviewed a youngish British person and they both took the decline as written for the sake of discussing it. My takeaways were that salaries are low and there are few technical jobs, and that smart people can either leave the country or go into banking.
Even banking isn't that great; I have a friend who was an investment banker and moved to London for a job offer (thinking the promotion opportunities would be better there than in Australia), only to realise rather belatedly that the pay was slightly *lower* in London and the cost of living much higher - he was flat sharing with 2 other Investment Bankers, which IMO is rather damning about living costs; how is any young professional meant to afford rent when one of the best paying jobs possible is only enough to afford a flat-share?!?
A friend of mine (in the US) has been in a long distance relationship with an English guy for many years. They’re engaged and he will be moving to the US soon. He recently visited and I met with them for lunch after not seeing him for about 5 years. The couple are your stereotypical progressive lefties who don’t follow politics very closely. It was fascinating to hear him talk about issues England is facing, because he was pretty much confirming things that 5 years ago were considered far right talking points. He specifically seemed concerned about the quality of doctors and nurses which they’ve brought in from other countries, along with general concerns about immigration, wages, etc.
What extraordinary there is US growth since 2007 relative to peer countries (in GDP per capita). Some of this is currency movement though, as it’s measured in dollars. Nevertheless it’s impressive.
There's a button on the sidebar that can change the view to PPP, and yes, it is much better than I expected from general media portrayals (and not just in the US).
That's not the claim I've heard. The claim I've heard is that, since 1914, Britain has declined from being the dominant economic power in the world to a middling European economic power –– which is consistent with the data you cite. Although to a citizen of Great Britain with a historical view stretching back to 1914, this might well seem like "terrible economic decline."
Noah Smith had a recent post arguing that people's opinions about economic decline are better interpreted as opinions about how they are personally being affected by cost of living rises.
>Threads is supposed to be just like Twitter except better in a few aspects
I haven't used Threads, but from what I've heard, there is no desktop website to use it on and your feed is entirely algorithm-controlled with no option to only get the post of people you follow. To me, that's not "better in a few aspects", that's "worse in a few deal-breaking aspects".
I'm pretty sure most of the people leaving twitter, or contemplating leaving twitter, have grievances that go *far* beyond "not liking the designated hitter rule". A fair number of them claim that the "home team" is now run by a Literal Nazi, who is stacking the deck such that his Nazi pals will always win and will violently harass people like themselves. Most are somewhat less histrionic in their terminology, and I think all of them overstate the real threat, but the nature of their objections is the same.
If they stick with twitter, believing that, they're chumps. And some of them will be chumps. But not all of them, and maybe not a Threads' worth of them.
I think the idea here is that those people don't literally believe there are Nazis running Twitter. They just *say* that, the same way people will say "I will not rest until the real killer is found" and still go to bed in the evening. Virtually no one is hiding Jews in their cellar or desperately applying for immigration.
If I had to guess, I'd say the most genuine motivation at actual play here is a ploy for good old fashioned attention.
I thought this post was going to be about reddit and its various floundering clones at first.
I'm skeptical about threads as well, but Zuck has a history of building social networks that obsolete others. Facebook ate Myspace, and Instagram ate Facebook when Facebook become the thing for uncool olds (and Zuck bought it). Now twitter is flailing, and Zuck is trying for a three-in-a-row.
Also, keep in mind that threads isn't a truly new platform that you have to abandon your friends for, it's something that all of your Instagram friends are going to be inevitably pop-up-bombed to join the moment the initial hype and exclusivity is no longer in place.
While there's some truth in that, I think you're underestimating just how successful Instagram has been under Zuck's ownership - specifically with gen Z. Facebook was *the* social network for millennials, and then they turned around and did it a *second time* with Instagram - basically everyone in Gen Z has an Instagram account, without ever having made a Facebook account. This isn't just them transferring network effects, they managed to fully capture a wholly separate demographic.
And they did this all with an app that in 2012 (when they bought it) was mostly a niche photo-sharing app, that they managed to turn it into the *default* social media app for a whole generation.
Go back in time and look at Friendster, followed by MySpace, followed by Facebook. In each case, it just took a bit of technical superiority for the network effect to come to attach to the new network. In the case of Twitter vs Threads, Threads starts with the superior network effect of Instagram, so it might not need much technical superiority.
It's harder to imagine both coexisting for a long time, because network effects will likely drag people to one or the other, unlike with the kinds of markets where long-term competition can be stable.
Threads isn't just an alternative to Twitter with some better features. It's an alternative to Twitter at a time when Twitter is in crisis. Imagine if the NFL had a Bud Light level PR crisis, and the XFL came around at that exact moment. Then, I could see the XFL succeeding.
The only problem that I have with Threads is that the name invokes involuntary memories of the deeply disturbing nuclear war docudrama that the BBC broadcast back in the 1980s.
I'm trying to imagine what Sudden Death Baseball looks like, and I've got the image of the pitchers from both teams pitching to batters from both teams; we turn second base into one team's home plate, and the other team fields the foul section, and then both teams pitch and hit simultaneously.
...This should probably be the standard inning for Baseball 2.
I still don't know why Facebook beat out Myspace. But I'll say, very few people give up baseball because of the people in the audience, whereas that's the biggest complaint I hear about Twitter. One advantage could be that the worst trolls stay on Twitter where they're comfortable, so you don't have to deal with them on Threads.
Here is what I wrote about this topic a couple years back
MySpace is one exemplar, and nearly decade on, the lessons are still not well understood. I’m old enough to remember just how buggy the site was. Error rates were as high as 30 or 40 percent at times. In contrast, commercial operations like Yahoo! or Salesforce had at most 1 percent at the time. Simply put, MySpace wasn't created with the kind of systematic approach to computer engineering that went into Yahoo, eBay, Google, and Facebook. Instead, the site slogged through upgrades to implement flexible data caching and the distributed architecture needed to scale. Here is how one article explained it at the time:
Every profile page view on MySpace has to be created dynamically—that is, stitched together from database lookups. In fact, because each profile page includes links to those of the user's friends, the Web site software has to pull together information from multiple tables in multiple databases on multiple servers. The database workload can be mitigated somewhat by caching data in memory, but this scheme has to account for constant changes to the underlying data…
And although the systems architecture has been relatively stable since the Web site crossed the 7 million account mark in early 2005, MySpace continues to knock up against limits such as the number of simultaneous connections supported by SQL Server, Benedetto says: "We've maxed out pretty much everything."
For those that haven’t designed a system to freely scale, and indeed many commentators in the tech regulatory space haven’t, it’s hard to impress just how difficult of a problem engineers face. This is what unites the biggest players in the space, the ability to innovate to solve problems at scale without interruptions. How each site has done it serves as a case study in its own right. Todd Hoff’s site High Scalability collects these ideas. For example, here is how Twitter solved its issues with scaling. Here is how Facebook did it. Here is how eBay did it. As for Google, well, the comapany’s solutions are legendary in the computing world.
My idea for Baseball 2.0 is that everyone but the batter, pitcher, and catcher are on horseback. In case the batter hits the ball into play, his horse is waiting for him next to first base.
Horses are fast, but they aren't as maneuverable as people. I think this would make it harder to catch the ball. And it would most certainly make it harder to pick up the ball off of the ground. And what happens when a horse gets hit by the ball? Or a runner's horse slams into a baseman's horse? (Come to think of it, what happens when a polo horse gets hit by the ball?)
"Getting hit by the ball" was a bit part of why I'd excluded the batter and catcher. I thought of also excluding the infielders, too, but then they'd be at a critical disadvantage in both speed and collision-survival ability to the base-riders. I'm not too worried about horse-on-horse collisions: horses tend to be more sensible than humans about barreling into one another at speed (i.e. they'll slow down and try to get out of one another's way if they notice a collision is imminent: a lot of historical anti-cavalry tactics rely on this kind of horse-sense), and polo horses rarely seem to get seriously injured by collisions (with a quick googling just now, I found a 2015 veterinary journal article that concludes that polo horses have a similar injury rate to the general horse population). I'd expect horse-baseball to be somewhat less dangerous in this respect than polo, for similar reasons to why regular baseball has fewer collision injuries than soccer.
Good point about picking the ball off the ground. Maybe give the fielders polo nets? Maneuverability is another good point, at least for infielders, although for outfielders speed is probably more important. Enlarging the playing field somewhat, especially the infield, seems like it'd give the horses' speed to be more of an advantage and their maneuverability less of a disadvantage. Positioning could also change to emphasize speed over maneuverability, with infielders (especially the middle infielders) taking positions further back to get more time to react before the ball passes them, then charging inwards to field pop flies and slow grounders.
Getting hit by the ball doesn't seem to be a big risk to the horses in polo, either, which is surprising since it seems that polo balls go as fast or faster than well-struck baseballs. I think the difference is probably the size and composition of the balls. Outdoor polo balls are 3-3.5 inches in diameter, in between the sizes of baseballs (9 inch circumference -> 2.9 inch diameter) and fastpitch softballs (11 inch circumference -> 3.5 inch diameter) and are lighter than either (~4 ounces for a polo ball, 5 for a baseball, or ~6 for softball). Tweaking the ball design for safety would probably be tricky, especially given my previous suggestion of enlarging the field, since a longer and lighter ball probably wouldn't carry as far from being batted or thrown due to being more susceptible to air resistance. Maybe outfit the horses with safety armor if necessary?
>You can pray that something (more) catastrophic happens to Twitter and its users don't realize that they've actually been freed and rush to get into another prison, I suppose. Maybe the login wall does it, but that's a risky gamble at best.
Social media companies have catastrophic things happen to them all the time. Reddit was basically "Digg 2" for a while, but then Digg pissed off their user base and suddenly Digg 2 was the best link aggregator in town. If there was some sort of massive scandal in MLB and we discovered that the World Series was as fake as pro wrestling, "Baseball 2" would probably get a lot more viewers.
Twitter has been slowly combusting for a while, but there wasn't an obvious alternative for people to converge on - Mastodon is confusing, Bluesky is still in development, Truth Social is full of nutjobs, etc. Threads has the potential to be that rallying point since it's run by a well-known company that already has a massive user base.
The login wall is the dumbest thing Elon has done. Maybe he thinks we will all sign up to read a tweet, but by hiding tweets twitter moves from the town square to a private club.
The limit on number of readable posts was probably dumber, but it seems to have been ephemeral. (So far.)
Having seen several Twitter-killers over the last year or so, I'll have to see it happen to be convinced that the next one is going to do it. But I wouldn't be surprised if Twitter is in its decline phase whether or not there's a New Twitter to go to. Social networks seem to have their cycles, whether they actually shut down or just go from the hot spot to the place young people only log onto to wish their grandparents happy birthday.
Rather than baseball, the analogy I would use is phones. Imagine someone started a competing telephone network, with the limitation that the new network only lets you call people who are also on the new network.
Like phones, social media is locked in by network effects. It's inherently more valuable the more people who are already there (insofar as you value it at all). In order to compete, you need a major disruptive innovation or some way to quickly acquire a similarly-enormous user base.
If you wanted to have competition in social networks, I think you'd need to do something analogous to phones where the networks are forced to be intercompatible so that someone using phone provider A can call someone using phone provider B. Which would probably be a technical nightmare, and which obviously the current dominant social networks wouldn't go along with willingly even if the technical obstacles were removed. (Currently it's illegal to write your own user interface for someone else's social network.)
This is more-or-less how MCI started. MCI originally stood for "Microwave Communications, Inc": they provided a network of narrow-beam microwave radio relay stations in order to allow trucking dispatchers to communicate beyond the reach of traditional two-way radios. They soon branched out into providing business-to-business voice and data communication between fixed locations via the same network.
They didn't get really big until some time later, when three things happened in succession. They added satellite and fiber optic lines to their ground-based radio network, giving more cost-effective nationwide coverage and better signal quality. They integrated with the telephone switching network so you could use them to make long-distance phone calls through regular Bell System telephones without needing the recipient to be a subscriber (i.e. you'd make a local call to an MCI endpoint, they'd send it over their network to another MCI endpoint, and from there they'd make a local call to whoever you wanted to talk to). And of course, they won a big regulatory and legal struggle with AT&T that concluded with AT&T divesting its control of the local Bell phone companies and those companies being required to let their customers choose whether their outgoing long-distance call would be routed through AT&T's or MCI's networks.
In that case at least, network effects were a big enough deal to keep MCI from getting AT&T-level big until they managed to first route around them and then to be given full access to the Bell local phone networks by regulatory fiat. But even before that, MCI had been able to run a good-sized and profitable business making a parallel competing phone network.
There was also the infamous XFL of 2001, begun as a competitor to the NFL. That crashed pretty fast. The idea never died, though, and it came back in 2020, which, like other things in 2020, showed the Best Timing Ever for starting a new major venture. It took some major investment from The Rock to save it from a second death.
From what I'm hearing, it's still experiencing what could be plausibly described as "growing pains".
> The WBSC invented B5 in part to increase the odds that a sport similar to baseball and softball will be played in the Olympics,[26] with B5 meant to contribute to an ultimate goal of having a billion fans in the baseball-softball community by 2030[27][28][29] and helping to demonstrate the global reach of the games.[30] It also intended to create a game that would be more accessible and cheap,[9] as the only equipment used in B5 is the rubber ball,[11][31] and the field is much smaller than a baseball diamond or softball field,[32] with the game being playable indoors or outdoors,[33] even during the winter.[6]...In addition, the game is meant to be simpler to learn,[40] have more youth appeal,[41] and be more exciting, with players participating more frequently throughout the game.[35] Another claimed benefit is that with B5 being introduced across the entire world at the same time, no one country is likely to dominate the game;[42] the lack of the pitcher role, which is highly specialized, contributes to this.[32]
Nick Cammarata jhana-pilled everyone like two years ago. Scott wrote about it at some point but running the below search in the Twitter UI is maybe a good start:
As an American currently living in Canada for reasons beyond my control (I'll be back in a few years) the America-bashing I hear from progressive family members drives me crazy. There's a reason why the Canada-->US per capital immigration rate is 10-20x that of the US-->Canada direction. If you want a slower easier life, sure move to Canada. If you want a career, money, technology, excitement, interesting people, network effects, etc, etc, then move to the US.
Okay, that second one should be catnip to me, though I really would prefer a transcript rather than listening to a lecture.
However, I'm just moments in and already I'm going "Well duh, the reason it was popular among minor clergy was (1) they were the literate ones and this was not a grassroots tradition, this was very heavily based on scholarship needing some knowledge of Latin, Greek and Hebrew and (2) the reason it was so uniform was *because* it was based on scholarship and to an increasing extent, lack of knowledge of original languages - you see this in the corruption of Hebrew terms over time as people are just copying without understanding."
I realise the guy is ten times more educated than I am and has genuine academic degrees, but I'm always very wary when someone starts off with "Gosh, whatever could be the secret hidden reason for Blinkin' Obvious Thing?"
I'll have to wait for further into this to see what he says about exorcism. I'm not holding my breath, though.
EDIT: Yeah, I'm five minutes in, and I'm not impressed. I should definitely refrain from commenting on him being a goyische convert to Judaism, but the winking and smirking at your audience about the Christianity you left behind you is not inducing me to keep watching.
I think there's the guts of an interesting topic there, but I honestly would prefer an atheist scholar to deliver it.
otter841 will be captured soon and the break will be the safer for it..
..but then otters will no longer ride the waves on stolen boards at Steamer Lane:
https://www.npr.org/2023/07/13/1187295769/otter-santa-cruz-surfboard-surfers-california
I only started reading ACX after the move from SSC. I have gone back and read the top posts from SSC but still felt like I was missing out on some of the old classics. I built a website to resurface old content from blogs by sending weekly emails. Let me know if you have any suggestions of other blogs / content you’d like to see, hope you find it helpful!
https://www.evergreenessays.com/
Some suggestions (note that I endorse the first the most):
- Overcoming Bias (by Robin Hanson), which has been around for ages and is very rationalist-relevant (it was in fact the birthplace of the Sequences!)
- Marginal Revolution (by Tyler Cowen and Alex Tabarrok)
- Wait But Why (by Tim Urban)
- Ribbonfarm (by Venkatesh Rao)
Also, I'm happy that you've made this website in the first place! I am subscribing to read the Slate Star Codex archives, since I considered reading them during the SSC-ACX gap (June 2020 - January 2021) but never really made it all the way. Also, I suggest that you make the "Success" dialog more prominent.
Thanks for the feedback, just updated it with your suggestions! Will send out first posts next week. Currently working on allowing for more granular control of which posts you want, eg only send posts in past year, or only posts of type x, or send y posts in a week, etc.
Just saw this report: Forecasting Existential Risks. I'm still reading it but it has a section on the existential risk of AI.
https://static1.squarespace.com/static/635693acf15a3e2a14a56a4a/t/64abffe3f024747dd0e38d71/1688993798938/XPT.pdf
Kevin Drum commented on this, and I have to agree with Kevin's assessment...
"Let me say from the start that I think this is kind of nuts. I'd personally put the overall risk of extinction from everything combined at about 0.1% or so. But that's just me. I'm an eternal optimist, I guess."
https://jabberwocking.com/some-ai-experts-are-really-rattled-by-the-whole-extinction-thing/
>Apparently there's a group of AI experts who are absolutely obsessed with extinction and have spent more than 1,000 hours thinking about it. These people seem to have thought themselves into a frenzy and they're obviously driving up the average a lot.
Wow, I wouldn't have come up with that interpretation if I tried really hard. We should trust people who spend more time on things more than people who spend less time in general. Otherwise, our way of doing science and policy implodes!
Just want to flag that the report highlights a divergence between "domain experts" and "historically accurate forecasters", which goes like this:
1. The median expert predicted a 20% chance of catastrophe and a 6% chance of human extinction by 2100. Superforecasters saw the chances of both catastrophe and extinction as considerably lower than did experts. The median superforecaster predicted a 9% chance of catastrophe and a 1% chance of extinction.
2. The gap between forecasts of experts and superforecasters was not uniform across topics. Experts and superforecasters were furthest apart on AI risk—and less so on the risk of nuclear war.
Asked this question in one of the threads, but I'm also curious about what the broader board population thinks.
Who do you consider a "AI Doomer?" If one person says "AI development should be banned" and another person says, "Regulators should require AI products be subject to testing standards before deployment - and periodically thereafter - to confirm they work as intended, and should require developers to ensure that AI-generated outcomes are explicable" are they both "Doomers?" Just the ban proponent? Neither (since neither of them are stating a belief that AI is going to kill us all)?
I'd personally limit application of the "AI Doomer" to perspectives along the lines of "humanity will go extinct if we don't ban AI," but a lot of the pro-AI circles also trend pretty libertarian, so I think the actual usage may be be more expansive, in much the same way that some people in our corner of the world consider zoning to be "tyranny," "food regulations "basically fascism" and so on.
It's in the name. "If ignored, [Thing] could be our doom." Anyone who thinks unchecked climate change will wipe out humanity is a doomer. Anyone who thought an unchecked Y2K bug could lead to societal collapse was a doomer. Any Pompeiian who was worried about Mount Vesuvius erupting was a doomer.
The difference between a Doomer and whatever the alternative is is whether they think something can be disastrous, or just unpleasant. (I guess the second ones are Gloomers? I'm calling them Gloomers.)
Is it "if ignored [thing] could be our doom" or "if ignored [thing] will be our doom?"
Probably just a typo but the climate change example used a "will" and the Y2K example used a "could" so I wnated to clarify.
I'd challenge that a doomer should have to breach some threshold of belief that the doom *will* happen. Otherwise anyone who acknowledges the existence of any X risk at all would be a "doomer" about it, simply on the basis of the acknowledgement. Which would make almost all of humanity "nuclear proliferation doomers" or "large asteroid impact doomers," etc.
Probably a matter of taste. I draw the line at "could". If you think civilization is at any level of risk, you're a doomer.
And yes, almost all of humanity is a nuclear asteroid doomer. Although I guess we could add a second required layer of "we have to do something"; most people think there's nothing to be done about nuclear asteroids, and don't spend any effort on it. But I'd say they don't get called doomers just because nobody dislikes their opinion enough to do so.
For me, a Doomer is someone who puts AI x-risk within the next century at higher than one in a million, more or less. I'm not in favor of any AI regulations at this point, but I'd reserve the Doomer label for those who focus on the supposed threat of human extinction by AI.
>> For me, a Doomer is someone who puts AI x-risk within the next century at higher than one in a million, more or less.
Wouldn’t that render many AI advocates themselves *also* Doomers? “Unaligned AI is an x risk” isn’t exactly an uncommon perspective in those circles, at least from what I’ve seen.
In that case someone who (a) advocates for AI research and (b) believes “unaligned AI is an x-risk so alignment is important” would nevertheless be an “AI Doomer” if they thought the chance of an unaligned AI nevertheless slipping through was 1/500,000.
Granted, definitions can be whatever people want, but it seems like an odd usage if I can simultaneously advocate for AI and be an “AI Doomer”
Yeah, I would consider anyone who considers alignment important to prevent x-risk to be a Doomer.
Now, I'm attempting to answer your OP, but I don't think it's particularly important to define who is and who isn't one. I use the term Doomer below because it seems like the online world I read is splitting into two camps with, say, Scott and Zvi on one side and Tyler Cowen and Marc Andreeson on the other.
>it seems like an odd usage if I can simultaneously advocate for AI and be an “AI Doomer”
Maybe I'm wrong, but I get the feeling that most AI Doomers are Futurists who are or were very pro-AI but who view Alignment as the big filter to get through. Some of them have become so pessimistic that they might not be pro AI anymore. I don't think many (any?) of them are outright Luddites in general. Genuine anti-AI Luddites are more worried about losing their jobs tomorrow (For example, The Writers Guild of America). Or people worried that the AI is going to steal the election somehow. Those people are not Doomers because they aren't worried about x-risk.
Doomer is an arbitrary term—and functionally it's an ad hominem attack. And as you pointed out the goalposts for doomer are likely to move depending on the opinions of the person with their ax to grind. I think a better question would be: what percentage of people that there's a significant chance that AI will cause an extinction event for humanity?
As we've seen with COVID, the term "COVID minimizer" has migrated from the GBD folks and their ilk all the way to encompassing T Ryan Gregory—who up until recently was one of the leading COVID and infectious disease alarmists on Twitter.
How is Doomer an ad hominem attack?
Doomer is an ad hominem attack in that it appeals to feelings or prejudices rather than the intellect. It's used to negatively stereotype a person or group of people rather than a way of directly engaging with or addressing their arguments.
Doomer may not be a particularly nasty label, but it's a label all the same. Supposedly, we're all rationalists here, and we're supposed to argue things rationally rather than emotionally. I say "supposedly" because I've seen a lot of people on these threads confuse emotional arguments with rational discussion. While I admit I haven't been a saint on some of these threads—having let my sarcasm get the better of me sometimes—I've tried to focus on arguments and data rather than attaching dismissive labels to people. The use of ad hominem labels is an unsophisticated method of arguing one's point—at the risk of being ad hominemy it's sort of juvenile. I think you (and we ) can do better than fall into that trap.
But having said that, I think a lot of AI experts vastly overrate the threats of AI. And I've never heard a logically consistent scenario that would lead to the extinction of humanity. They seem to be channeling their fears rather than working through scenarios.
Some people from Metaculus have asked for clarification on some of my 2023 prediction questions. In order to say everything as publicly as possible so nobody gets an advantage, I'm putting it here:
>> 1. Will US CPI inflation for 2023 average above 4%? Are you comparing CPI at the end of 2023 vs the end of 2022, the average CPI inflation across 12 months (CPI Jan 2023 vs CPI Jan 2022, CPI Feb 2023 vs CPI Feb 2022, etc.), or something else?
Let's say https://www.bls.gov/cpi/ will be the canonical site. If they don't update in time, it will be some other measure of year-on-year inflation.
>> 2. In 2023 will there be more than 25 million confirmed COVID cases in China?
You mean happening in 2023 right? Not if the total number will reach 25 million in 2023.
Yes.
>> 3. In 2023 will Google, Meta, Amazon, or Apple release an AR headset?
Is the definition of AR that the user can see through it, i.e. that the screen is transparent, and therefore MR headsets where the user can see the environment through cameras don't count?
I will defer to Dan Schwartz from Metaculus, who wrote this question, but if he doesn't weigh in, then my interpretation is that any headset which allows AR capabilities counts, including if that happens through projecting the outside world onto an internal screen through cameras.
>> 4. In 2023 will a cultured meat product be available in at least one US store or restaurant for less than $30? Could it be a special offer that lasts for a limited amount of time? Do giveaways count? Does imported meat count? Does partial cultured meat count and if so is any percentage ok? If it's part of a menu do we divide the cost of the menu by the number of dishes or does the whole cost count?
I'm going to rule that promotions where it stops being available later won't count, but that if it's available just like any food, and some sale of the same sort that other foods sometimes get causes it to be on sale for below $30, I'll allow it. Imports are fine. Partial cultured meat is a good question; I'm going to rule that animal meat+cultured doesn't count, but plant-based-meat+cultured does if the cultured significantly changes the taste or texture or something. I'm going to rule that you have to be able to purchase it with less than $30 in your pocket, so if it's part of a package meal that costs more than $30, it doesn't count.
>> 5. In 2023, will SpaceX's Starship reach orbit? Does a transatmospheric orbit count or does it have to be at an altitude of more than 100km?
I'm not an expert on the definition of orbit. Google suggests an orbital spaceflight is one "in which a spacecraft is placed on a trajectory where it could remain in space for at least one orbit." I will consult with people who know more about orbits but try to generally use this definition.
>> 6. Will the Shanghai index of Chinese stocks go up over 2023? Are you referring to the SSE Composite Index? Because there are others, such as the SSE 50.
Yes, SSE Composite.
>> 7. On January 1, 2024, will an ordinary person be able to take a self-driving taxi from Oakland → SF during rush hour? Is the person being selected from a waitlist ok? You're saying "self-driving taxis in SF only operate at night" in your clarification, but these are waitlist only (as per a commentator).
I think the plain text of the question gives me one hour from the time I start worrying about this to make the trip, so if I have to worry about waitlists more than one hour before, no.
>> 8. In 2023 will a successful deepfake attempt causing real damage make the front page of a major news source? What's a major news source? E.g., would this count? What about TV news shows or radio programmes or news sites in general that don't exactly have "front pages"?
CBC seems major. I would count anything linked from the front page of their website, ie cbc.ca. Realistically this is a bad question because of articles like this where the deepfake itself isn't of major interest but is being used to make a wider point about deepfakes. I will probably be forced to count it anyway.
“ >> 5. In 2023, will SpaceX's Starship reach orbit? Does a transatmospheric orbit count or does it have to be at an altitude of more than 100km?
I'm not an expert on the definition of orbit. Google suggests an orbital spaceflight is one "in which a spacecraft is placed on a trajectory where it could remain in space for at least one orbit." I will consult with people who know more about orbits but try to generally use this definition.”
100km is the internationally accepted minimum altitude of space, and you more or less cannot remain in a stable orbit lower than that. In any case I doubt you’d event try to put Starship on an “orbital” trajectory there so it’s probably moot.
Obviously the Starship actually completing one or more orbits would seal the deal.
For a suborbital flight, I would say you should only count it if the flight is on an orbital trajectory at some point in the mission but intentionally deorbits via retro burn.
Maybe you could count if the rocket follows a full mission profile but (intentionally) does not go orbital.
What do I mean by this? Most often, launch vehicles going to orbit fire their engines multiple times - first a boost to apogee and then a restart and “circularization burn” to raise the perigee into a stable orbit.
Now a so called “direct insertion” where you do only one burn is possible (though less flexible) and I believe Falcon 9 uses this at least some of the time. That said, to do all the things Starship is supposed to do, it definitely needs the capability to restart its engines (probably multiple times) in space. This is not a trivial problem, so I would argue it needs to be demonstrated before you really call the rocket “orbital”.
If you really want to be lenient I guess you could count a launch where they achieve a successful full duration burn on stage 2 but it stays suborbital intentionally. But I don’t like that and I don’t think the planned profile of the test flight would have counted.
I'd also probably count it if it were on a trajectory with an apogee above 100 km and a perigee no worse than -100 km; that would demonstrate performance close enough to orbit as makes no difference, but give them a guaranteed reentry over open ocean if everything goes dead half an hour after launch.
That’s fair, assuming that’s where they aimed it.
Hello, Dan Schwarz from Metaculus here.
>> >> 3. In 2023 will Google, Meta, Amazon, or Apple release an AR headset?
>> Is the definition of AR that the user can see through it, i.e. that the screen is transparent, and therefore MR headsets where the user can see the environment through cameras don't count?
>> I will defer to Dan Schwartz from Metaculus, who wrote this question, but if he doesn't weigh in, then my interpretation is that any headset which allows AR capabilities counts, including if that happens through projecting the outside world onto an internal screen through cameras.
Agree with Scott, we should interpret "user can see through it" to refer to the capability, not the literal transparency of the material. Mixed Reality (MR) devices that do this with cameras would count.
This is a bad question because if MR headsets count then it should have resolved before it was even asked thanks to Meta’s Quest Pro, which released in 2022.
Following on from the "Americans created 'Italian' food" piece, here's Max Miller of "Tasting History" re-creating the Pompeian 'pizza' from the recently discovered fresco:
https://www.bbc.com/news/world-europe-66031341
https://www.youtube.com/watch?v=eEgSGkp8nMA
I'm re-reading Albion's Seed for the first time in a few years. This passage triggered my Gell-Mann amnesia alarm bells:
"The Puritanism of eastern England was not all of a piece. Several distinct varieties of religious dissent developed there, each with its own base. A special strain of religious radicalism which put heavy stress upon the spirit (Antinomianism) flourished among Puritans in eastern Lincolnshire. The more conservative and highly rationalist variant of Calvinism (Arminianism) found many adherents in London, Middlesex, and Hertfordshire."
I'm no expert on the different strands of the Reformation, but I've always understood Arminianism as being diametrically opposed to Calvinism, not a variant of it.
Maybe I'm wrong and Fischer is right? How much should this make me distrust the book as a whole?
If you want to get a good fictional account of religious conflict in England, try "Micah Clarke" by Arthur Conan Doyle.
It covers the Monmouth Rebellion and the scandal of the Taunton schoolgirls.
I think the confusion is that both Calvinism and Arminianism arise out of what were (are?) the Reformed churches of the time, and that it's debatable if Arminianism was a movement within the broader Calvinist grouping or something entirely separate.
They did eventually come to an official breaking apart and setting up as opposed camps, but within the specifically English context of the later Dissenters, probably they're best viewed as a strain of Calvinism or at least a breakaway group of Calvinists.
This is all pure ignorance on my part because I can't keep the finer points of which Reformed churches are what between France, Switzerland, the Netherlands etc.
Diarmuid MacCulloch's "Reformation" is helpful here; the Dutch split arose out of the conflict between the hardline Calvinists and those led by Arminius who had been trained in Calvinist orthodoxy but was questioning the theology. The English version were something different, at least according to MacCulloch:
"Many clergy and laity in England would have been horrified to think of themselves as Puritans, but they still regarded their Church as part of the international Reformed consensus; they were highly distrustful of the new tendency in divinity, and the new Archbishop Abbot shared their distrust. During the 1610s they gave avant-garde conformists a foreign name in allusion to the disputes about predestination that were currently tearing apart the Dutch Reformed Church – ‘Arminians’.
This was not so much because the English ‘Arminians’ were interested in the issues about salvation and predestination that preoccupied the Dutch. Although many of them did reject what they saw as a Calvinist doctrine of predestination, and there were friendly contacts between Dutch and English Arminians, the English party was more concerned with promoting ceremonial worship, use of the sacraments and a high view of clerical vocation (shockingly to many they had no hesitation in regularly calling ministers of the Church of England ‘priests’). None of these issues had any parallel in Dutch Arminianism. The Arminian label was a symbol that England was witnessing a rebellion against European Reformed orthodoxy as serious as that of the Arminians in the Netherlands. Some also nicknamed the developing Arminian clique ‘the Durham House set’, since from 1617 Richard Neile was Bishop of Durham, and Durham House, his London residence in the Strand, became an unofficial headquarters for the group, the shadow equivalent of Archbishop Abbot’s Lambeth Palace across the river Thames."
MacCulloch, Diarmaid. Reformation . Penguin Books Ltd. Kindle Edition.
MacCulloch, Diarmaid. Reformation . Penguin Books Ltd. Kindle Edition.
Thanks, this clarifies a lot. Glad to hear it's a sensible interpretation. Basically my whole understanding of American history is underpinned by Albion's Seed. If this was an outright mistake like I had assumed, I would have had to go through all my assumptions with a very fine toothed comb.
"The Vicar of Bray" is a satirical song from the 18th century and very good at showing the twists and turns in the Church of England from the early 17th to the early 18th century (reigns of Charles I 1625-1649 through to George I 1714-1727):
The Vicar of Bray - Lyrics
1. In good King Charles' golden time, when loyalty no harm meant,
A zealous high churchman was I, and so I gained preferment.
To teach my flock, I never missed: Kings are by God appointed
And damned are those who dare resist or touch the Lord's annointed.
(Chorus):
And this be law, that I'll maintain until my dying day, sir
That whatsoever king may reign, Still I'll be the Vicar of Bray, sir.
2. When royal James possessed the crown, and popery came in fashion,
The penal laws I hooted down, and read the Declaration.
The Church of Rome, I found, did fit full well my constitution
And I had been a Jesuit, but for the Revolution.
(Chorus)
3. When William was our King declared, to ease the nation's grievance,
With this new wind about I steered, and swore to him allegiance.
Old principles I did revoke; Set conscience at a distance,
Passive obedience was a joke, a jest was non-resistance.
(Chorus)
4. When Royal Anne became our queen, the Church of England's glory,
Another face of things was seen, and I became a Tory.
Occasional conformists base; I blamed their moderation;
And thought the Church in danger was from such prevarication.
(Chorus)
5. When George in pudding time came o'er, and moderate men looked big, sir
My principles I changed once more, and I became a Whig, sir.
And thus preferment I procured From our new Faith's Defender,
And almost every day abjured the Pope and the Pretender.
(Chorus)
6. The illustrious house of Hanover and Protestant succession
To these I do allegiance swear -- while they can hold possession.
For in my faith and loyalty I never more will falter,
And George my lawful king shall be -- until the times do alter.
(Chorus)
So you can see that Puritan, Calvinist, etc. in the Church during the immediate aftermath of Henry's Reformation, through the reigns of Edward, Mary, Elizabeth and then James on, would have been very fluid categories with some, but not total, resemblance to the Continental models 😀 (and with clergy of various sympathies in official positions as bishops, so you had 'Puritan' bishop and 'Arminian' bishop in the Church of England).
I'm an outsider here, so I do see Arminianism as a variant of Calvinism, where it grew up within Calvinism and along Calvinist principles but the major disagreement is on predestination. I think Arminians would have identified as belonging to the wider Reformed, even if the Calvinists had rejected them.
Under the rule of the Parliamentarians, there wasn't the same force holding things together, so a lot of little groupings were free to do things their own way:
https://en.wikipedia.org/wiki/History_of_the_Puritans_from_1649
"From 1649 to 1660, Puritans in the Commonwealth of England were allied to the state power held by the military regime, headed by Lord Protector Oliver Cromwell until he died in 1658. They broke into numerous sects, of which the Presbyterian group comprised most of the clergy, but was deficient in political power since Cromwell's sympathies were with the Independents.
With the abolition of the Act of Uniformity, even the pretense of religious uniformity broke down. Thus, while the Presbyterians were dominant (at least theoretically) within the established church, those who opposed Presbyterianism were in fact free to start conducting themselves in the way they wanted. Separatists, who had previously organized themselves underground, were able to worship openly. For example, as early as 1616, the first English Baptists had organized themselves in secret, under the leadership of Henry Jacob, John Lothropp, and Henry Jessey. Now, however, they were less secretive. Other ministers – who favored the congregationalist New England Way – also began setting up their own congregations outside of the established church.
Many sects were also organized during this time. It is not clear that they should be called "Puritan" sects since they placed less emphasis on the Bible than is characteristic of Puritans, instead insisting on the role of direct contact with the Holy Spirit. These groups included the Ranters, the Fifth Monarchists, the Seekers, the Muggletonians, and – most prominently and most lastingly – the Quakers.
The Puritan movement split over issues of ecclesiology in the course of the Westminster Assembly. In the course of the 1650s, the movement became further split in the course of a number of controversies. With no means to enforce uniformity in the church and with freedom of the press, these disputes were largely played out in pamphlet warfare throughout the decade."
Asking for a friend: Is there a point to fixing small mistakes in review contest submissions that are finalists, or will they be posted as originally submitted?
Is there now a culture war over AIs? I admit that I feel somewhat irrationally angry at Doomers. Why do I?
Is it because I am tribally aligned with Tyler Cowen? I don't know why that would be the case, but it's a possibility.
Is it because the AIs killing humans just seems like The Latest Hysteria -- along similar lines as Overpopulation, Climate Change, Trump, Covid and Satanism? Yes, Doomers do strike me as the drama queens of The Latest Hysteria, and that may well be why I hate them.
Is it because there is no escape from Culture War, and the more you try to avoid one the more likely you end up in another?
I don't know the answer. I don't really hate Doomers. At least not most of the time. They may even be right. They just annoy me. I need to get over it, I guess.
What is your working definition of culture war? I'm skeptical of just dropping the google result in but I don't have a working one myself, so just as a starting point - "a conflict between groups, especially liberal and conservative groups, that have different cultural ideals, beliefs, or philosophies."
Personally I find this definition unsatisfactory because it kind of renders *any* policy debate a culture war. Find me an ongoing debate, from BLM to where to put a local road, where there is disagreement, and people *aren't* dividing into groups to support or oppose particular options based on their different "beliefs."
Personally, I think there has to be some kind of threshold of scale for something to become a culture war, and I don't think AI is there yet. It's generating news coverage, and it's a surprising and new divisive issue within the kinds of niche "smart on the internet man" communities like this one that (a) were already interested in the issue, and (b) are usually aligned on many subjects such that its startling to see the break, but it feels like the average person on the street is going to have more and stronger opinions on trans rights and race issues, which is why those things seem to more to fit the "culture war" moniker than AI does (at least to me, at least for now).
[Edit] Also, who do you consider a "Doomer?" If one person says "AI development should be banned" and another person says, "Regulators should require AI products be subject to testing standards before deployment - and periodically thereafter - to confirm they work as intended, and should require developers to ensure that AI-generated outcomes are explicable" are they both "Doomers?" Just the ban proponent? Neither (since neither of them are stating a belief that AI is going to kill us all)?
There is absolutely a culture war, and doomers have a classic Eschatology. While they probably have a more rational view than Christians thinking the Rapture is soon, for instance, I think that the doomer movement needs more scrutiny.
When a group has all the hallmark traits of a doomsday cult, it makes sense to be more skeptical of their conclusions.
I think the comparison to the Rapture is quite apt, but that saying one is more rational isn't. Both stem from strong priors and not a lot of empirical facts. If we are valuing empirical facts (as implied by saying one is more rational than the other), then I feel like we need actual empirical facts, not strong priors. Like, an AI system recursively improving itself, or trying to kill someone. You know, actual things beyond a theory.
I relate to this, and have determined there’s a big fear component for me. On some level, validating the debate opens me up to considering the end of the world.
Additionally, adherents this particular hysteria feel more out of touch — partially because many of them I know are genuinely out of touch people overall. At least with overpopulation, Covid, climate change, etc, everyone can sort of see things slowly going wrong and picture what things would look like if they do go very wrong. This one, however, has a sort of nerdy ivory-towerness that feels completely removed from the lives of everyday people, so despite being a nerdy ivory-towerer myself, I feel reflexively off-put.
Lastly, the huge gap between the threats that currently exist from AI and the end of the world are so massive and poorly explained that even I as an ML engineer have trouble picturing exactly through what means AI will cause my demise. Obstacles that seem very critical to me, like the requirement for massive scaling and improvements of difficult robotics are brushed off in nearly every explanation I see. This plays into my aversion to putting resources towards it too — even if the expansion of alternative energy isn’t necessary or doesn’t save the world, at least the air smells a bit better, geopolitical dependence on oil is reduced, and I save money with my electric car. However, if AI turns out to never be an issue, what was the use of all that money, brainpower, and worry? It just feels like a waste already (especially the worry part).
I’m still totally torn on this issue, but think you really described a cultural/emotional aspect I hadn’t been able to put my finger on.
Well it seems to me that what's happening to you is *not* that you're getting sucked into a culture war and hating on the outgroup, and also is not that you've mentally joined a cult. Sounds like good old-fashioned fear to me, and I don't think it's silly to be very confused and uneasy. I am. It's just really figure how plausible the pessimistic view is.
Am I allowed to find both sides infuriating?
The doomers, yes. Anyone who thinks p(doom|unaligned AI)=1.0, or anything even close to that, is too enamored of their own smartness. Thinking that this doom is imminent because their buddies made ChatGPT, and then that they can avert doom with their Extra Smart AI Alignment Research, yeah, massive overconfidence.
But, I'm just as skeptical of "highly valuable new technology". Particularly for any AI that is going to come out of anything we are talking about now, I'm not expecting it to be any more valuable than e.g. social media. Some legit use cases, but mostly noise and social strife and generally making life a little bit worse for everyone.
I'm pretty sure AI is going to get us cheap, mass-produced mediocrity *long* before it gets us Canned Super Genius, and cheap mediocrity isn't going to bootstrap itself. It's just going to generate noise, and be used as a weapon by people playing status games against one another (and against me, whether I want to play or not). If and when we do get a Canned Super Genius, I think it's very likely to be persistently misaligned, and if I think that's more likely to mean that it keeps going catatonic or self-destructing than that it tries to paperclip us all, that still doesn't add up to "highly valuable".
Highly Valuable, is not impossible but requires a lot of things to line up just right. Doom, is not impossible but requires a lot of things to line up just wrong. Mass-produced mediocrity, that looks to be surprisingly easy.
I think you have what I think of as a hate structure going on regarding people who think there'a a high probability of doom. You imagine they're all mini-Yuds, and that their sense of doom is just the delicious opposite side of the coin that, when heads, shows them being smart as fuck. I doubt that's true. Some I've seen posting on Twitter seem to be just very scared people who expect catastrophes of many kinds -- we get crispie crittered by climate change, covid wrecks our bodies and soon half the world will be bed-bound, nukes, pesticides, etc etc. Lots of very scared people around these days. As for the tech bro's who think they and the other smart tech bros will make AI safe by clamping the alignment attachment to their dicks and aligning that bitch -- eh, that's just one school of optimism. Actually, most of the people I know have heard that some are worried about AI, but believe that the scientists will figure it out like always and things will proceed as usual.
BUT I agree that mass-produced mediocrity looks pretty easy to achieve. Sort of like building a mall.
Lately I've been watching old Johnny Carson Tonight Shows from the early 70s and am fascinated by how foreign the culture seems. The past is a foreign country, they say. A striking difference is how sexual so many of the jokes are. Many of the things Johnny or the other guests say to the female guests would be considered misogynistic today, because they visibly react and comment upon the attractiveness of a women in comedic, salacious ways. It's hard to tell whether the female guests would prefer if the males behaved differently.
For instance, Johnny will often ask the guest, male or female, about their dating life. It seems to be considered natural that these famous people might want to talk about their private dating lives. Johnny might also talk and joke with male guests about good lines for picking up women. References to getting drunk and having one night stands is normal.
In one episode Johnny makes a joke about having had VD during the war, and then it hit me. I had assumed I was seeing a cultural product from the peak of the Baby Boomer era, and that explained why the show was so loose in discussing sexual matters. But then I realized Carson himself had been in WW2.
It got me thinking that maybe the WW2 vets changed the culture more than even the Boomers. After all, nobody was more worldly than the vets. Consider all the cultural leaders circa 1970 who were WW2 vets: Johnny Carson (with the biggest TV audience ever in history), Hugh Hefner, Kurt Vonnegut, Norman Mailer, J.D. Salinger, Joseph Heller -- all who pushed a countercultural take on America to various degrees. Carson himself isn't normally considered part of the "counterculture", but I dare you to watch those 1970s Tonight Shows and report back that they don't seem countercultural compared to your image of the early 1960s or the sexually conservative world that is 2023.
Maybe the Baby Boomer were overrated when it came to cultural change in the 60s and 70s. Maybe it was more about their veteran daddies.
It’s interesting to think about the Silent Generation contribution to the Sexual Revolution, but I think the GI/ prostitution angle is a red herring here. Casual attitudes to prostitution can easily coexist with quite strict sexual ethics around family formation and the behavior of “nice” women. Victorian London, for example, had huge numbers of prostitutes of all ages and sexes, with relatively destigmatized access for men who could afford it, while famously enforcing a tight code of sexual purity and chaste speech around its marriageable middle-class girls—and importantly, that ratcheting-up of domestic purity norms happened *with* the return of world-weary veterans from the Napoleonic Wars.
To take another example, Shakespeare’s plays joke openly about fun with whores in a way that would shock right across the political spectrum if written today, but there are also multiple plots that turn on the importance of female chastity in the aristocratic marriage market (note, again, as imagined by a middle-class person).
It doesn’t feel revolutionary for Carson to joke about catching the clap, but it does feel new for that to be perceived as a great topic for humor on a mainstream TV show aimed at a mixed audience. If the middle-class moms in the audience laughed excitedly, rather than finding it disgusting and offensive, then presumably it was something the moms wanted to hear and something the producers wanted to have said to them. I suspect it’s probably about legitimating a fantasy of libertine consumerist pleasure as a high-status coastal norm, to a willing audience of Middle American rubes. (Look at these Hollywood people, so rich and well-dressed, so morally relaxed and carefree! *They* understand that it’s OK to have a bit of fun now and then.) Feels very much of a piece with the mainstreaming of _Playboy_ as an openly accepted part of the suburban dad library, and that’s definitely also about class and consumption (smoking jackets, satin sheets, marble pillars...).
Didn't Scott have a piece on this at some point? At the very least, I seem to remember an extended conversation about how the cultural figures associated with a given generation are often actually members of an earlier generation (most of the famous rock stars and artists of the 60s were members of the Silent Generation or--as you pointed out--members of the Greatest Generation who actively participated in the Second World War. The Sex Pistols and the Ramones were Boomers. The bubble-gum pop stars and spiky-haired rap-rockers of my Millennial youth were mostly Gen-Xers). Which makes sense. The members of a generation generally aren't going to be old enough to be artistically influential to their own peers. Teenagers might be the prototypical consumers of youth/mass/counter culture, but they by and large aren't big producers of it. Even teen pop idols and child stars don't really count, because their artistic output was mostly designed by the adult producers and songwriters around them (although I'm curious how this is shifting with the current generation of teenagers. Current social media stars/influencers seem to be much closer in age to their audiences than the rock stars of yore, with much more control over their own messages than earlier teen stars)
I wonder about those young influencers and the material they release. It's clearly less polished, more repetative, and often drops the ball at key moments. It's the opposite of highly choreographed shows like you see on TV, especially as related to reality TV, which is heavily edited to take real (or "real") events and hype the excitement.
My kids started watching Mr. Beast recently and I've seen a few episodes. Climactic moments often fall apart. Filming seems random and out of order. Props fail in highly predictable ways. This is bad TV, bad storytelling, bad writing. It reminds me of hanging out with my friends as teenagers. I'm thinking that my generation liked the polish because we could see the improvement over the haphazard stuff we could do on our own, but the upcoming generations get too much polish and appreciate the authenticity of young influencers who don't have adults running the show and making everything perfect. It may be related to the fact that kids are spending a lot less time hanging out with their friends in random unstructured ways. That's an itch not being scratched, so they get it from digital sources instead of physical.
I was thinking specifically about the effects of the war on the psychology of these people. For instance, Carson's attitude toward sex was probably formed in those years. Casually referencing sex with a hooker was no big deal to him or much of his generation. Prudery likely doesn't survive major wars.
But taking that and adding it with your point makes me wonder if the so-called Sexual Revolution was merely a transfer of values from ex-GIs to the younger generation.
Are these old episodes of the Johnny Carson show online?
I'm getting them on the Roku app.
I have never done any research into climate change. I have just always deferred to people smarter than I am and the consensus seems obvious.
I'm now looking at an investment in a climate sensitive region and would like to understand more.
Does anyone have a recommended primer / link / book etc. (that is politically motivated)?
> I'm now looking at an investment in a climate sensitive region and would like to understand more.
Can you expand on this?
97% of AI Risk researchers agree that AI poses serious risks.
Therefore, we should end all AI research, and you can’t disagree with me because that’s the scientific consensus.
You want something politically motivated?? Do you have a political point-of-view preference?
New article on my Substack - it's a review of "Innovation Breakdown: "How the FDA and Wall Street Cripple Medical Advances". This is a rather obscure book which tells the story of a skin cancer detection startup that got screwed over by the FDA, delaying the entrance of their product to market by several years and severely damaging the reputation of company in the eyes of dermatologists.
https://moreisdifferent.substack.com/p/book-review-innovation-breakdown
Maybe I've missed this discussion in previous threads, but, um, why Threads? To me, it seems that Twitter is a) money-losing enterprise, and b) huge magnet for unwanted regulatory attention. Why is Meta trying to replicate this seeming trainwreck?
One needing another Meta account (Instagram) to use another Meta product (Threads) is a red flag.
Another is the requirement to delete the Instagram account to delete Threads. Meta claims to be fixing this, but who knows how long that will take. Perhaps signing a suicide pact with users’ Instagram accounts contributed to the alleged 100 million registrations.
I suspect we think that with better execution we can make it genuinely positive, for the users and for the business. We have more expertise and resources in content moderation than Twitter ever did
> more expertise in content moderation
Imagine someone hearing this at the peak of the Cambridge Analytica & troll farm Fake News scandals of 2017... and yet it's 100% accurate.
Cambridge analytics was more about decision making around 3rd party API access than it was “content moderation” per se… but yes fake news troll farms were / are a content issue
I’m not sure I completely understand the question -- are you claiming that harassment is permitted from progressives but not to progressives?
Is it a money losing enterprise if it just piggybacks on the existing Meta infrastructure?
Good point. I don't know! Do we have any Meta accountants in the audience?
Are there therapists who focus on helping people who get into conspiracy theories and distrust experts, which leads them to avoid doctors and medication, which worsens their condition?
....alright, let's avoid the uncharitable take.
In my very limited experience, the big issue isn't finding an appropriate psychologist, it's finding an appropriate caretaker. People with serious issues like this don't need a lot of complex therapy, they need to take incredibly powerful pills every day for the rest of their life that make them miserable and they hate, otherwise they will quickly become incredibly paranoid, like won't stand within 20 feet of phone lines, and in general just degenerate incredibly rapidly. The issue isn't, ya know, finding a solution, the problem is generally making sure that the person is consistently taking these pills, that again they absolutely hate, on the worst day of the year. People who are, underneath it, just people with all our failings. From the outside, their internal experience of skipping a pill seems like no more of a moral failing than buying a pizza and candy when you're depressed, just with horrifically worse consequences. This generally means there needs to be a caretaker person, who needs to ensure this person is on their meds, every day, for like 10 years, which is inevitably going to lead to a lot of screaming fighting matches between the caretaker and the paranoid. It's a brutal, thankless task that consumes decades of a persons life. Finding someone who will do that is difficult, finding someone who will diagnose the paranoid schizophrenic as a paranoid schizophrenic and get him pills and weekly meetings is easy.
I don't think that's a specialty. But most competent therapists, talking with someone like that, would make a special effort to stay out of the role of expert, to not give any advice,to try to be particularly transparent and honest, and to have a stance of "how can I be helpful?" I have sometimes been able to develop a working relationship with someone who had that stance. However, if the person is really deeply furious and mistrustful that is not going to work either. The last person of that kind that I saw said that being a psychotherapist was the ideal profession for imposters with no degree, because everything we say and do is bullshit anyhow. Expressed no interest in seeing proof of my degrees, because said it was still possible that I had stolen the name and the identify of the person who had actually gotten the degrees. I ended up suggesting he consider co-counselling, which I think is actually sort of high-risk, because this man was so utterly incapable of trusting authorities and yet was greatly in need of help.
Yes, it’s quite addicting to be proven right again and again after being told you’re crazy for years.
Does anyone have any clear examples of someone skillfully avoiding Gell-Mann Amnesia? I personally just avoid news in general, but I wonder if there are other options. I will preemptively say that my news abstinence is trivially compatible with all strategies of the form "don't trust a news item unless you [take some extra, effortful step]"
I think you are making too big a deal about GMA.
Some stories are covered by sober minded experts and others are covered by complete idiots.
If the NYT has a piece about portage canoe trips that I know for a fact is not fully baked I don’t thInk I should regard every story on the site with suspicion.
Michael Chrichton came up with the idea and named it after the guy who discovered the quark. The physicist was brilliant and Crichton is a so-so sci fi writer. Now that naming is a bit of deception in itself. How much traction would the Michael Chrichton Amnesia meme have gained?
The point that GMA is making is not "all news articles are bullshit". It's "some news articles are bullshit. Remember this and be appropriately skeptical, especially when you are not an expert in the topic".
We know that _some_ articles even in well respected outlets are garbage. So if you are not capable of telling (which most people are not on most topics), then the natural response is to be wary of articles.
I agree we should remain skeptical and pay special attention to the source. Do your own homework. My quarrel is with the particular name, GMA. Crichton wasn't really coming up with anything new and naming it Gell-Mann was a bit of cheesy sleight of hand to make it seem like it was something cool and new.
It is especially cheesy because Gell-Mann was famously skeptical and very conscientious about getting his facts right. And he knew a *lot* of facts. I remember a Seth Lloyd book in which Seth said that meeting GM for the first time is intimidating because GM knows about 10 times as much as a typical person on any given topic. Another eminent physicist (it may have been Stuart Kauffman) opined that Gell-Mann knows more facts than any other human. It's impossible to adjudicate such a thing, but FWIW I can't think of anyone I'd put ahead of Gell-Mann at his peak in that regard.
Zvi recently linked to a tweet from lawyerBrian Manookian describing how he got off the hook for his student loans: https://twitter.com/BrianManookian/status/1674963884703088642, and saying anyone whose student loans have gotten passed from lender to lender can do what he did.
The approach is to demand proof of chain of title, i.e. proof that the loan company now servicing the loan now owns the loan, which I believe they prove by showing a bill of sale. He says that these companies that buy up big batches of student loans are sloppy, and do not do the proper paperwork, and probably will not be able to supply proof that they own a loan. The person who got the loan then petitions the court to declare that the person does not owe the loan company any money. Court does so, person sends judgment to loan company and they cannot pursue collection any more. (Seems to me that nobody else can either -- because the previous owners of the loan have sold it to a collection company, and really don't own the loan any more.)
This sounds plausible to me, and my daughter and her boyfriend both have substantial student debt and are interested in trying it. However, we would like to ask a lawyer for advice before trying it. Four lawyers have now turned me down when I have asked them to first advise us, then carry out the process if we are not confident we can do it correctly. We would be paying them for their time, of course. Lawyers have said it's outside their realm of competence. The steps Manookian describes sound to me like extremely simple basic stuff, and in fact Mazookian has supplied copies of the letters he used as models, so it's hard to understand why lawyers keep turning me down.
Anyhow, can anyone help with information? I am open to any possibility. You can give me general info then fade away, or you can actually do the filing required. You can give info in a post here, or you could send me an email or we could talk online. We can do it anonymously or using real names. I will happily pay your hourly fee, though I think I can ask all my questions in about half an hour. If you are law student or legal aide and understand the process Manookian is describing I am happy to work with you also.
And, by the way, I am sure there are many people on here dealing with school loans. You could help a lot of people at once. Or you could do them one-by-one and have a lot of clients you can process fast and easily. Maybe you can bring down one of the loan shell companies.
This is fascinating. I don't have student loans, but I do have a 98k mortgage that has been passed around three times, probably under similar circumstances as student loans (ie, as part of a big sloppy batch of loans).
On the one hand, you would think that even crappy mortgage companies would be more careful about bills of sale given the large amounts of money involved.
But on the other hand, mortgages come with a lot more paperwork than student loans do, and if they're buying mortgages in huge batches, maybe they actually might be sloppy enough to lose critical ownership paperwork.
Huh.
During the mortgage collapse of 2008 there were a couple news stories related to this I think -- some seemed to be about the inability to address the crisis because it couldn’t be determined who owned the loans, and some about a lender/loan servicer that didn’t have its paperwork in order and so couldn’t evict people who didn’t pay.
I've been doing some reading since originally posting my comment - the TLDR is that mortgages seem to take documentation much more seriously than student loans and there are lots of additional protections for mortgage lenders, at least in Washington State (where my condo is). From what I can tell, even if they can't put their hands on the original paperwork, they can just do a affidavit swearing that they lost the paperwork, but, like, they totally do own the loan. But on the other hand, it looks like it's occasionally worked for people in foreclosure?
Either way, this seems like having such a low chance of working that I'm not going to seriously pursue the idea. From what I can see, "produce the note" is much easier to do with student loans because, unlike real estate, there is no underlying asset or public records about it.
I can tell you that these "Chain of Title" arguments are very important in patent litigation. In some jurisdictions "incorrect recording of inventorship or ownership is, in itself, fatal to the validity or enforceablity of a patent." (quote from here: https://blog.patentology.com.au/2014/02/inventorship-and-ownership-importance.html)
I don't know if the law is as clear when it comes to chain of title for loans, but maybe a lawyer with experience as a patent attorney would be more comfortable making these arguments?
Thank you, that's a good suggestion and I will follow up on it.
I’m not sure about student loans, but I did use a similar technique to get rid of credit card debt a long time ago. The collection agency was not able to provide any proof of the original debt so they had to forgive it.
Ooh! I like that story.
I am not a lawyer, and cannot offer legal advice. This seems like it's not going to work effectively; if the shell company doesn't have record of the paperwork, they can get it from the people they bought it from. If they don't remember that, once you take them to court they can demand you tell them who that was as discovery, and then get their paperwork. They can probably make you testify under oath that you don't believe you owe the debt, with whatever perjury penalties that incurs. It really seems like you're just banking on them not showing up and losing by default, which a lawyer isn't going to want to do.
Tennessee seems to have a $25 paywall on their court cases. Might be worth your money to look up exactly what happened with that linked case, or find others like it.
Thanks, I will look it up. Guy who gave the advice said that it's not cost-effective to these companies to defend themselves so long as claims like this are rare. They have a huge number of loans, and know they will lose many to defaults. Losing to me is just one more. Calling me into court and having their lawyer grill me sounds like it would cost them a hell of a lot of money. As for getting the paperwork from the previous company -- I think a lot of them just collapse and go out of existence, then sell the loans they are holding cheap. Also, I would not be claiming that I (actually my daughter, but let's say it's me) did not know who had made the original loan and do not believe I borrowed money. I would just be saying yeah, I borrowed it and will pay it back to the organization that can prove to me it is now the owner of the loan. But I'm not sending $30,000 or so to a company that cannot document that I owe them money. No perjury involved.
Recently convinced a relative to hire a lawyer for an eviction proceeding; it cost them $4000 in total. Not chump change, but much less than the property value. The guy's arguing it's not worth the collector's hassle, which is banking on them not showing up.
Was going to say my loan servicer has changed at least twice since they were paused for COVID (and possibly more than once before that) and I could maybe honestly say I don't know who the previous ones were since I've kinda not been paying attention until they're a more pressing problem and the repayments are starting up again.
No you would not need to lie in any way. The way this works is not that you pretend you don't know who the loan company is, or who had it before, or that you owe a loan. All you are doing is saying that yes you took out a loan, and you will repay the loan to a company that can prove via paperwork that it bought the loan and is now the owner of it -- BUT that you are not sending money to a place that can't even prove it owns your loan. It is reasonable to demand that a place trying to collect a large sum from you can document that its claim is legitimate. If it can't document it, it is reasonable not to pay it, EVEN IF you think it is extremely likely that they did buy the loan, just didn't do the paperwork. I mean, it's like if you bought a rug for $5000 online, and start getting emails from the place saying send us the money -- but they can't tell you the date you bought it, the item number, and they can't prove that they are the online rug store. It's just a guy writing from his private email saying "you bought the rug from my store a while ago, and we shipped it. No I have no records of the date you bought it and no proof that we shipped it, but I'm telling you we did."
Right, I was just saying if it was a question of being asked to provide the paper trail on the loan servicer's behalf I would be able to give an honest "I dunno".
Anywho I'm sort of concerned that Brian Manookian might be a bit of an internet clout chaser, looking at the rest of his Twitter feed. Seems worth noting that he's qrt bumping a tweet from August of last year and keeps saying he's going to set up a website detailing his process and has yet to do so.
Like I don't think he's lying or that he didn't do what he's saying he did per se, just that I sort of suspect there's probably details and circumstances that are being obfuscated in a way that prevent this from being something scalable at all.
Yeah, he's definitely got a big ego and no heart of gold, and I noticed too that he hasn't set up a website, so bad at follow-through. However, one of his Twitter threads contains what's probably the crucial document, which is the one you send first. I think there's a fairly good chance that this would not work, but it looks promising enough to me to put some effort into figuring out for sure. A couple of the lawyers, before they declined to be hired, have said things in passing about why the think it would not work:
-This is not a way of getting your loans discharged. (But I already knew that. They will still be on record in various places. On the other hand, there will be no record of ongoing failed attempts to collect. The loans will exist forever in a sort of limbo: You borrowed the money, but the original lender sold the loan, and the buyer resold it, and the last organization that bought it can never collect it because the court has declared their claim permanently invalid. And none of the earlier loan-owners can collect because they sold the loan, and really do not own it any more.)
-Manookian's posting about something he did in 2016. Loan collection companies are different and better organized now. (Have no idea whether that's true, and it's not clear that the lawyer who told me that has every tried this maneuver with somebody's student loans. How would he know that the loa collection companies are better organized now, given that he said does not handle student loan debt in this practice. His field is "debt defense," but he seems to mostly help people who have very large debts and are in big trouble.
-Most student loans are federal loans. The loan companies that take them over don't really own the loans, they are being hired by the US government to do the collection work. (He would not tell me how he knew that or how to find out more about it.)
The thing is, I have had a lot of middle-aged white men (please forgive me if you are one) in authority tell me something is wrong or impossible, and my observation is that if I argue, politely, or ask probing questions a lot of them double down and start throwing out reasons why they're *positive* their view is right, and in fact these back-up arguments turn out not to be based on solid info when I look them up on google scholar or wherever. They seem to be incapable of just saying, "based on my experience and intuition that's wrong, but I can't really back that judgment up and it's possible I'm wrong." And about several important matters it has turned out that I was in fact right.
If you are interested enough in this possibility to look into it on your own, let me know.
Was thinking about this the other day and couldn't quite recall where I'd seen it. Can confirm I'd also be interested in knowing more.
I have little interest in the outcome of this, but I'm curious why four lawyers have turned you down.
My first two guesses would be 1) it's too novel ("outside my realm of competence" sounds like the professional phrasing of that) and 2) it violates some professional norm against lawyers being highly lawyerly. More interesting to me is whether any of them relied on consequentialist reasoning, like this attempt to apply Econ 101 concepts:
'If this becomes a trend, then lenders will probably take notice and then take steps to ensure that each buyer of each loan can demonstrate proof of chain of title with ease. Won't these additional steps taken by lenders raise administrative costs at all lenders? Won't that increase the interest rates that these companies charge future borrowers?' (Similar to the model in which shoplifting by some increases the average prices that consumers pay.)
Are plaintiffs' attorneys known to make decisions on that sort of basis? It doesn't sound like the sort of thing that they do, but then I have never asked.
Where is the AstralCodexTen for people interested in law, anyways?
Part of it may be that lawyers have an enforceable profession duty of candor to the court. They're not allowed knowingly to put perjured testimony on the stand, present forged documents, present spurious legal arguments in briefs, etc. You don't have to volunteer evidence or arguments against your client's interests, and you can definitely present things in the most favorable light to your client's claims, but you aren't allowed to outright Saul Goodman it up.
A lawyer could probably defend an action brought by a loan servicer by demanding proof the servicer owns the loan and nit-picking every shortcoming in the documentation the servicer provides, sure, but bringing an action on the client's behalf seems like it might require more of a prima facie case. If getting such a case heard would require the lawyer to file an initial brief saying "my client doesn't think he owes this money to the respondent", then an ethical lawyer (or at least a lawyer who doesn't want to risk being sanctioned by the court) needs to reasonably believe the claims he puts in his brief.
Wait, you are misunderstanding to plan I have learned about. There is no perjury involved. There is nothing that is even sort of vaguely *like* perjury. The person with the loan does not say, "what loan? I never took out a loan." They say, "yes I took out a student loan and I will repay the loan to the original lender, or to any other organization that has bought the loan. However, the organization that asks for me to make loan repayment has to be able to prove they are actually the owners of the loan. They should have a bill of sale or similar for my particular loan." If you win, it is because the loan organization bought a bunch of loans in a big batch and did not do the required paperwork. There is actually a law about this sort of thing to protect consumers -- something along the lines of, you can't dun somebody unless you have documentation of the proper kind showing they owe you money.
Anyhow, some may think this approach is kind of sleazy, because it's getting out of a loan one really does owe on a sort of technicality. But my daughter and her boyfriend have oppressively large student loans and want to do it, and I am not about to be judgmental about that.
I know one lawyer pretty well, and he's actually a highly ethical person who worries about things like that in real life. But his view is that the job of a lawyer is to make the best case he can for his client, and that considerations like those you mention are irrelevant. It's OK to take the OJ Simpsons of this world as clients and do your best to get them off the hook -- not because they deserve to go free, but because a system where everybody gets a good defense protects the justice way better than letting each lawyer's squeamish scruples rule. I agree with that point of view. And I can tell you that none of the 4 that turned me down sounded like they gave a shit about that kind of thing. I think the reason might be that as lawyer jobs go mine is really puny. Manookian estimated that a lawyer charging typical hourly fees would make maybe $1000 learning about situation from client and then doing all the paperwork involved in this project.
If lawyers were in the habit of making decision based on the type of logic you describe, we would have dramatically fewer lawyers. The majority of the lawyer profession exists to take advantage of exactly these kinds of bureaucratic overheads and process inefficiencies.
For anyone who has quit their job to take a break to pursue personal projects, how was it? Do you regret it?
I am currently torn between burning desire to quit my job and pursue a startup idea, and ruthless pragmatism that just considers the opportunity cost of lost paychecks.
I have no kids, no rent/mortgage, and many years of living expenses in the bank, but somehow can’t rationalize to myself actually taking a risk. If you ever did quit to pursue a personal interest or purely take a break, I’d love to hear from you!
I took a three month break last year to explore career options and writing, and it was the best time of my life!
I got a job again soon after just from networking in my local area, but I do not regret it at all. I think we should normalize taking sabbaticals, it's extremely useful for mental health and breaking out of rigid mental models.
I don't think people are morally or ethically against taking sabbaticals. I think we're financially against it. I would love to if I could afford to just be jobless for three months. The privilege...
I've taken at least three year-long breaks from work, first at age 20 and most recently in 2015 when I started a profitable side business manufacturing electronic products. Now that I have children, I don't expect to have that sort of opportunity again.
I grew up in near-poverty and started work in my mid-teens. My early life left me extremely frugal and with a horror of debt, so living off savings and finding work when money ran out was a naturally self-motivated process.
I pursued hobbies, did some traveling, and had some adventures, but frankly I never accomplished as much as I expected (except in 2015). Romantic partners were generally supportive but unimpressed.
I feel that I was lucky to be stably employed during periods of economic turmoil like 9/11, 2008, and the pandemic. I would caution you that various economic indicators don't look so hot, and seem likely to decline.
Probably there were opportunity costs, but certainly opportunity gains. No apparent career impact, and no regrets. All of the trite homilies about non-fungibility of time and the advantages of youth are true. Have a plan, keep to a more-or-less heathy sleep schedule, set expectations low and curiosity high. Do it!
How hard will it be to get another job if you need one?
I've never quit to pursue a passion, but my last batch of unemployment lasted six months, during which I made no progress toward any goals. So, don't assume the time will make things happen.
I have a relative who started their own company, successfully, and they mentioned mostly working 20-hour days seven days a week for about two years. Expect pain.
On the pro side, I think of Shamus Young's autobiography; https://www.shamusyoung.com/twentysidedtale/?p=13953. "It’s been easy over the last year to just keep coding at home and punching the clock at Taco Bell. This isn’t my dream, but it’s a tolerable substitute."
I'd say, go in with the expectation that every dollar you put into the thing is completely lost.
Thanks, I’ll take a look at that. I’m an ML engineer, so I don’t expect difficulty finding a job again, though I have a pretty chill and pleasant remote gig now that I don’t want to throw away. I am definitely in the (very) tolerable substitute territory, which is part of what makes it hard to walk away.
I certainly expect any startup process to be grueling, but feel this might be my best time to give it a go. I’m in my mid/late 20s so if I’ll ever have the energy for that sort of pace, it’s now.
I’ve done it multiple times, though not always to pursue personal projects. Sometimes just to live life. It’s paid off every time with the possible exception of quitting in January 2020. I wrote more about job-quitting in my recent Substack post titled “the fifth bomb”
You might also find the recent book “die with zero” an interesting perspective on the time vs money problem. The author frames it more as an energy problem, with your time/energy being more valuable when you’re young and losing value over time. So selling your time/energy for dollars when you’re young is a bad exchange rate (from that perspective) which supports quitting jobs often.
Also recommend the YouTube short film Slo Mo (I think it’s called) about a surgeon that loses his vision and quits working.
Interesting, thanks for sharing! I’d love to take a look at your blog. I’m in my mid 20s now, so I’m certainly curious in the long-term opportunity cost equation for risky endeavors.
I do suspect growing up in a working class family one generation removed from extreme poverty has made the opportunity cost numbers seem gargantuan, though in reality, a six month attempt that returns 0 would really just mean working an extra ~3 months at the end of my career, which sounds trivial.
It’s tough stuff to grapple with so I appreciate the alternate perspectives!
Yeah I also grew up in a working class family which definitely left a chip on my shoulder too. Definitely skews perspective a bit and great that you’re aware of it
I’m late thirties and hesitate to give genetic advice, but generally think “do what you want to” seems to play out better for high agency people vs “suffer and grind”
it’s worked for me 🤷♂️
I try to do a 24 hour fast twice a week, sometimes three, if I'm feeling ambitious. I go noon one day to noon the next, so I'm asleep for 8 hours in the middle of it. That seems to make it much more tolerable, in my experience.
Whoops. I meant this as a reply to Nolan Eoghan below.
Handicapping a potential Musk-Zuckerberg MMA fight. For some reason I had the impression that Zuck has been training for years, but I recently read that he only started BJJ 2 years ago. 2 years is.... not a very long time to be training, at all. At this point he probably doesn't know what he doesn't know. Reportedly Musk weighs at least 70 pounds more. I would personally not recommend that a white belt with 2 years of training and 1 competition take an extremely high-profile fight with someone 70 pounds heavier than them. For one thing, MMA fights start standing, and takedowns in no-gi take a lot of energy and skill unless you grew up wrestling. No-gi takedowns against someone much much larger is going to be very, very tough, and very, very exhausting.
Like, I don't any think of the say 170 lbs. high school or college wrestlers in my old gym could just easily take down a 240 lbs. white belt. In the gi maybe for a judoka, but not in no-gi.
Anyways, there's not a ton of fields where I know substantially more than Mark Zuckerberg, but this is 1 of them- I would personally not advise this fight. Contra Hollywood, size really matters in fights
70 pounds make a difference talking about actual athletes, where you can assume most of it is muscle. Elon looks chubby and soft. His cardio is probably terrible too.
Has Elon trained in martial arts at all? Because 2 years is a hell of a lot more than 0. I'm bigger than Elon but I've never trained in any sort of fighting so I'd hesitate to go up against the Zucc.
Respectfully disagree. You seem to argue that Zuck can’t win but that isn’t an argument for why Elon can. I can’t see him landing a KO punch so not sure his path to victory. Plus, Zuck is 13 years younger if it goes the distance.
The more interesting question to me is who will be on the undercard!
Update for the zero people who care: the combo diet of keto/intermittent fasting continues to work well. I'm down about 20 pounds in about two months. It's been remarkably easy, too. The keys to success are a) try to eat lean meats but don't be afraid to splurge now and then on some bacon or whatnot b) get some exercise but don't overdo it, cause that's when falling off the wagon is most likely, and c) don't be afraid of a little or even a lot of caffeine.
Does not keto allow all the bacon you can eat? And meats as fatty as you want?
Perhaps, but I found eating lean meats lead to much better results. For Memorial Day, we had a cookout, and an entire package of ground beef patties was left untouched, so I spent the following week eating basically nothing but hamburger patties, and the weight loss slowed or stopped entirely during that time, so...ya know.
How intermittent is the fasting?
I try to do a 24 hour fast twice a week, sometimes three, if I'm feeling ambitious. I go noon one day to noon the next, so I'm asleep for 8 hours in the middle of it. That seems to make it much more tolerable, in my experience.
Do dried psilocybin mushrooms loose potency over time? Yesterday I ate some that had accidentally popped up in a Tupperware container that I was using to store damp coconut coir a couple years ago and the visuals were just not there.
Yes, the psilocybin decays over time. Just to add a little more detail over other replies, the chemistry involved is an oxidation reaction: the psilocybin reacts with oxygen from the air to give non-active blue compounds. https://www.chemistryworld.com/news/mystery-of-why-magic-mushrooms-go-blue-solved/4010870.article I.e. fresh mushrooms turn blue when you bruise them because air gets in, and this represents a slight loss of potency. (only slight - the blue colour is very strongly blue and a little goes a long way)
Things other people have suggested to stop this happening sound reasonable to me. Keep in an airtight container, under vacuum or with one of those oxygen-remover sachets (not just desiccant which only removes water). Keep moisture out because being thoroughly dry can be a physical barrier to oxygen.
Chemistry happens slower in the cold so I would normally suggest that the freezer is good. I think the counter argument (that Eremolalos is thinking of) is that when you take them out of the freezer, they are so cold that water from the air condenses onto them, which makes them damp and reduces shelf-life if you're not using them straight away. Same if your container is not completely airtight in the freezer (get some good containers with a clip-on lid with a rubber seal).
If you do thaw and you're planning to refreeze, maybe thaw completely to room temp before opening, or even better, thaw the containers by putting them in an even bigger airtight container with desiccant.
Forgot to mention the other thing from that paper - part of the oxidative loss of potency is assisted by enzymes in the mushroom.
This is analogous to green tea leaves turning into black tea - the catechins/tannins in green tea turn black in an enzyme-assisted reaction with oxygen. You manufacture tea leaves by destroying the enzyme with a quick blast of heat, once the desired level of oxidation is reached.
It may be possible to do something similar with shrooms, where a quick pasteurisation-type treatment just after harvest increases shelf-life dramatically. I'm sure people have looked/are looking into it.
You're supposed to not chop them up because that exposes more surface to the air, & store them with little envelops of desiccant stuff. Last and fussiest advice I read was something about protecting them from oxygen? Or maybe air? and recommended a vacuum storage system.
Or you can just grow some more ; )
Yeah they were dried and run through a coffee grinder. Each little batch stored in a small jar with desiccant packs.
I once wrote some code to calculate the decline in potency of pharmaceuticals. The big deal in that case was ambient temperature. They held up better when stored at lower temps. Not sure if this is true of mushrooms but I could store the jars in the freezer.
Or wait for some more to accidentally pop up in some damp coco coir stored in a Tupperware container as I’m sure meant by growing some more. :)
I read somewhere convincing that you should not put them in the freezer, even after they are dried. Up until then, I dried them, put in desiccant envelopes and stuck them in jars in the freezer. I absolutely cannot remember the reason why that's a bad idea but whatever it was I read convinced me. So yeah, damp coco coir
Yes they definitely lose potency over time. Anecdotally I’ve heard 7% loss of potency per quarter but 🤷♂️
That's a really specific/precise number!
Yeah, the number will be that precise only under a similarly precise set of conditions; temperature, amount of oxygen and moisture, mixing and so on. Often though, its not just X percent of the original potency per quarter, its X percent of the remainder which is why chemists talk in half-lives (half of the remaining potency disappears in every Y-month period).
Psilocybin is not a simple example though - from the paper I linked above:
"However, no single blue compound exists in Psilocybe mushrooms. Various oxidative pathways contribute, to different extents, to the reaction. Each of them produces various chemical (including isomeric) species. [...] The substrate to oxidant ratio seems to affect the mechanism and color."
In my experience, fresh bruised mushrooms are a "pure" blue colour, with dried mushrooms being a more greenish-blue. This is consistent with slightly different compounds being formed in fresh vs dried, and these will have different rates of formation depending on conditions. Likely more difficult to model with code than Gunflint's pharmaceutical example.
Thanks!
Why is there so much less generative AI music than other forms of media? Is it because music that's been created to-date is all heavily copyrighted, so it's not available as training data? Generative AI obviously can create text (ChatGPT, etc.) and images (Midjourney, Dall E etc.). I've seen/heard a very small amount of AI-generated music, but infinitesimally less compared to these other media. My understanding is that existing generative AI was trained on an enormous amount of free/open-source/maybe kinda stolen media that was found online. (I mean Midjourney will randomly throw in watermarks to some images, so obviously watermarked pictures were used in the training data). Was this just not an option for music, whose copyrights are generally protected by ruthless highly litigious large companies?
I mean, intuitively to me music seems a lot simpler than imagery- if an AI can create pictures, it seems that technologically it should be able to create music too. So I'm guessing the barriers here are legal, does that sound correct?
There’s a lot of music in the public domain, but the existing recordings are copyright, and so is the published sheet music . It’s a little weird. But an AI can learn a lot of public domain music by ear. Why not? Hook up a microphone play a song. It should have perfect pitch… after that it should be able to play with what it has. But I agree with others in this thread that I doubt it’s worth it because of the nature of music and how people want to consume it, and what people usually want out of making it. Also, I don’t know how much it gets you over a modern synthesizer.
I remember a funny experience at the Lincoln library for the performing arts years ago. I went in looking for sheet music for the song “streets of Laredo”, it one in a long line of descendants of a 16th century folk song.? The rakes lament, Saint James infirmary, a young soldier, cut down in his prime etc. anyone can use that tune, and write their own lyrics to it basically.
I found the sheet music, but they would not let me Xerox it. It was perfectly appropriate for me to copy it by hand onto another sheet of notation paper though (which they provided me with) and then do what I wanted with it. Most of the western Canaan of classical music is free from any copyright. The actual recordings are a different story though; anyone can do Don Giovanni if they want to, and they don’t have to pay Mozart a nickel. So you could get an AI to do variations on a theme by Mozart, for instance and knock yourself out. You have to stay away from Marvin Gaye, though.
It's possible the sheet music _was_ copyrighted even tho the 'info content' wasn't.
Exactly.
This is something I've long noticed. Generative music is highly feasible: spend some time listening to random OpenAI Jukebox samples https://jukebox.openai.com/ and note that this is generating raw audio from scratch, posted in *2020* using 2019 stuff, and not as scaled-up as GPT-3 was, and there's vastly more accessible audio data. You could also get fairly reasonable samples from non-raw-audio data sources with a few tricks (https://gwern.net/gpt-2-music) which sound even better if you put a tiny bit of effort into them (https://soundcloud.com/theshawwn/sets/ai-generated-videogame-music). So, where is it all?
It seems to be a combination of: (1) the copyright situation is much scarier. Music comes with *all sorts* of bizarre complex IP laws and rights that don't apply to text or images, and this is because of (and empowers) groups like the RIAA. Every commercial startup doing music-related generative work seems to keep things very quiet, as blackbox as possible, restricting to licensed datasets when possible, and definitely not releasing stuff. (2) lack of major hobbyist interest - people may like to listen to music, but a lot of it is parasocial, there's superstar winner-take-all effects, and people don't generally want to make their own music the way they do their own images & text - think about porn/hentai as a huge driver of image synthesis. There is no porn/hentai of music generative models. (3) relatively demanding compute compared to other modalities. Audio is just very, very bulky: a piece of music is made out of millions of individual datapoints (X frequencies times Y milliseconds), while you can make a recognizable image with just 64x64=4096 pixels. Jukebox wasn't as demanding as GPT-3-175b was... but it's still a lot more compute than most people want to spend. Even sampling Jukebox is pushing it: early use of the released Jukebox model deterred hobbyists when they realized it (a) didn't fit in their consumer GPUs and (b) would take like half a day to generate a full audio file. Yikes. It'd be a lot better today, of course, but is nowhere near the instant gratification of a GAN or near-instant of the OA API or heavily-optimized diffusion model.
I'd also note that images specifically has benefited from luck: you would be seeing way less generated images today, and you would be unimpressed by image progress, if Emad hadn't done what all of the giants refused to do and bankrolled a decent FLOSS image generation model. (You would be wrong to be unimpressed, because there were, and still are, many much better image generation models than the Stability ones, and trying to judge image generation SOTA by what an SD model can or cannot do is a serious mistake - but they are all locked up behind barriers so you can't use them.)
Thank you for saying this. I had the same thought, and it is one of my exhibits that I point to when I try to make the case that AI's impact is a lot more marginal relative to its perception. The nothingburger that is AI music is proof that AI is much more incremental than people expect. AI seems to frequently arrive in a context where generation is already very cheap, potentially free, or even negative. In the case of music, the negative costs are borne by all the striving artists donating in-kind time to become hits.
The music industry has already gone through multiple waves of being concerned about cheap, fake and derivative music, and sometimes those controversies ironically become foundations, such as the initial aversion to autotune until current entrenchment.
There is a recurring ebb and flow of new innovations, such as even something simple as analytics-driven producers cracking the code as to what it takes to make a hit in the 1960s, to the modern Rotten Tomatoes-like algorithmic PCA to find the right sample banks, producers, writers, and overall secret juju to produce that true derivative—but extremely catchy—hit.
https://tedium.co/2023/02/04/why-do-modern-pop-songs-have-so-many-credited-writers/
Gen Music is simply not as profound or disruptive as other techs like autotune. There is still no secret to creating the perfect hit, just as there is still no secret to creating the perfect trader. (something something arms races)
And this extrapolates to other domains. I find it interesting that Gen Art came at roughly the same time we were just getting used to free high-quality images from Unsplash. At first, it was like, "Wow, Stable Diffusion solves all my clipart problems" to "Wait, these problems were already solved by Unsplash."
Or that GPT-based coding is downstream from a world where we were already googling 90% of our issues to find a good Stack Overflow link and then blindly copying-and-pasting. The existence of GitHub Copilot initially seemed like, "Wow, this is going to end all programming", and yet here I am now spending instead of 90% of my time fishing for Stack Overflow, I'm spending 80% of my time. Or, correction, I'm sending more time, because in parallel with these innovations we have the further advancement of community-heavy frameworks and packages, which seem to lend themselves to a dependence on Stack Overflow. For example, you cannot separate the rise of React from the rise in our dependence on Stack Overflow, unlike earlier eras where PHP programmers only looked up issues occasionally.
The future is boring. Object Detection's improvement is logarithmic to current AI R&D expenditure. Moore's Law is downstream from Rock's Law. Secular stagnation puts a hard cap on demand. 1960: The Year the Singularity was Canceled (by Scott Alexander), etc.
You can drastically reduce the size of generated music (and presumably the compute and data set needed) by having your AI generate sheet music instead of audio.
That just occurred to me the other day, too. And, it would improve the sound of generated music a lot. Right now it sounds like a muddy mishmash. Rather than AI having to try to discern "what a guitar sounds like" it would make more sense to have it try to discern "what notes and rhythms are enjoyable." It could generate MIDI files and/or sheet music instead of audio files.
Thanks, yes I didn't think about answers 2 or 3 at all. The universe of paying consumers who want to make their own music at home is probably pretty small. (And most of the ones who do probably want to copy existing, copyrighted artists- 'make a collaboration of Taylor Swift and Drake and Tupac' or something).
I'm skeptical the music industry will ever want to license their IP to AI either, just to ensure that they're not cannabilizing their existing artists
I remember hearing a talk about a music-generating AI something like a decade ago (possibly a TED talk?). Some musician had creative block while working on an important commission and decided to write an AI that could sample a bunch of sheet music in a given style and then compose new songs in that style, then turned it on his own past work to fulfill the commission, then gave a talk about how he did it.
I don't think copyright is an issue; most of the training data in the current popular generative AIs is already copyrighted, but the stance of the companies is that using something as training data is not a copyright infringement (and I think they are correct).
But "there isn't a lot of sheet music posted online in places where you can download it for free" might be an issue.
How about IMSLP? There's quite a bit of free sheet music on there
"If an AI could generate new Mozart operas on the fly, I’d listen to them a lot! But it can’t"
It has been able to generate "classical" music for 3 decades. Nearly as long for jazz. There is no market. Even less for opera.
> Caveat: I have no idea what I'm talking about but this sounds pretty good)
Well you sure had me fooled. Geez I was ok with temporal, but tension, motifs and climaxes! I was sure you had to know what you were talking about.
Hmm, but AI isn't too bad at (very) short stories. You'd think they'd be able to produce a few bars of nice music (are they? I haven't kept up with the music side of things).
I would love to read a linked AI-generated short story that wasn't absolutely horrible, formulaic schlock.
a couple bars can be enough to get a piece going. "Scrambled eggs" did end up turning into "Yesterday".
Scrambled eggs
are such an easy meal to make…
Gimme a break gimme a break break me off of that… I am drawing a total blank - The Office
https://www.tiktok.com/@theoffice/video/6914076535381150982?lang=en
I don't think that's a difference. A piece of sheet music only "changes over time" in the sense that it has a beginning, middle, and end, and text also "changes over time" in that sense.
Questions about affirmative action:
- How does affirmative action in college admissions actually work? Are there any concrete examples of how admissions would look with and without affirmative action?
- Do we know specifically what changed in UC admissions policy after Prop 209?
https://en.wikipedia.org/wiki/1996_California_Proposition_209#Effect_on_enrollment,_graduation,_and_income
- How was Prop 209 enforced? If admissions policy is totally opaque, what's stopping the university from "considering race"?
- How was the Bakke 1978 ruling against racial quotas enforced?
- How will the SFFA v Harvard ruling be enforced?
- Do people understand what affirmative action is? In this Pew survey, a plurality said affirmative action is "a good thing", but "only 17% say college admissions should take race into account."
https://www.pewresearch.org/race-ethnicity/2023/06/08/asian-americans-hold-mixed-views-around-affirmative-action/
- "The Copenhagen Interpretation of Ethics says that when you observe or interact with a problem in any way, you can be blamed for it."
How have college administrators avoided being blamed for racially skewed student populations?
https://laneless.substack.com/p/the-copenhagen-interpretation-of-ethics
- "[Opponents of affirmative action] claim that making room for an applicant of one race necessarily requires a university to reject a different applicant."
Don't supporters of AA claim this as well? Isn't this the *whole point* of AA?
https://www.vox.com/policy-and-politics/23405267/affirmative-action-supreme-court-ruling-race-harvard-unc-chapel-hill
The University of Arizona stopped the consideration of race in admissions in 2010. Its most recent incoming freshman class was 47% minorities.
People who claim the Supreme Court somehow diminished the inclusion of minority races in college admissions are mistaken.
- How does affirmative action in college admissions actually work? Are there any concrete examples of how admissions would look with and without affirmative action?
Sort of. What we have from discovery are (1) evidence a lot of this was just open racism from admissions committees and (2) theoretical class statistics in a counterfactual world where things like personality scores were removed.
- Do we know specifically what changed in UC admissions policy after Prop 209?
Yes. There was a brief surge in white and Asian percentages (and to some extent women over men) before it returned to normal. One of the UC administrators was recently caught on tape admitting that they did this by illegally ignoring the ruling and simply avoiding keeping any records of it. So now he's probably getting removed and there's probably a lawsuit incoming.
- How was Prop 209 enforced? If admissions policy is totally opaque, what's stopping the university from "considering race"?
- How was the Bakke 1978 ruling against racial quotas enforced?
Poorly. See previous example. However, now a bunch of people have cause of action against the UC system. Remember, though, that discrimination is not an intent issue. If you can simply prove disparate impact (such as through admissions numbers) that is sufficient.
- How will the SFFA v Harvard ruling be enforced?
Probably about the same. There's already a bunch of administrators walking right up to the line of saying they're going to find work arounds. Which probably means more lawsuits.
- Do people understand what affirmative action is? In this Pew survey, a plurality said affirmative action is "a good thing", but "only 17% say college admissions should take race into account."
Affirmative action doesn't just apply to college admissions. It's perfectly consistent to argue that you want (say) a representative bureaucracy but not representative college admissions. In fact this is more normal. In places like Singapore or Lebanon it's considered necessary for social peace to make sure there are minority administrators and politicians but who gets to be a doctor or engineer is entirely race/religion blind.
- "[Opponents of affirmative action] claim that making room for an applicant of one race necessarily requires a university to reject a different applicant." Don't supporters of AA claim this as well? Isn't this the *whole point* of AA?
They deny it because if no one is harmed no one has standing to sue. Of course people are in fact losing out on slots and so do have standing. Supporters just don't like that fact because they want to be able to do racial balancing (which was illegal even before this decision) and they're going to say whatever they need to say to try and get it past the courts. Freddie DeBoer's a good source of honesty: he believes it's a reparations program and so do most of its supporters. But such racial programs are illegal and so they (broad they, not Freddie specifically) lie to get around that fact.
The long term goal here, as they tried in California, is to repeal the parts of constitutions that require race neutrality so they can do it legally.
- Does the "ban on affirmative action" mean that it would now be illegal for a politician to (as Biden did) say they will appoint a black Supreme Court justice? Would it be illegal for a corporation to deliberately appoint a black board member so their board looks more diverse (if someone could prove in court that this was their logic?)
No to the first question. Politicians have wide latitude to select people for whatever reason they wish. The second one is already illegal unless the business can show a compelling interest in why having a member of that race on the board is important. And diversity is now not an acceptable compelling reason.
The literal phrase diversity is downstream of legal decisions saying that while racial discrimination was illegal without a compelling interest that diversity was a valid compelling interest. Because these people are lawyers or lawyer adjacent they just copy-pasted (in some cases literally) the justification the Supreme Court said was allowed.
"discrimination is not an intent issue. If you can simply prove disparate impact (such as through admissions numbers) that is sufficient."
That is certainly not the case re employment discrimination. Rather, disparate impact establishes a prima facie case for a violation, and then the employer has the burden of showing a legitimate business reason for the policy that creates the disparity. Note also that the individualized use of race endorsed by the court will likely result in disparate impact; a black applicant is more likely to be able to show how race affected him than is a white applicant.
"diversity is now not an acceptable compelling reason."
Not necessarily. After all, the court left open the possibility that diversity could be a compelling interest for military academy admissions, so there was no holding that diversity is per se never a compelling interest.
1. Correct, once you have shown disparate impact you then have to find a safe harbor for why the disparate impact is allowed. If you can't then it's discrimination. The previous safe harbor was diversity. That is no longer allowed.
2. The opinion specifically pointed out this is not allowed and that the theory advanced by the liberal justices it might be allowed is wrong.
3. The court simply pointed out that it was not being asked to consider military academy admissions and that it wasn't certain that appointment to military academies, which is done through politicians much of the time, are the same thing as general university admissions.
"2. The opinion specifically pointed out this is not allowed and that the theory advanced by the liberal justices it might be allowed is wrong.'
I am not positive what "this" refers to, but the decision says: "At the same time, as all parties agree, nothing in this opinion should be construed as prohibiting universities from considering an applicant’s discussion of how race affected his or her life, be it through discrimination, inspiration, or otherwise." It seems to me that a policy implementing that suggestion will likely have a disparate impact.
"3. The court simply pointed out that it was not being asked to consider military academy admissions and that it wasn't certain that appointment to military academies, which is done through politicians much of the time, are the same thing as general university admissions."
What the court said, in footnote 4, was this: "The United States as amicus curiae contends that race-based admissions programs further compelling interests at our Nation’s military
academies. No military academy is a party to these cases, however, and none of the courts below addressed the propriety of race-based admissions systems in that context. This opinion also does not address the issue, in light of the potentially distinct interests that military academies may present" That certainly sounds like they are saying that diversity might constitute a compelling interest in the context of military academies, and hence that the court did not hold that diversity can never be a compelling interest. And there might well be other instances in which diversity might conceivably be a compelling interest, such as hiring police officers in a diverse community. I am not saying that that is likely, but only that your statement is too broad.
More questions:
Does the "ban on affirmative action" mean that it would now be illegal for a politician to (as Biden did) say they will appoint a black Supreme Court justice? Would it be illegal for a corporation to deliberately appoint a black board member so their board looks more diverse (if someone could prove in court that this was their logic?)
Re the Supreme Court appointment, no. AFAIK, no court has held that the Fourteenth Amendment limits the appointment power of the President or state executives. Re corporations, board members probably are not employees for the purpose of federal anti-discrimination law, so I doubt there are any limits on such appointments.
"If admissions policy is totally opaque, what's stopping the university from "considering race"?"
Being able to prove from other evidence that racial discrimination is happening. Completely coincidentally I'm sure, there's a (successful) push to get rid of standardized testing and other "objective" measures of academic merit.
They go into detail as t how it worked in both Harvard and UNC in the opinions on the latest SC case:
https://www.supremecourt.gov/opinions/22pdf/20-1199_hgdj.pdf
The podcast 99% Invisible had an episode which I think would be of interest to this crowd: The Frankfurt Kitchen: https://99percentinvisible.org/episode/the-frankfurt-kitchen/
The Frankfurt Kitchen was the first modern kitchen - designed with flat, uniform countertops, an electric stove instead of a hearth, and a layout carefully calculated to minimize the number of steps between each action in cooking. The architect, Margarete Schütte-Lihotzky, took inspiration from the compact kitchens on trains or ships and assembly-line management ideas (very modernist). It was also politically motivated - she hoped that it would liberate women by allowing them to finish their household work quickly and have time for themselves.
(The podcast also discusses a related architecture-as-politics idea - designing homes with no kitchens of their own. The cooking would be done in communal kitchens, elevating it from unpaid "women's work" and reducing the amount of labor needed overall. Unlike the Frankfurt kitchen, this didn't catch on.)
And on the one hand, the political dreams didn't pan out. The labor that women saved mostly went into more household labor - now that you can cook really quickly, you're expected to make fancier meals, or spend more time on other chores like ironing. And there were also some issues with adapting to the new design - the isolated design made it harder to watch the kids, the electric appliances were a liability if you were too poor to pay the bills, etc. And this was in a public housing project, so if you had an issue with the layout you didn't much choice. (Once again, High Modernism packs everyone into evenly spaced rectangular grids!)
But on the other hand, the Frankfurt kitchen is the ancestor of all modern kitchens. The basic ergonomic design is still around, and everyone designs around electric appliances now. Architects treat the kitchen as the centerpiece of the home instead of writing it off as something for the women to deal with. And of course, women do indeed work outside the home these days, and labor-saving appliances are a part of that. So maybe give Modernism some credit for this one?
I'm having a hard time imagining that before the Frankfurt kitchen, the layout of kitchens was not designed to maximize the ease of cooking. Putting things together that are frequently used together is a basic organizational principle. What were kitchens of the past designed for, if not that?
Besides what others have brought up already, the big difference was a systematic effort to apply time-and-motion-study analysis to home kitchen design. Things had been optimized before, but through ad hoc creativity and analysis by individual householders rather than an attempt to design and popularize the One True Kitchen design. It's more or less the difference between learning to play a complex strategy game by everyone playing around with it, seeing what works, an occasionally copying someone else's strategies; versus someone making a bunch of spreadsheets based on the game mechanics and trying to work out optimal "meta" build orders and unit designs.
This article (https://www.apartmenttherapy.com/brief-history-of-the-kitchen-from-1900-to-1920-247461) shows what kitchens before the Frankfurt kitchen. It makes me think the Frankfurt kitchen was more of an incremental innovation, partly made possible by other recent innovations of the time.
I suspect it's less the layout itself, and more about furnishing the room specifically for cooking - having countertops instead of tables, and building those counters with dishracks and places to hold utensils, and so on. The Frankfurt kitchen was designed for a new public housing project, so they had the opportunity to mass-produce furnishings for a new style of kitchen.
The ancestor room / area of the ‘kitchen’ served many purposes, not only cooking.
Heating of the household, and managing the movement of fuel, air, and radiant heat, was an important role of traditional hearths and early ‘stubes’, as an example.
So you can imagine the optimal arrangement for some purposes might be different from the optimal arrangement for other purposes, such as water storage and access, or food preservation, or cooking a meal to be eaten the same day.
That does make sense, it didn't occur to me that what I think of as a kitchen didn't necessarily exist prior to the Frankfurt kitchen. Thank you (and also the other commenters) for the explanation.
There are other changes, both social and technological, that happened at the same time:
-Historically people who were wealthy enough to live in a house with a separate kitchen also had servants to cook for them. So the kitchen would be "below stairs" in the servants' area of the house.
-In a house where some but not all of it has running water, the kitchen needs to be somewhere that has it. If none of the house has running water, the kitchen needs to be somewhere where water can easily be brought.
-If solid fuels like coal are being used to cook, the kitchen will be somewhere that's close to where the fuel is stored. You also have considerations of exhaust, and possibly proximity to larders.
Do libertarians or other heavily market-oriented types have a more free market solution for providing social welfare programs? Right now retirement payments (Social Security) and healthcare for the poor & elderly (Medicare & Medicaid) are the largest parts of the US budget by a huge huge margin, and I would assume the same for other countries. The simple benefit of the private sector somehow providing these goods is simply saving the federal government a staggering amount of money.
I mean, we know how to have the government provide them- the Nordic/Western European model where income tax rates are 40-60% for the middle class, plus a VAT of 20+%. I don't have a huge philosophical objection to this, but it doesn't really seem like this is the final form of government that humanity has invented- I kind of doubt 'the government has to take half of everyone's paychecks forever' is like a Fukuyama-ish end of history and we will never invent a superior governance model. If it's the 25th century and we're zipping around to Alpha Centauri in a warp drive, will the government still have to take half of everyone's paychecks in order to provide the social welfare goods that citizens demand? Seems a bit unlikely?
So- what are the more market-oriented solutions? Forced 401k savings for everyone as opposed to Social Security? (Don't some countries like Singapore already do this?) Mandated private sector insurance program covering all citizens? I think in practice you do need a government mandate for people to purchase these, otherwise free loaders will refuse and then turn to the state when they need medical care- I don't think this is a solvable political problem otherwise
About two years ago, I wrote on the application of libertarianism to healthcare, on Quora. Much of that could apply to social welfare in general.
https://www.quora.com/What-is-the-libertarian-view-on-healthcare/answer/Paul-Brinkley-1
The main theme in that answer is arguably that whether your market is free or not, the libertarian notes that you will have to give people an incentive to want to provide services, for health, welfare, or otherwise. You will also want to give them an incentive to improve that service, either by raising the quality or lowering the price (or both), and since it's impossible to make anyone want to work for less money in a vacuum, you need another incentive, which turns up in the form of competition.
The problem with a government-forever system like you describe is that there's almost no incentive to improve. There's one provider, it gets paid regardless of how well it does, so its remaining incentive is to exert as little expenditure as possible to avoid revolution. Elections only matter if providers themselves are elected, and it's empirically harder to improve or maintain high service than it is to simply change the election rules to make their positions immune. So the entire system calcifies. All ailments treatable today are all there will ever be. Same for all financial troubles. Assistance is impersonal and austere; see the Department of Motor Vehicles in large cities for a model of what the optimized end will be.
For more specific libertarian-style reforms to US health care, I also wrote the following:
https://www.quora.com/As-a-libertarian-how-do-you-think-the-US-should-reform-its-health-care-system/answer/Paul-Brinkley-1
I like all of them, but 4 is the classic area where government regulation is needed. Providers deliberately don't publish prices now so that consumers can't compare them and price-shop. It's the same reason every B2B company doesn't list their prices on their website, you have to call & chat with their enterprise sales guy- they don't want to be commodified and have their prices driven down.
I mean- 'Price transparency makes the market more efficient - and less expensive'- you get why existing providers are not going to do that willingly, right? It requires government action
Why do any private providers provide prices?
If an outbreak of Teutonic Bronyvirus-K happens, Medicare announcing it will pay for the first $200 of treatment (or everything over $5000 or the middle $1000 on alternate Wednesdays or whatever) will do bugger-all if everyone with any expertise in immunology knows that it will cost more than that to develop any sort of vaccine against the Teuts.
Generally, anyone sick from something new, is screwed, because there's no incentive for anyone knowledgeable to cure it, by government decree.
Moreover, anyone who suspects a way to generate insulin for only $3/kilo won't bother to look into it any further if they know the government rate is higher and they'd just be leaving money on the table or competing with Uberpfizer. In other words, if it costs $X to fix your health problem, it will always cost $X. If a company knows how to do it for $X/2, they'll still charge you $X. And they won't tell you except in private, so that anyone else who might know how to fix it for $X/2 won't try, because they don't know what X is.
Generally, the current state of the art in medical will be frozen in time.
Heavily market oriented doesn't necessarily preclude government action in this area. Libertarians disagree with these government actions but for philosophical reason separate from their market orientation (though there is some overlap). Milton Friedman suggested a negative income tax as a replacement to many social programs. https://en.wikipedia.org/wiki/Negative_income_tax
Charity (including churches and non-religious groups), family, and community would be the answer to how to provide these services without any government action. Groups like the Salvation Army or St. Judes do this already
FWIW, NIT & UBI are functionally equivalent under some conditions.
I'm not so sure about that. NIT scales with income and is only available to people making below a certain level. Isn't the main point of UBI that everyone receives it and its the same amount for everyone? Those are two major factors that change a lot of dynamics of either proposal (not to mention that UBI would probably be taxed in the US while a NIT would be a tax refund, more or less, so after tax).
Combining an income tax where you pay $T(I) depending on your income and a UBI where you unconditionally receive $U is the same as a single, possibly-negative income tax of $T'(I) = $T(I) - $U
I would think the only real heavy hitters here are finding some way to reduce healthcare and housing expenses for the elderly (or ideally, for everyone). Regardless of how the payment is provided, as long as it remains expensive, you achieve only abstractly moral benefits over just taking 40-60% of everyone's paychecks. Maybe they can find some more voluntary means of getting that money, but it would still divert an enormous volume of the economy toward keeping people who are past working age healthy and off the streets.
I supposed the tech bro stereotype thing is replace all the doctors with GPT-15 and house people in Tokyo-style highrise cubbyholes with a hospital across the street.
We have great solutions to both of those problems already, we just reject them as unthinkable. For housing, we can tell people to move to cheaper places (which Libertarians may tell you is already what the market is telling people who cannot afford their current housing). Old people on Social Security will get much further in random rural places than big expensive cities.
Healthcare is even easier, but much more fraught. You run an analysis of life expectancy compared to cost of treatment and let people die if they are too expensive.
I don't advocate these options, but everyone knows they exist and they would be pretty effective. There's also no other way to lower either of those costs more than marginal amounts.
If you get rid of forced saving you need to be willing to let the people who fail at this actually fail.
And to tolerate the impacts of that. Mostly these days people are not interested in watching grandmas starve in the streets in their 60s even if the grandmas were extremely profligate and short sighted with their money for the 40 years before that.
So if we are going to be forced to take care of people then they need to have a certain level of responsibility forced on them.
To me it seems like the only libertarian solution is charity, i.e. consensual giving. A market-based approach would be to auction off those who are unable to take care of themselves as wards to the lowest bidder. Forced saving mostly just seems like a way to pretend that taxation isn't taxation.
Indentured servitude is very market based indeed!
Indentured servitude only works for people whose labour is worth more than the cost of keeping them alive. Welfare is for the people whose labour isn't. https://en.wikipedia.org/wiki/Fattigauktion describes how this worked in practice.
I'm trying to get more readers at https://codingwithintelligence.substack.com/
It's a weekly link aggregation of projects, papers, news, repos, demos and products in the "LLM space". Basically for Machine Learning Engineers or AI Engineers or Prompt Engineers however you'd like to call it.
Recently I've covered a lot of what's happening with Open Source language models that you can (mostly) run on consumer GPUs.
Why do I want more readers? The more people I reach the more I'm motivated to invest time & resources in it. I like the flywheel.
Anyone need a top-tier graduate student in either Economics or Computer Science? I have one to recommend! He's extremely smart and self-motivated. But he's unfortunately from a developing country, so he has extremely limited opportunities, so I'd like to get him into a program in the global north.
He's the kind of student of whom you can ask "I need to apply a Wald estimator to this data, but don't have time. Can you take a look and do it this week?"... and then he came back THREE HOURS later, saying, "I figured out how they work, and went ahead and wrote the code that applies it to your data files".
If interested, please ping me!
Why do attendants at full-service gas stations insist on topping off your gas tank? As long as I've been alive, the general recommendation has been that it is a bad thing to do, because it screws with internal portions of how the fuel system works in various ways that I don't understand. It also may be a waste of the extra gas. And yet even if I ask the gas attendant not to top it off, he usually will still do it about half the time, maybe just because they're used to doing it and it's a muscle memory thing.
My question is, why do they even do it at all? Every single official recommendation from all auto manufacturers is not to do this. So do the attendants do it because they think customers want it? Do they know that it messes with the fuel system? Are they just trying to squeeze out a few extra bucks out of each customer?
You should buy gas for a given amount of dollars. If you always fill your tank, you buy a set amount of gallons.
First, due to the cost averaging, you'll buy more of the cheap gas and less of the expensive gas.
Second, with a completely filled tank, you drive around some extra weight in gas, which will raise your consumption.
> Second, with a completely filled tank, you drive around some extra weight in gas, which will raise your consumption.
That's a very interesting take on it! However, I expect that, for myself at least, any minor cost saved by driving around with slightly less weight with fewer gallons would easily be outweighed by:
1. The amount of extra miles I'd be driving to and from the gas station more often
2. The monetary value of my time spent driving to and from the gas station
3. My own stress about having to fill up more often and procrastinating on it
I am curious, though, if you have any figures for how much gas consumption might increase per pound that the car carries around.
You're right... apparently 100 kg of additional weight can cost up to 0.3 liters for 100 km. https://www.adac.de/verkehr/tanken-kraftstoff-antrieb/tipps-zum-tanken/sprit-sparen-tipps/ So the 20 or 30 kg of gas that you drive around won't do much of a difference. Especially if you need to drive a detour to the gas station.
But nevertheless, if filling the tank always costs you e.g. something between 50$ and 60$, you should always ask for 50$ of gas. https://en.wikipedia.org/wiki/Dollar_cost_averaging
When I fill my tank, I wait 15-30 seconds after the pump auto-stops, then start it again. I can get another .2-.5 gallons in there. This is due to air escaping from the tank. I guess you can call it topping off but I don't think there is risk of damage to the system. I have a 14 gallon tank and if i drive it until the estimated range is below 10 mi, I usually only get ~13 gallons in before the pump stops itself.
Boy, in the US full service is pretty rare since, oh, 1980 or so. They just suddenly started making most of us pump our own gas. It was a pretty abrupt change.
So here is the last appearance of a full service gas station in a major American movie. Dan Ackroyd offers to wash the dead bugs off Twiggy’s windshield in the Blues Brothers 1980. (don’t worry it’s just a 24 second clip)
https://m.youtube.com/watch?v=sy3qvBPr5c4
Still mandatory in Oregon. I think? https://www.opb.org/article/2023/06/30/oregon-self-serve-gas-pump-law-governor-tina-kotek/
They just changed/repealed the law, right?
Unless something's happened since I last heard, I'm pretty sure they only relaxed it for rural areas at night, to make it easier to stay open (having struggled to find an open station driving through Clatskanie at 11pm with the needle kissing E, I'm in favor of this).
Too bad – so ridiculous
From Zip recruiter: "As of Jul 2, 2023, the average hourly pay for a Gas Station Attendant in New Jersey is $15.76 an hour. While ZipRecruiter is seeing salaries as high as $20.16 and as low as $8.85, the majority of Gas Station Attendant salaries currently range between $14.76 (25th percentile) to $18.46 (75th percentile) in New Jersey."
Why do they do it?
Because likely they are general poorly trained and selected from the bottom of the barrel.
At least where I live in NJ, gas station employees are mostly Sikhs, so you've got a very restricted labor pool on top of that.
There aren’t any where I live. Nevertheless I assume it’s a useful income for some people, students perhaps.
I could be wrong, but I thought topping off your gas was only bad if you did it poorly, like extra gas leaked out from the full sensor not picking that up because the nozzle was inserted incorrectly, or something. I thought it had nothing to do with how the car runs, and I don't see how it could.
Of course, it is assumed professional full-service attendants would not have this problem, and you can go a little bit farther without needing a fill-up.
Is there really such a thing as a "professional full-service attendants", with emphasis on "professional"?
They've almost certainly pumped gas more times in the past week than most of us do in a year. If there is *any* amount of practice that helps with doing a feature of the job, they've got more of it. It's only if you think that there is *no* way that practice could improve that you wouldn't call them "professional" in the relevant sense.
Practice makes permanent not perfect.
All work has dignity but profession and professional have a meaning and the meaning is not synonymous with being employed even for a long time at the same job.
Surely no one working above a minimum-wage job believes that all work has dignity. The other day I watched a Target employee clean up some piss from an "emotional support" chihuahua a customer brought in. I'd love to watch a Gandhian explain the dignity in that to me.
"Professional" indicates being paid to do something, so yes. It implies, but doesn't guarantee, expertise, since they do it a lot.
Are there volunteer "full-service" attendants? Being paid make you a "professional"?
I guess you're asking questions rhetorically, since you seem to have already rejected any answer other than your own definition. One definition of "professional" is someone with a higher degree, like a doctor, lawyer, engineer, architect, economist, etc. Another definition is simply "someone who gets paid to do that thing." If you make music to entertain yourself and your friends, you're an amateur musician. If you land enough paying gigs to support yourself, you're a professional musician. If you put tile just in your own bathroom, you're a do-it-yourselfer. If your job is going around tiling other people's houses, you're a professional. I'm not sure why you're quibbling with simply narrower or wider definitions of the same word. Tea can mean any beverage brewed with water and some plant material, or tea can specifically refer to Camilla senensis. It's not super complicated.
My own definition?
I pump my own gas, I'm an amateur gas pumper; but the the person who pumps in for me is a professional gas pumper?
I'm laughing out loud at the absurdity.
The person who flips burgers at McDonald's is "professional" burger flipper?
https://www.merriam-webster.com/dictionary/professional
All three definitions 2a, 2b, and 2c apply.
I cannot speak to volunteer attendants, but I would guess...very rarely? Amateur full-service attendants?
No they do not apply.
It does have to do with how the car runs to some degree, and the worry is not just about extra gas spilling out. I've heard several explanations, but I don't know much about cars so I don't really understand any of them, or know how much extra gas would cause the damage. But there's are concerns about the extra gas causing a cracked tank, or getting into the car's charcoal canister or carbon filter, which usually are only supposed to get gas vapors, not actual gas. Supposedly this can cause engine damage.
I'm no car person, but in order to crack the tank or something the gas should have be be under pressure, and I don't see how that could be with how the nozzle inserts into the tank. There's some kind of metal seal, but I don't know how tight that seal is, or why the gas cap would need a tight seal if it is already well-sealed, and in any case, I think the gas would equalize the pressure as the nozzle is removed.
Note I'm explaining how I don't know, not claiming expertise. I would be grateful for an explanation proving me wrong.
I'm just explaining what I've seen people say online. I can't speak to anything's veracity, because I don't understand cars. I'll bet a cracked tank is rare but does happen in certain circumstances, but then why even take the chance?
Also what about the carbon filter or charcoal canister issue?
I'm kind of a car person, and I do understand cars. At least I've restored some cars, trucks and tractors from the ground up. I have lots of friends who are fulltime mechanics currently working in repair garages. I asked them, and none of them has ever heard of a "cracked tank" or other malady from completely filling the tank.
It's like a lot of things with a claimed potential for harm, but the actual evidence for genuine damage is practically non-existent. Like your cell phone harming a jet's avionics. People top off their tanks so frequently, often just to round off the change they have coming back, or to get to the exact amount they pre-paid for, to keep from having to walk back into the convenience store and wait for their change, that if there really were significant problems from a few extra squirts of gasoline, where are all the damaged cars? If this was a real thing, there would be real cars brought into repair shops every week all over the country. There aren't.
There's got be more significant things to worry about than this hypothetical issue.
Okay, forget cracked tank, maybe that one's not real, I haven't seen it cited too much, so I could believe it's fake. But Google for "should you top off gas tank", and almost without exception, every site says no, you shouldn't. Most commonly, they seem to indicate that it can irreparably damage your car's "vapor recovery system", and possibly also cause pollution.
Back in the day, the attenants tended to try to round the price up to the nearest dollar or at least quarter.
Given that I can't remember the last time I paid cash for a fill-up, I see no benefit in them doing so.
It's moot for me - if at all possible, I choose a self-serve pump.
Are there even any full-service stations in the US outside of New Jersey?
Oregon also requires all stations to be full-service only.
Not for long though! I think they voted it out this year.
No to full service gas stations. Yes to shrooms.
Haha yes. NJ mandates them, but they exist as optional all over New England and the tri state area, at least.
Affirmative action: this seems really simple to me— create a race-blind application process that records SES information. Take every decile of the SES range and make a pool of the top 10% of students from that decile. Take an equal random sampling of students from each pool. I bet you would end up with a pretty darn racially diverse student body, with the bonus of also having economic dIversity. If this isn’t “elite” enough, make your pool from the top 1% of the first SES decile, top 2% of the second SES decile, etc.
Even simpler: selective admission policies are obviously more aligned with the institution's interest than the public interest. Institutions that use such policies have self-selected into the Hillsdale College path of no federal funding. Adult education often does scale with technology, and the default policy should be admission of all qualified applicants (where 'qualified' = 'demonstrates the prerequisite skills to succeed at the program'). Or where physical constraints prevent open admission of qualified applicants, a lottery of qualified applicants is an acceptable fallback.
The most likely outcome of an SES schema like you describe is that all of the best white and Asian students dominate at every decile. Black and Hispanic students will be heavily reduced. There are lots of poor white and Asian people, many of whom are higher-achieving than black people at the same SES. We see consistently that white test scores outpace black test scores at the same SES levels (and several levels above), such that rich black people have similar or worse testing scores to poor white people who in turn far outpace poor black people.
I'm in favor of race-blind applications, but those interested in AA will never accept it because it would mean even fewer black applicants get into elite colleges.
I was under the impression a fair number of states already did this, guaranteeing admission for certain state schools to anyone who is in the top 10% (or possibly some other percent) of their own school's graduating class. Assuming here that, for the most part, people in the same school have similar SES and this divides students in roughly the same way.
Some do, and also grant guaranteed scholarships to the top x% of each school, but this may be more common in states where the state university is average or below, not for the top ones. New Mexico does this, for instance, but UNM is truly a state college, for people in the state to learn things, not very useful for signaling, or drawing many people from different states or countries.
<mild snark>
Rather than optimizing for economic diversity, how about ideological diversity? Rather than splitting up the population by economic deciles, how about having all the candidates be classified by one of the 2D political ideology measures? Then split that 2D space into bins, and pick from the bins in proportion to the fraction of the overall population in each bin (with the same top 10% from each bin).
</mild snark>
Because the desire the universities have expressed is to achieve racial diversity, but they cannot achieve this outcome by discriminating on the basis of race. But, since, as I understand it, family wealth/SES is heavily correlated with race, using that as a proxy could perhaps achieve similar results.
"family wealth/SES is heavily correlated with race, using that as a proxy could perhaps achieve similar results"
Yes, I've read of that too. I'd characterize using a proxy in this way as underhanded.
What do you mean by underhanded?
The usual meaning "acting or done in a secret or dishonest way".
In this specific case, choosing a proxy for race, so that the university can continue to racially discriminate against white and asians and in favor of blacks, SCOTUS or no SCOTUS, while concealing this discrimination under a proxy which purports to be racially neutral but is actually chosen precisely because it is racist.
It would add a unique perspective to the mixed views in their class. :-)
I saw someone arguing recently that IQ for whites earning below the poverty line (or similar) is higher than blacks earning 5-10 times as much.
If this is true then I think your system might select for even less black students at medium to high selection schools.
The problems with black educational attainment go much deeper than absolute family income.
Punishing children for the hard work of their parents sounds awfully unfair to me, as does rewarding children for having parents who know how to game the SES calculation, and I think a lot of people would agree with that assessment.
I mean, so does allowing children to benefit from the hard work of their parents. If you believe in equality of opportunity, then students shouldn't be able to have their parents buy tutoring services.
I disagree. I think parents supporting their children is entirely fair, though I don't think it has that much of an effect, and it's presumably in the interest of schools to have entrance exams that measure underlying ability rather than the amount of tutoring received.
There are lots of easy solutions. The problem is people are wedded to the current bad solutions so they are not interested in different ones.
Also lets say you do what you suggested and in the next class there are 0 Latinos lets say. You program gets stopped immediately because academia is 60% is racists these days who only see skin color and don't care about the people underneath.
That seems interesting to see you use the term "racist" for someone who sees the complete absence of a major group and assumes that this is prima facie evidence of a hidden unfairness. Are you really so convinced of the absence of any possible unfairness that you want to use the word "racist" for this?
If you just mean to use the word "racist" in a morally neutral way, then maybe that's fine. But it should then also be fine to use the word "racist" for the kind of structural injustice that these people want to talk about.
I am using "racist" for a group of people who judge others based on skin color uber alles and fail to consider people's individual backgrounds or any other pragmatic considerations besides fractious counter productive identity based nonsense.
Universities are the late medieval Catholic Church/monasteries of today, and the sooner the institutions can be torn down and reformed the better. They fail at both of their main goals (educating citizens and conducting research).
For the amount of money plowed into the system there is incredibly little return. Its too bad the MOOC model didn't work, as that would have been a huge boon to society.
Anyway when a group of people makes it clear they are the enemy to you, to truth, and to fairness, it is best to treat them as such.
I'm sure there is a social scientist who can explain that there is no "person underneath." After all, people only exist at intersections of oppression axes, n'est-ce pas?
Even if this worked as intended, it would still create a huge black underrepresentation (especially among descendents from slaves).
But it wouldn't work since the rich and educated would game the system. "Record SES information" is not a simple atomic task. What's the SES of a person who's parents used to be high-earning doctors but that are currently working for minimum wage at a local nonprofit? What's the SES of someone who claims to be disowned by their rich parents and that are receiving the dole living in the poorest neighborhood?
My guess is that those are pretty rare edge cases. Most parents who have money plan to use it to help their kids pay for college. I mean there could also be some poor kids whose uncles have won the lottery and plan to pay for their college. I think that the marginal impact of these scenarios is pretty small.
They are rare edge cases right now, but if doing a year of volunteer work increased your kids chance of getting into Harvard by 860%, it would quickly become very popular. There would be industries of consultants who design the best strategies for upper- and middle-class parents to minimize their SES score (just like how there's already an industry like this for our current university admission system). This is blatantly obvious to me, I'm a bit perplexed that you don't seem to get it.
This is already a thing in Italy where universities give free accommodation and/or free cafeteria access and waive tuition to the poorest segment of the student population. Anecdotally, rich parents who care enough about this manage to temporarily hide their wealth through accounting tricks. It is however generally considered that this is just a price society has to pay to support the students who are the intended target of these measures, and the fact that there are some free riders is not perceived as a big deal (we are not talking admission here).
Because for college admissions purposes, SES is referring to SES of the applicant’s parents.
"SES" is not a well-defined thing. Neither is "SES of the applicant’s parents". You will have to set up an algorithm calculate the "SES score" of someone. You might use parental income or capital or postal code or whatever, but you need something concrete. Whatever algorithm you set up, people will game it, and since SES is very mutable it's likely that the algorithm will be easy to game. Once again, I'm perplexed that you don't see this.
If you specify how the SES score of an applicant will be calculated, I'm very confident that I can find way for rich people to easily game it.
Once again, what's the SES of a person who's parents used to be high-earning doctors but that are currently working for minimum wage at a local nonprofit? What's the SES of someone who claims to be disowned by their rich parents and that are receiving the dole living in the poorest neighborhood? Don't you think applicants will manufacture these circumstances if it increases their acceptance rate by x10?
Where is scholarly aptitude in this? Aren't you assuming everyone is equally qualified, and only divided by SES (social-economic status, unless my Googling came up with the wrong meaning)?
Scholarly aptitude is how you create the pools of top students from each SES decile. Accept a random sample of students from the top X percentile (in scholarly aptitude and any other combination of desirable factor) of each SES decile.
Isn't it true that poorer people have less opportunity to learn? It may be that the top 1% if the lowest bracket has less aptitude for academic success than everyone in the top bracket, simply from never having been exposed to the curriculum before.
I think this is a difficult problem to solve, for many factors enter into it. No one is actively prevented from a good education, AFAIK, but still some don't get the opportunity.
If there is less opportunity to learn, that would mean that the student is less academically sound and capable, and therefore wouldn't belong in a higher education institute in the first place, *unless their test scores show otherwise.* Maybe... just maybe... we should select students based on test scores! How radical.
What test scores? It has been said that if an aborigine wrote an IQ test everyone in civilization would flunk it.
How is someone supposed to get a high test score when they haven't the opportunity to learn and practice what might be on it?
"It has been said that if an aborigine wrote an IQ test everyone in civilization would flunk it. "
First of all, they wouldn't. We certainly have more general knowledge than an aborigine. But even if we did, it's because the test would be designed for a world that we don't live in, and certainly not a world that secondary education would prepare you for.
"How is someone supposed to get a high test score when they haven't the opportunity to learn and practice what might be on it? "
They aren't expected to. If someone doesn't have a solid primary education, they don't belong in a secondary institution, until they develop that knowledge.
I went to the Philly Rat Fest on a whim and had a good time. I didn't know anything about critical rationality (based in Popper and Deutsch) nor that it was critical rationalists rather than a gathering for rationalists in general. Any recommendations for critical rationalist online places or sources in addition to Deitch? I like the optimistic attitude and the way human experience gets centered. I don't think Bayesians (a fair name for rationalists who aren't critical rationalists?) completely ignore human experience, but they can get very abstract. Those questions of whether humans are even a good idea....
One of the talks was about any early image breeding program. Does anyone remember the name? It offered random shadings, and people could "breed" them to make interesting images. The point of the talk was that sometimes exploration according to personal taste can pay off better than deliberate problem-solving.
I've started reading _The Fabric of Reality_ I'm just on the chapter about about explanations being of central importance and predictions being no substitute for explanations. I'm not sure that Deutsch's thought experiment of a perfect oracle compared to having explanations is fair. I think his point is that If you had a perfect oracle, you still couldn't use it alone to build a spaceship unless you understood enough to ask the right questions.
I do think that wanting (settling for?) predictions without explanations might be part of a pattern of trying to make conscious thought unnecessary.
I'm not sure whether Deutsch makes this explicit, but if all you have is a prediction, you have to guess about how much scope it covers.
I do have a real world example of what can happen if you just want predictions. If someone gave you three great stock tips or predictions of horse race winners, what would you pay for a fourth?
In this context, you might see it coming. This is a scam. The scammer sends out different sets of predictions to different marks. Some of the sets of predictions will be correct. The mark is willing to believe in a secret system.
Deutsch is heavily influenced by Popper. They both believe that scientific progress goes like this: first conjecture, then refutation. Deutsch rejects "instrumentalism", the claim that all that matters in science is the prediction of experimental outcomes. He believe, then, that what "matters" is the explanation (a conjecture), the experiment being a means to reject bad explanations (refutation).
So, yes, if an alien species gave us a perfect oracle, for example a computer that predicts the outcome of any experiment we can describe, we would not stop doing physics (or science in general) because we would still wish to understand why things work the way they work.
The instrumentalist view is surprisingly common, even from people unaware of these distinctions. It permeates the Copenhagen Interpretation of Quantum Mechanics, for example.
Deutsch rejects "instrumentalism", the claim that all that matters in science is the prediction of experimental outcomes.
Deutsch also rejects empiricism, prediction and induction . That makes it pretty difficult to o combine his approach with Yudkowsky 's.
This might only be tangentially related but: https://eprints.lse.ac.uk/100335/1/Dual_Inheritance_Theory_Perspective_final.pdf
is an argument that an underlying problem in the replication crisis is that a lot of science doesn't have an underlying theory and therefor can't be properly tested. It's a similar to the complaint (I think?) you're making that a sufficient number of people making sufficient predictions will eventually predict the entire works of Shakespeare.
My dad has a wrist injury where he can’t put any compression on it for the next few months, but he sorely misses weightlifting and is going a bit stir crazy now that he can’t do it. Does anyone know of any devices that you can put on the forearm which can have dumbbells or barbells attached to it?
Many of the exercise machines at gyms allow you to exercise your upper arm, shoulders, upper back etc without grasping anything in your hands, or even using your forearms. I think it's probably also possible to substitute exercises with elastic bands for weight lift exercises, adapting the arms exercises by connecting the band to the area right above the elbow, rather than holding the end on one's hand.
When I shattered my forearm years ago, I shifted all of my weight lifting to core and lower body. Anything that could be done without the broken arm like machine squats, leg presses, back extensions, etc. My doctor claimed I had the most successful recovery he had ever seen for such an injury. IMO trying to workout the affected arm seems like a bad idea. Let it heal and do whatever exercise you can elsewhere.
Forearm weights? https://fitnessvolt.com/forearm-weights/#:~:text=Forearm%20weights%20are%20a%20modified,any%20time%20during%20the%20day.
No personal experience. My wrist weights only go up to like half a pound. But unless you can fit your forearm into a cable handle it's the best I can think of for bicep curls.
Bro squats, deadlifts, pullups, high pulls, side bends... there a lot of lifts that don't involve wrist compression, you know?
Wouldn't the solution be to find him something else to get interested in? It's a just a few months. Surely, there is muscle between his ears which he could also work on. Or the heart muscle?
I don't know about Gruffydd's dad but I find that sudden changes in the amount of exercise I get tend to be quite physically unpleasant. It's not an unreasonable request.
Really? You hurt your wrist you supposed to cool it with the weightlifting temporarily. I think trying to find a work around to continue weightlifting is pretty unreasonable.
"sudden changes in the amount of exercise I get tend to be quite physically unpleasant" Well if the change is to do a lot more exercise that could be physically unpleasant. To do less? Physically unpleasant? (Maybe mentally or psychologically unpleasant if you weirdly conditioned yourself.)
I am recalling my father in the 70s seeing an adult jogger and remarking: he obviously isn't working hard enough.
Finding alternative exercises that don't use the wrist doesn't seem like it wouldn't actually have much risk of worsening the injury if it's not even involved, so I don't see why that's unreasonable.
Yes, doing less is physically unpleasant. This is probably restless leg syndrome and I don't know how well it generalises.
He still wants to lift weights with his arms. That is unreasonable if you have a wrist injury.
RLS, not mentioned in connection with wrist injury, being idiopathic would not be generalizable to anybody including someone else with an RLS diagnosis.
Since I basically only have my own experience to go on, I assumed that most people would get at least some symptoms upon discontinuing an exercise routine even if it's not the same sort, but if that's not the case, fair enough.
When it comes to weight lifting, you can lose a LOT of progress in "just" a few months.
Is the where SMH would be appropriately used by people between 25 and 55?
Possibly. What meaning are you trying to convey?
I exercise using elastic bands. Might that work?
Is it true that kids today have worse fine motor skills than what kids had say 15 years ago? (Presumably this is because they spend more time today with digital devices.) I've seen this claim thrown around allot but I haven't seen any studies to back it up and I can't find any when I Google.
As some others have said, fine motor skills are very task-specific.
I was in the Scouts, and could tie a knot or make an eel spear. However, the first time I ever used a computer mouse was as a teenager, when a couple of friends and I played the Macintosh game "Dark Castle" (1986). We were terrible at it, so much so that in order to make any progress, we had to split the controls so that one person used the arrow keys (like in a "normal" game at the time) and another could put their full attention on aiming with the mouse.
My kids have grown up using WASD and a mouse, so if they ever experienced something like the above, they were probably too young to remember.
I don't know about studies, but yes, it probably depends on the specifics of what they mean.
Sometimes I attend Elementary art teacher meetings, where the older teachers will talk about how the current generation of little kids don't really know how to use things like scissors, tape, glue, paintbrushes, threading, twisting, play dough, and so on. Some of them may sometimes shorthand this to "fine motor skills," but they mean a very specific subset of those fine motor skills. It's no help to them if kids are way better at operating devices with their fingers, but struggle terribly with tying a knot. Other people would care about different things -- maybe typing vs handwriting or something. I would expect typing to be better, handwriting worse, but am not certain.
I think motors skills are pretty diverse. Kids might be worse at writing, better at working a mouse or touch screen, and way worse at throwing/catching balls.
What does that say about their "fine motor skills"?
So suppose we found a definitive answer to this question. What then?
Institute some commission to monitor the issue and publish yearly reports. Profit.
This Finnish study found no significant change 1983-2018: https://pubmed.ncbi.nlm.nih.gov/37050948/
Note that they examined kids and teens aged 7-18, so the study might have ended too soon to catch the full force of any "ipad at 2"-type effects.
Probably depends on precisely what you measure. I think videogames on net improve fine motor skills, but 15 years is a short enough time that one is looking more at how videogame control schemes have evolved than at the shift from playing outside to playing video games; I could believe that touch screens develop motor skills less than controllers
I thought I'd mention one of my recent podcasts!
I have been fascinated by the story of Anson's voyage around the world back in the 1740's for years. It was a British attack on Spain's Pacific coast possessions in South America during the absurdly named War of Jenkins' Ear. Eight ships were sent round the Horn into the Pacific. But the weather they faced was truly horrendous and two turned back and one was wrecked. The Spanish squadron sent to stop them was basically destroyed.
My podcast with David Grann focuses on the Wager perhaps the unluckiest ship in the squadron (it was a competitive field!). At the time scurvy was not really understood and the death rate was appalling - both going round the Horn and for the surviving ships as they crossed the Pacific. I believe one of the worst outbreaks of scurvy in naval history.
It didn't help that a shortage of manpower had led to the British making up numbers by putting on board retired and invalid military personnel. Pretty much all of these died before they got round the Horn.
As I say Grann and the podcast focuses on the wreck of the Wager - the subtitle of the book is Shipwreck, murder and mutiny which gives you an idea of what happened to them there. But the story of their various attempts to escape from the freezing, deserted island they were trapped on is the stuff of movies (Scorcese has bought the rights and has just turned another of Grann's books into a movie so may well happen).
The rest of the squadron suffered appalling losses and by the end only a single ship was made it back with some 256 crew. But astonishingly they had succeeded in capturing the Manilla galleon. The treasure was about $80m in today's money and the crew got half. The Manilla galleon itself is a fascinating thing with one voyage once a year from South America to Manilla with silver and other treasure and another ship going the other way with trade goods from China. The Manilla galleons sailed continuously for about 250 years. The British dreamed of capturing them but succeeded only (I think?) four times. Here are my two podcasts (on the Wager and on the Manilla galleons) and David's book:
https://www.buzzsprout.com/207869/13167893
https://www.buzzsprout.com/207869/10198722
https://tinyurl.com/bdz7nknv
So when Anson returned with the treasure everyone thought it was a huge success but one newspaper printed a poem casting a more sceptical eye . . .
Deluded Briton! Wherefore should you boast
Of treasure, purchased at a treble cost?
To purchase this, think how much treasure’s gone;
Think on the mighty mischiefs it hath done.
In this attempt, count o’er the numerous host
Of Albion’s sons, unprofitably lost.
Then will your boastings into sorrow turn.
And injured Britons, Albion’s fate shall mourn.
true stories are the best!
Ignoring the possibility of future nuclear war, has the existence of nuclear weapons so far been net good or bad?
It's hard to imagine how, if you discount the possibility of future nuclear war, the existence of nuclear weapons has not been an extraordinary boon to humanity. We haven't fought a World War in almost eighty years, or even a direct Great Power conflict of any sort. This in spite of half a century of Cold War between two superpowers, of the sort that would certainly have fought a Hot War in any prenuclear era of human history. The cost of establishing and maintaining the world's nuclear arsenals to date, is trivial compared to the cost of even a conventional World War.
So, yes, ignoring the only major down side of the whole thing, it's been a good thing.
Do you mean, do the current accrued benefits outweigh the current accrued costs?
Since you're eliminating the largest negative, it must be net good. Without nuclear weapons we wouldn't have nuclear medicine, I think, or other, possibly less obviously-related technologies.
But I'm not sure the question makes sense. Kind of like, "If knives couldn't be used to kill people, are there any downsides to them?"
I find their existence is net bad, since it creates the possibility of global nuclear war in the first place, which would be rather catastrophic for much of humanity.
- Any real or perceived "gap" in nuclear weapons power inherently leads to mutual buildup, reinforcing itself in a positive feedback loop. This loop can only be stopped by disarmament and inspection agreements, which are subject to revision by whoever is currently in power in any of the signatories.
- The lesson learned by every two-bit dictator around the world is that if you have even one nuke able to reach your main rival, you become basically untouchable, so there is a strong incentive for proliferation.
- The decision whether or not to use strategic nuclear weapons is heavily informed by technology and intelligence, both of which can fail and create a situation of accidental nuclear war, which is in effect indistinguishable from intentional nuclear war.
- On the plus side, they reduce the chance of war between the great powers, but industrialization is already doing that by reducing the incentive for large-scale war in general.
Nuclear weapons, or the nuclear industry in General?
Looking only at Nuclear Weapons, the only good they can do is to deter violence.
Since the only Nuclear weapons used solar were. in Japan in WW2, and caused less damage than the Tokyo regular bombing (Not a historian), and my impression is that they hurried Japan's surrender (but by how much?) You could argue that Nuclear Weapons stopped Violence by the Sheer promise of even greater violence.
The Whole MAD scenario of the Cold War seems to support this, though I am not convinced that with the Alternative history of no nuclear weapons we would have a more direct NATO-Soviet conflict.
I think averting WW3 is really obviously very good, and almost all the downside is in the risk of a future nuclear war. I do think the USSR and the USA would have ended up triggering WW3 absent MAD, too
Also on the positive side of the trade-off: Nuclear weapons made ICBMs make sense, and ICBMs were precursors to access to orbit and deep space. Yes, Goddard and Peenemünde preceded Trinity site, but, in the absence of nukes, I think it would have been a _long_ time before we got weather satellites and spy satellites and GPS (and human spaceflight, but the benefits of that thus far are much more debatable). I strongly suspect that the lives saved by better weather predictions alone probably outnumber the dead of Hiroshima and Nagasaki.
There are weaknesses with the Intelligent Design argument in biology, but those weaknesses don't exist with Fine Tuning argument in physics. For example:
1. There are no fossils in physics.
2. Physics is fundamental while biology is derived from chemistry.
3. Scientists have never observed other universes while they have observed other planets.
If you're looking for the strongest rational convincing argument for the existence of God from science, look to physics. Listen to the episode of Physics vs Biology at https://www.physicstogod.com/podcast or you can find "Physics to God" on any podcast platform.
My favorite 'meta-GUT' is that ALL possible universes exist, so God is entirely superfluous – of course we'd find ourselves (whomever 'we' might be) in a universe 'fine-tuned' for our existence!
Somebody fine-tuned mine by putting beaches right next to oceans. The knew I'd dig that.
That's Max Tegmark's Mathematical Universe Hypothesis. We discuss that in the multiverse miniseries, but it's a very flawed theory which to some degree undermines all of science.
'Undermining all of science' seems like a fine thing for a (meta-)GUT to do!
Nah – close, but I'm thinking of Stephen Wolfram's "Ruliad" specifically: https://writings.stephenwolfram.com/2021/11/the-concept-of-the-ruliad/
The need for fine tuning is usually taken to show that theories are inadequate or incomplete. Theories often happen to have these issues. I wouldn’t want to rely on these characteristics of fallible human theories to prove the existence or to constrain the nature of God.
We're not making an argument of God of the Gaps. Listen to episode 3 and 5 of our podcast "Physics to God" for a detailed explanation of this point.
Why does it matter? There are plenty of reasons to believe human engineers may one day be able to create universes if our species makes it long enough. That would make us God by this definition, but doesn't make us worthy of worship, has zero consequences for how inhabitants of these universes should behave, doesn't mean we're imbuing them with immortal souls that will survive the death of their physical bodies, or any of the other reasons believers in God typically actually care that there might be a God. We wouldn't even be able to observe what happens in these universes and wouldn't know if life ever developed. If we had a creator, but one who was not infallible, omniscient, omnipresent, or omnipotent, and likely became causally disconnected from our universe within picoseconds of its creation, what is this supposed to mean for us?
There are many reasons why God matters, separate and aside from religious issues of worship. First and foremost, is about truth about reality. Is the ultimate fundamental reality an intelligent agent or is it an unintelligent law(s) that generates an infinite multiverse? Truth about reality has mattered to many people throughout the ages.
It's 2023 and some people still believe in Sky Daddy, or that their souls will one day live forever in a beautiful place, or that some kind of sentient, intelligent design created the universe and the Earth, where 99% of all creatures that have ever lived are extinct? Hahahaha! Sad.
We don't discuss religion on the podcast. It's about the proper philosophical idea of God to infer from the argument based on science.
"Why does it matter? There are plenty of reasons to believe human engineers may one day be able to create universes if our species makes it long enough."
I can't think of any reason to believe that, other than "Humans seem to get better at making things over time".
This does not deal with the most important piece of evidence from fine tuning.
We observe that there are a handful of physical constants that would have to be within some pretty narrow ranges for this universe to look the way it does, namely, with us in it.
There are two hypotheses to explain the apparent coincidence: a creator, or anthropics.
If the answer was anthropics (that is, for us to make the observation we must be in a universe with parameters in certain ranges, and no observations can be made in universes that don't) then we should expect the minimal amount of fine tuning necessary for us to be able to exist and make predictions.
If the answer is a creator, then we should expect omnipresent fine tuning according to precisely whatever the goals or values of the creator are. Not just a handful of parameters, but all of them, and everything else. The more the universe looks obviously tuned and designed, the more evidence for a creator.
Therefore, minimal fine tuning is evidence for anthropics, and maximal fine tuning is evidence for the creator.
We look at the universe and see that is extremely minimally fine tuned. The only way to say this is evidence for a creator is to say that the creator just so happens to want exactly the same minimal fine tuning that anthropics would end up in. And even so this gives it precisely zero explanatory power _more_ than the anthropics hypothesis. This is worse in terms of Occam's razor and is obviously motivated reasoning.
>There are two hypotheses to explain the apparent coincidence: a creator, or anthropics.
There's also a third hypothesis: deeper order. Under current established physics models (General Relativity and the Quantum Standard Model), the constants in question appear to be free variables that could be anything but happen to be fine-tuned to allow the existence of atoms, molecules, stars, planets, and complex planetary life. But we know those models are incomplete, and we don't know for sure what's really going on under the hood. Maybe there's some deeper relationship between the now-seemingly-independent physical constants that explains what now appears to be fine tuning.
Agreed, there's always the possibility the apparent coincidence is no coincidence at all. But with the info we have now, and to grapple fairly with their argument, we can exclude it.
"We look at the universe and see that is extremely minimally fine tuned."
Nice! I was aware that this is the feature one wants to test, and I was aware that in the case of the Higgs field value, it indeed looks minimally fine tuned (in fact looks likely to be metastable). Is there a url that documents the minimal fine tuning of the other Standard Model parameters? Many Thanks!
You are right up until a certain point. When we look at the universe we find that there is way more order than is necessary for an anthropic interpretation. Richard Feynman made this point and it goes back to Arthur Eddington. The minimal amount of order needed to explain our observations is a single brain with all your (false) memories surrounded by a universe of total disorder. But in fact we find that the universe is filled with order, complexity, and structure everywhere in it we look. This is called the Boltzmann Brain problem. We'll discuss that in our series on the multiverse.
This is resolved by a proper application of Occam's razor, or Salomonoff Induction. An ordered brain in a disordered universe is actually more apparent coincidence than simply a slightly ordered universe which begets a slightly ordered brain. The "total amount of disordered stuff" is higher in the disordered-universe-with-ordered-brain, but the simplest (shortest) description is of an ordered universe with a similarly ordered brain. The boltzmann brain problem is primarily used as a tool for comparing and assessing different specific cosmological theories. It certainly cannot be interpreted on its own terms to find a creator more likely than the anthropic interpretation, and the actual expert cosmologists certainly wouldnt agree with it being used that way. Theres no creator in the boltzmann universe either.
And, frankly, the universe does NOT appear to be filled with order and structure everywhere we look. The universe appears frighteningly disordered with tiny tendrils of coincidental order. I cannot stress this enough - the idea that the universe is actually largely ordered is in total contradiction with what any ACTUALLY ordered thing looks like. The universe is almost entirely empty space. Of the mass it contains, the vast majority doesn't even interact with light. At the same time that it contains a maximum speed, it also is expanding from within such that given time the universe will become filled with little pockets that slip entirely out of each others light cone and so can never interact whatsoever forever. These are totally bizarre qualities to suppose a creator would deliberately design for any purpose. They certainly do not appear ordered.
There's the apparent coincidence of the well-tuned constants, there's the still incomplete mathematical descriptions that reality seems to follow, and there is the tendency for competitive spaces to become filled by ever more competitive entities (which Scott has helpfully named Moloch), none of which especially suggest a creator over random chance. I have not seen, and you have not presented (though I admit I have not gone through ALL or your rather prolific work) a fundamentally different kind of order than these.
I think you are mistaken. The second law of thermodynamics says that all physical processes bring a system from a low state of entropy (less probable) to a higher state of entropy. Therefore, an ordered universe which begets a single brain must logically have lower entropy than the single brain itself. It is therefore more likely to get a single brain by chance alone than an ordered universe which begets a brain.
With regards to your second point that you don't think the universe has enough order, complexity, and structure - I understand what's bothering you. On the other hand, it's hard to play God and decide just what the most perfect, beautiful, ordered universe should be like. However, it seems very reasonable to say that our universe is amazing and incredible with wisdom everywhere on all orders of magnitudes - from atoms and molecules, to planets, stars, galaxies, and life. It is certainly much more interesting than a universe filled with black holes which is the most likely state of the universe (a state of maximum entropy). That point is all we need for our argument to work.
Theres a mistake in the logic. The second law states the tendency toward low entropy. But it doesnt ask about initial conditions.
For whatever reason, the universe started as a massive concentration of energy - A big ball of negative entropy (negentropy). From there the universe evolved (evolved here in the mathematical term, that is its change in state over time) into states of higher entropy (lower negentropy, spending it down). The second law absolutely does not preclude any process by which entropy is decreased within a particular control volume - it only precludes a universal decrease entropy. Entropy may decrease here as long as it increases by more there.
At the start of the universe when it was a near uniform distribution of energy in a very tight space inside a massive volume, it evolved into states such that that energy was concentrated into particular clusters of a size determined exactly by the physical constants. (Stars and Galaxies). For eons, those large balls of negentropy slowly spend themselves down fusing elements. All this is described without much requisite "order". Only the constants themselves and the tendency for exclusive spaces to be filled with competitive entities. (Stars and galaxies "sequester" energy into small pockets relative to their zone of influence. This happened because while certainly other structures were possible in an even distribution of energy, anything that didnt sequester would naturally end up sequestered)
In the shadow of a giant ball of negentropy, however, strange things are possible. In the presence of so much negentropy to spend, and enough matter and energy to play around with structures of meaningful size, random chance can temporarily spring up all sorts of incredible "orderings".
The problem of the totally disordered universe with the single brain is describing how it ended up in ONLY and EXACTLY a single brain, along its evolution from a single massive concentration of energy (and therefore negentropy) to what it must logically become. An even distribution with a volume containing a human-brains-worth of particles lined up just correctly is NOT a likely trajectory of that evolution. Not because the alignment of so many particles is so unlikely, but because it is (statistically) precluded by the things that happen first.
So no, it is absolutely consistent with the laws of thermodynamics that a universe which begets a brain can be more likely than a brain existing among disorder, simply based on the mathematics and observable initial conditions.
Finally, I do not agree on the face of it that the universe is amazing everywhere on all orders of magnitude, nor that it is reasonable to say so. It certainly seems more interesting to me than a universe of zero entropy, where all energy is exactly evenly distributed. But then, I can't exist in that universe.
The one we inhabit certainly does not seem optimal by any metric. I certainly have some complaints. Dont get me wrong Im enjoying being alive. But it certainly does not seem designed to be amazing or incredible or with wisdom everywhere or to be interesting, the criteria that you hang your hat on for your argument to work. But lets note - weve totally gone away from the physics. You are appealing to human values. Gesturing at some cosmology we havent figured out yet, saying, "look! theres room enough inside this confusion to allow you to believe in God and wouldnt that be nice since everything is so amazing and incredible and full of wisdom".
You can believe whatever happy things you like but it certainly isnt coming from the physics.
The second law of thermodynamics states the tendency toward higher entropy, not lower entropy. Perhaps you just misspoke.
Also, the second law clearly indicates that the initial conditions of our universe had very low entropy. This is because of the apparent contradiction between the observations that lead to the second law and statistical reasoning. This is something that physicists from Richard Feynman to Roger Penrose clearly say.
In episode 8 and 9 of this series we'll discuss this all in depth. We just released episode 6 so it won't be too long for those.
And lastly, it is not an appeal to human values to say that a universe filled with atoms, molecules, planets, stars, galaxies, and life is a more interesting universe filled with more wisdom (chemistry, biology, astronomy, etc. in addition to physics) than a universe with only black holes (and just fundamental physics).
But, um, what is your definition of God? It certainly does not seem to me like an argument supporting an existence of the entity described under that label in the Bible, or in the Quran...
We're not trying to prove any book or religion. Just a God who intelligently fine tuned the laws of nature. We're going to have a separate miniseries about the idea of God that emerges from the fine tuning argument in physics.
Oki, that separate miniseries sounds really interesting!
Science is all about using theories to make testable predictions.
Does the God fine-tuning theory make any testable predictions?
We're arguing that God is the proper philosophical interpretation of the physics.
Adding a 'God' to the system doesn't remove the 'problem' of fine-tuning, it makes it much much worse - now you need to explain a super-powerful extra-universal creator! That's way way more complicated that suspiciously narrow viable ranges for a handful of free parameters
Not really. Unless you posit literally infinite regress of causes or contingencies, there must exist at least and at most one non-contingent, eternal, immanent, et cetera Thing that can cause everything else that follows: this we understand to be God.
Religion can be very complex with a lot of moving parts that have to be justified, but the metaphysical first cause is not just parsimonious, but really necessary unless you embrace endlessly elaborate and frankly—to me—incomprehensible worldviews.
The concept of cause is itself an abstraction from experience. Trying to make it precise and workable (e.g. in the context of machine learning) is far from trivial. I don’t think it is necessarily a good idea to extrapolate an ill defined notion of cause to the beginning of the universe without dealing with the philosophical intricacies of the concept first.
Is it? I don’t see a strong argument that the concept of causation is hazy or suspect at all.
This can be discussed from a bunch of different angles. Classical mechanics essentially boils down to equations such as Hamilton’s, that allow us to calculate the coordinates and momenta of a system at time t given their values at time 0. It has no obviously defined notion of cause and effect. The system just does its thing; reversing its time evolution would still yield a valid solution to the equations. So it’s hard to say what causes what. Quantum mechanics is not too different in unitary evolution (basically the smooth evolution that happens when no one is looking). Cool stuff -irreversibility- happens if you take measurements, but I understand it’s an open problem still. All in all it seems that a meaningful notion of cause has to emerge statistically, rather than being fundamental; pretty much like irreversibility in statistical mechanics emerges. To think about it, this seems reasonable: you wouldn’t say that smoking causes cancer if you had access to only one person that smoked a fixed amount of cigarettes. Either that person gets cancer or he does not and in both cases there is no counterfactual. So ‘smoking causes cancer’ is a statistical statement. There has been quite a lot of debate in statistics around the status of causal statements like this. Until recently it was consensus that statistics can just detect associations. The notion that we can do more than that is fairly recent and still somewhat controversial (you can look into the work of Judea Pearl if you are interested).
Non sequitur; this is an expected result of both a clear understanding of causation or a confused view where we claim that we don’t know anything.
"Not really. Unless you posit literally infinite regress of causes or contingencies, there must exist at least and at most one non-contingent, eternal, immanent, et cetera Thing that can cause everything else that follows: this we understand to be God."
Why not posit that it's the laws of physics? The laws of physics are, as far as we know, non-contingent, eternal, and immanent.
The laws themselves are not noncontingent; their action is contingent upon the existence, too, of space time itself. Likewise spacetime is contingent on some laws that define how it behaves.
While one may be able to posit that, the laws of physics were discovered to have fine tuned constants which indicates that they have an intelligent cause. We take up this possibility and develop this argument more fully in earlier episodes (3-5).
Do we have any reason to believe there is NOT an infinite regress of causes? Whenever we look closely enough, so far, we have found things to be more complicated. We have also not proven anything to be fundamental (which proof may be impossible, like proving God's non-existence).
>Do we have any reason to believe there is NOT an infinite regress of causes?
Short answer: No.
Long answer: The notion that the arguments for God require denying an infinite regress of causes is a misconception. The traditional First Cause argument is actually based on the notion that there cannot be an infinite regress of *essentially ordered* movers. This is because an essentially ordered sequence is defined partly by the fact that, if the first mover in the sequence is removed, all of the objects after it will cease to move. Therefore, if any object in the sequence is moving, then the sequence must have a first mover.
However, this entire line of thought is irrelevant in the modern world, because the concept of an essentially ordered sequence is based on a debunked Aristotelian notion of how motion works. In reality, any time you have a chain of moving objects and you remove one sequence in the chain, the remaining objects will still keep moving until an outside force stops them. So the argument for why you can't have an infinite regress of causes is no longer valid.
This is a good point. It seems to me that infinite regress of causes means nothing when it comes to the existence, or not, of God. We can't prove the assertion one way or another. So it comes back to simply whether or not one believes in God without proof.
The path of scientifc knowledge goes from the complex entities and events we observe, and explains them based upon simpler and simpler fundamental particles and laws.
"In the whole history of science from Greek philosophy to modern physics there have been constant attempts to reduce the apparent complexity of natural phenomena to some simple fundamental idea and relationships. This is the underlying principle of all natural philosophy." – Albert Einstein, The Evolution of Physics pg.52
Sure, but once upon a time we discovered the basic building blocks of everything were photons, protons, neutrons, and electrons. Oh, and neutrinos. Oh, and muons. Oh, and anti-matter, and quarks, and gluons, and mesons...
And what are any of those made of?
Good question. Depending on who you ask, they are believed to be fundamental. Or perhaps they're vibrations of strings...
Do you think it's reasonable to have an infinite chain of simpler and simpler particles? Though my intuition is that it's not, I cannot prove it. Perhaps it's simpler and simpler turtles all the way down?
We will discuss this more on our miniseries on God.
I would say that infinite regress has other worse problems, but yes, it’s a pretty bad idea.
Additionally, at best it gets you to deism, an absent creator who doesn't touch the universe after it's created; not many religious implications to that kind of creator, frankly
It also doesn't get you to a conscious creator. The "creator" it gets you to could be as dead and mechanical as a rock--or, as I said in an earlier post, the laws of physics.
The argument gets you to a Creator who intelligently fine tuned the constants of nature to result in our complex universe which contains atoms, molecules, stars, planets, life, etc as opposed to consisting of a chaotic sea of fundamental particles (the result without fine tuning).
Note that a benzene ring fetishist creator is equally compatible with fine tuning. Some entities just like organic chemistry... :-)
We address your question in the last couple of minutes of this episode, albeit in a short manner. We’ll elaborate more in the miniseries about God. But you’re right, that we’re not dealing with religion. Only God at this point.
Hi everyone! :) I conducted several interviews with a bunch of Berkeley AI alignment researchers in February. I distilled them into vignettes which include themes such as (perceived) social pressure, status dynamics, and self-censorship. I'm curious how many of you can relate to them! https://www.lesswrong.com/posts/yXLEcd9eixWucKGHg/the-seeker-s-game-vignettes-from-the-bay
(if you prefer the EA forum - https://forum.effectivealtruism.org/posts/WxqyXbyQiEjiAsoJr/the-seeker-s-game-vignettes-from-the-bay )
this is the kind of content I'm here for. thank you for putting it together!
This is another update to my long-running attempt at predicting the outcome of the Russo-Ukrainian war. Previous update is here: https://astralcodexten.substack.com/p/open-thread-277/comment/16392854.
15 % on Ukrainian victory (up from 14 % on May 22).
I define Ukrainian victory as either a) Ukrainian government gaining control of the territory it had not controlled before February 24 without losing any similarly important territory and without conceding that it will stop its attempts to join EU or NATO, b) Ukrainian government getting official ok from Russia to join EU or NATO without conceding any territory and without losing de facto control of any territory it had controlled before February 24 of 2022, or c) return to exact prewar status quo ante.
46 % on compromise solution that both sides might plausibly claim as a victory (up from 45 % on May 22).
39 % on Ukrainian defeat (down from 41 % on May 22).
I define Ukrainian defeat as Russia getting what it wants from Ukraine without giving any substantial concessions. Russia wants either a) Ukraine to stop claiming at least some of the territories that were before war claimed by Ukraine but de facto controlled by Russia or its proxies, or b) Russia or its proxies (old or new) to get more Ukrainian territory, de facto recognized by Ukraine in something resembling Minsk ceasefire(s)* or c) some form of guarantee that Ukraine will became neutral, which includes but is not limited to Ukraine not joining NATO. E.g. if Ukraine agrees to stay out of NATO without any other concessions to Russia, but gets mutual defense treaty with Poland and Turkey, that does NOT count as Ukrainian defeat.
Discussion:
Much has happened from previous update, but imho none of it fundamentally changed the situation.
What finally prompted me to slightly change the odds was Biden’s decision to supply cluster munitions to Ukraine (btw. Russia already uses cluster munitions against Ukrainians). And that is for two reasons: a) it signals slightly (not much) deeper US commitment to support Ukraine that I would expect, and b) those might be actually practically fairly important on the battlefield (for a quick explanation see e.g. here: https://www.youtube.com/watch?v=deK98IeTjfY&t=3459s).
*Minsk ceasefire or ceasefires (first agreement did not work, it was amended by second and since then it worked somewhat better) constituted, among other things, de facto recognition by Ukraine that Russia and its proxies will control some territory claimed by Ukraine for some time. In exchange Russia stopped trying to conquer more Ukrainian territory. Until February 24 of 2022, that is.
When the Ukrainian counter-offensive began I was hoping for an early decisive breakout, but it became clear in the first two weeks that wasn't happening.
My next hope was that Ukraine would make broad, slow progress across the whole front, and slowly attrit Russia forces. It's evident that isn't going to happen either. All we've seen are minor tactical gains. Russia has a seemingly endless supply of convicts to man the front lines, a compliant populace, and a military industrial base that is more than capable of replacing lost equipment. All of the talk about an impending Russian collapse has turned out to be false.
The current lines probably aren't going to change much, barring a direct NATO intervention or a 10x increase in Western assistance. I hope I'm wrong, but it's looking like Ukraine is going to have to agree to a negotiated peace on Russia's terms.
I mean, yea, that is what I have been saying the whole time.
I feel like Prigozhin's Midsummer adventure warrants a mention here. How does that not affect your probability of a Russian defeat as a result of a collapse in leadership?
From the example of Erdoğan, who survived an attempted coup in 2016, it's possible to end up stronger than ever as a result.
We understand so little about what actually happened (still happening?) that the prudent course of action is to ignore this event for now. No way to incorporate it into a forecast today without having to update it tomorrow.
I understand why that seems relevant, but now it looks like Putin has roughly the same level of control as before that exciting adventure.
Of course the very fact Prigozhin was allowed grow his private army to such an extent is sure sign of the weakness of the regime, but it is not a new information that Russian regime is weak compared to liberal democracies; that is already factored in the prediction (btw. that weakness is imho the main reason why Putin felt the need to externalize its problems by invading Ukraine, but that is a different can of worms).
Isn’t it surprising that Prigozhin did not wake up with novichok on his underpants? I think it’s quite surprising that he is still alive. That suggests either this whole mutiny is some elaborate ruse, or Putin does not have the influence to simply execute him.
I mean, second explanation is rather obviously correct; Putin needs Wagner mercenaries, or more precisely, people who are currently Wagner mercenaries, to cooperate with him both in Ukraine, and also in Africa. These guys are reportedly basically running the Central African Republic and are attempting to do the same with Mali and other African countries. It is unclear how they would react if Prigozhin would be offed. Which doesn't necesarilly mean it won't happen, of course.
Also Prigozhin has some support among Russian public, which counts for something. Putin is a dictator, but he cannot afford to completely disregard public opinion.
For people like me who only occasionally read these, it might be helpful to mention how things have changed over the long term. Maybe a graph showing the three numbers change over time?
I'd love too, but I don't think substack interface supports graphs in the comments (or am I wrong?), and probably aproximatelly no one clicks on the links, so...
I'd click a link to a graph
Ok, maybe next time, if I get over my laziness.
I started a blog on meditation and spirituality. One of my pet peeves was that what most people think about meditation comes from marketing: https://shant.nu/most-of-what-you-know-about-meditation-comes-from-marketing/
Where meditation is sold as this happiness pill - just pop one and all your problems will be solved! And so one of the most common questions on meditation forums is: I’ve been meditating for years and still don’t feel happy.
This myth is harmful because people are sold a dumbed down version of meditation and never stick around with it to see its real benefits.
Two philosophers walk together through the cobbled streets of a university town.
“Did you know”, says the first scholar, “that one hundred multiplied by a million is less than one multiplied by a billion”
The second scholar is stopped in his tracks.
“You are right. It is so!”, he declaims “but, but …. this is a repugnant conclusion!”
OK, I don't get this. I understand the maths (obviously, 100,000,000 < 1,000,000,000) and I understand the "repugnant conclusion" reference to the the happiness of populations, but I don't get why this phrasing is funny, or thought-provoking, or whatever it is meant to be...
That’s all really. The maths works out.
100 or 1 = happiness level
1M or 1B = number of people
The math is representing the repugnant conclusion, but also the arithmetic itself it counterintuitive making the equation even devoid of context "repugnant".
(If you still don't think it's funny don't worry, you're not alone)
Yep, I realised the happiness level and number of people were being represented in that way. OK, I didn't miss anything, fair enough, thanks.
It is interesting that there are people who scorn spherical cows but will accept the usefulness of "utils."
Wouldn't cubic cows be more efficient, since they could have less wasted space?
https://twitter.com/ilplantdoc/status/823736868974039042
Love it!
My favorite protein bar maker is being sued for lying about nutritional values (e.g. a claim that their bars have three times as much sugar as stated). They claim they're telling the truth and the prosecutor must have tested a bad batch or something. Is there a way I can test these bars myself?
(Context: this is an Israeli company that I don't think sells internationally, so the only authorities that would check it would be local ones. Otoh the bars definitely don't taste like they're high in sugar, so I'm inclined to believe the company that they're not).
The problem probably hinges on the exact definition of "sugar". Like: Where exactly is the cutoff in chain length between sugar and non sugar carbohydrate? Do all sugars count (fructose, lactose, maltodextrin...)?
Quantifying glucose in a home lab is quite possible, though probably not practical for you if you have to ask about it. Doing a complete sugar profile would be um... a project. Finding a commercial lab to do it is certainly easier and probably cheaper.
Re exact definition: Doesn't the government (in the US the FDA, in Israel presumably some similar body) have specific definitions for how to measure things for the nutritional information on the back of the box? (to be clear we're going by that, not by some ad making vague claims).
An example of what you could do: Precipitate CuO from acidic CuSO4 solution, weigh the precipitate to quantify. If all sugars do this, how oligosaccharides react, if other contents of the protein bar would be a problem and so on, I would have to look up.
There actually turns out to be a distinction between https://en.wikipedia.org/wiki/Reducing_sugar which do include glucose, and non-reducing sugars like sucrose. "In an alkaline solution, a reducing sugar forms some aldehyde or ketone, which allows it to act as a reducing agent, for example in Benedict's reagent." https://en.wikipedia.org/wiki/Benedict%27s_reagent indeed contains CuSO4 (along with sodium carbonate and citrate), and indeed yields a copper oxide (albeit Cu2O rather than CuO) in the presence of reducing sugars (and probably other reducing agents).
Could it be a difference between "sugar" and "added sugar"? That difference can be pretty huge if the bar contains milk or fruit.
I hate the big “No High Fructose Corn Syrup” on products that list fructose corn syrup as the first ingredient. But it not the high concentrated stuff you see so obviously a healthy choice.
I think how to measure nutritional content for the text on the back of the box is standardized though? At least, I would have assumed so.
I have been surprised by the amount of sugar in 'healthy' or 'health food' bars, and this is something I have to watch out for.
Just looking online, here's a lawsuit that I had no idea about (and it involves a dairy co-op from round these parts):
https://topclassactions.com/lawsuit-settlements/consumer-products/food/think-class-action-claims-protein-bars-falsely-advertised-as-containing-no-artificial-sweeteners/
"Sugar" covers a lot more than just "glucose" which is what most people think of when they hear "sugar". Unless you have a list of ingredients, it would be hard to tell if they mean "low in sugar", "no added sugar" (but plenty of naturally high-sugar ingredients), artificial sweeteners or the like.
There is no such thing as a “healthy” food bars. They are useful as a dense source of energy on long hikes and the such.
See also: trail mix, for applications not involving a trail.
They're certainly being presented as 'better for you than a candy bar or bar of chocolate' with all the fibre etc. and then I read the nutritional info and go pale.
Yep! I wouldn't be surprised if an actual bar of dark chocolate were "healthier" than many food bars.
I know that “no added sugar” is common on some produce with high normal sugar content. People read it as no sugar.
A patient of mine saw a psychiatrist who had him do genetic testing to get additional information about what meds would help him. I quickly looked up the company that does this, and found some pretty limited research into efficacy by people not with the company. Upshot seemed to be that it helped some with choosing medications for depression, at least, but not a lot and only on certain measures. Anyone know any more about this approach and whether it's currently worth using or mostly hype?
Definitely not worth using. Currently, it is totally useless to help a given person, cf for example this recent review
https://www.sciencedirect.com/science/article/pii/S0165032722001719
I just read the abstract, which is pretty detailed. They didn't conclude it was useless. In fact, they seem to have sort of bailed, saying studies used different endpoints so no meta-analysis was possible. (That seems odd -- studies measures never match up perfectly in meta-analysis, there are ways to work with the differences statistically to make it possible to compare results.) But they are saying, PRS might prove to be quite useful. If there's an article saying approach is bogus I'd like to know about it. Are you sure this is the one you had in mind?
See https://slatestarcodex.com/2017/03/06/antidepressant-pharmacogenomics-much-more-than-you-wanted-to-know/
In 2019, Scott wrote an article saying that a lot of it was based on bad science. No clue if they have improved since rhen.
https://slatestarcodex.com/2019/05/07/5-httlpr-a-pointed-review/
Anecdotally, a friend of mine did that, after trying a bunch first. Did not seem to make a difference for them, the "most genetically suitable" psych medication still did not work well enough. But maybe better than the dr trying randomly.
Anyone know of a really good free flow chart tool like Visio for iPad?
app.diagrams.net ?
If everybody in the world suddently and permanently became more physically attractive (according to the world's current standards), setting aside any correlated benefiits such as improved physical health, would this:
A) ...simply be a positive change long-term, since we can all enjoy looking at more attractive people around us, and enjoy the benefits of being more attractive to others?
B) ...possibly be a positive change short-term, but eventually become a neutral change in the long-term steady state, because everybody's standards and preferences will simply adjust?
C) ...actually be a net-negative change, for some reason?
I think C). If everyone is more attractive, everyone has much more of an extra incentive to cheat on their significant others. It would be harder to stay professional at work with members of the opposite sex. Sexual harassers and rapists would commit even more sexual offenses, due to experiencing even stronger urges.
Want to say C*, but it's probably not actually C. With the caveat that we're literally just talking about a perception change**, have to go with B. Unless you're nailing everybody to the peak of beauty, the differences will still be maximized by society. You won't be more attractive to others, you're still just on the scale.
*(Sex is a distraction at work. If everyone looks more like Tom Cruise then seeing them will more likely remind you of how crazy Tom Cruise is. You don't see the drug use from across the street anymore, you find out when they're in stabbing distance.)
**(If we're talking about physical effects in ability, it's C. Every powerlifter in the world would suddenly lose 300 pounds of liftable weight.)
(Also the cat ears would be hard to clean.)
I think it must be B, since attractiveness will turn out to be relative. It could be worse, but then again, it could be better, to paraphrase Gandalf in The Hobbit. Even if you could quantify attractiveness and magically make everyone equally attractive according to your measure, people would find differences between them, no matter how small, and then a slight flaw would be amazingly ugly, and the top people would be amazingly beautiful.
B. People will get used to the higher level of physical attractiveness, and those born and growing up where everyone is not-ugly will take it as a matter of course.
So then attractiveness will shift to be even higher; now you have to be perfectly symmetrical in features and have the toned body and flawless skin even without makeup etc. to be considered beautiful.
I feel like it's A, but we must be much more attractive than past generations when everyone was malnourished and got lots of diseases, and I would have guessed that people in the past were just as horny and interested in physical beauty as we are.
Perhaps some of the causes of variation in physical beauty (e.g. smallpox scars) have disappeared?
On the other hand, obesity became considerably more common. I wonder if there were ever as many people seriously disfigured by diseases or malnourishment as there are today by obesity.
That's a good point! ( My personal tastes happen to run to heavier women, so I personally don't react to extra weight as detracting from beauty, but I recognize that this is atypical. )
>If everybody in the world suddently and permanently became more physically attractive (according to the world's current standards),
This question is based on a false assumption that there is such a standard. I'd argue that there is no universal standard that hold for the entire world, but rather many standards that are held by different cultures (and different groups and subcultures within these cultures, and then there are also personal preferences that can differ even more).
These standards overlap to some extent, but there are significant differences, and if you remake everyone according to just one of these standards, it will be a downgrade in the eyes of many who did not hold that standard. Also, physical attractiveness is usually only a part of attractiveness in general, and I really doubt most people focus only on physicall attractiveness rather than on general impression that combines it with other factors (behaviour, clothing style, etc).
And that's the case even now, after all the globalisation which did push Western beauty standards to other cultures worldwide, with noticable effect. The gaps used to be much bigger in the past, especially if you are willing to look some centuries back. If you look up that was considered to be a standard of attractiveness in say Russia, Japan, France and India in 12th century, you will find even less in common.
So I'm not sure if that's C (because it's hard to quantify if there will be more winners or losers), or B (because whatever effect it takes will be temporary, people will invent new standards pretty soon).
Of course, if there are people who are ugly by all standards, it would be a good thing to make them more attractive by any standard. I'm not really sure to what extent that would be possible though, and it would be a rather small effect overall, so I don't think that would change the conclusion.
As others, I tend to say A. But B: Just saw some pics of pretty Norwegian women (me cis male). Nice. But: Swedish look even better - so: meh. Also: Could Hugh Hefner remember which of his "bunnies" was who? https://media.radaronline.com/brand-img/kI15V7hJA/2160x1130/hefner-pp-1661461380772.jpg
C: your scenario says: health et al. stays same. But different looks - esp. a broader spectrum of less and more pretty - did signal "better"/different health/genes. Thus: loss of valuable information. (As studies seem to show, the closer to the "average" a portrait-pic was, the more "attractive" it was judged. )
Among his many flaws, Hugh Hefner did not include the last woman willing to have sex with him in his will.
I think most people do not include "the last person willing to have sex with them" in their testaments because that would set up incentives resulting in a rather undignified death. :-)
Also relevant to Mark:
IIRC (and if my gossip sources are correct) of his four "girlfriends" that accompanied him in public/lived with him/slept with him, only one of them would have sex with him. I don't know if she got a bonus stipend compared to the other three, but if she didn't then not including her in the will makes me dislike the man in the same way I dislike lousy tippers.
I assume he paid up-front (his bunnies had it materially good, I heard. Mostly hating each other). If "my last" should not be my wife - I doubt, I would or should change my will for her.
I believe A. Generally when people to a more scenic sea shore or forest, they still consider it scenic years later and not this is boring and everyplace else sucks.
Seems like this would reduce the difference in attractiveness between real people and fictional representations optimized for attractiveness, which seems like a good thing to me. (Because I expect it would tend to make people happier with their actual partners, and also make it harder for advertisers to manipulate us.)
If you're asking purely about the direct hedonic effect of seeing more total attractiveness in daily life and whether the hedonic treadmill would eventually negate it, then I'm not sure.
You made me change my mind from (B) to (A). We already spend a lot of time looking at the most beautiful people that exist, and presumably this change would narrow the gap between those people on the one hand and ourselves and our partners on the other, which seems like it should increase happiness.
You know what, I think the Barbie movie is a great test for this! I haven't seen it and wild horses couldn't drag me to it, but the basic plot line is this example: everyone is Barbie-pretty (be they blonde Barbie, BIPOC Barbie, Ken, etc.) and they live in the Barbie-world of sun and fun and happy days and contented nights.
The story seems to be that the less than perfect dolls get exiled to the real world, but of course - what makes them less than perfect in the first place? Why are they not satisifed with their utopia?
Would we be happy with Barbie World forever, or would there always be the less than perfect among us?
About the Barbie movie. It was totally below my radar but I just read a lengthy sympathetic article about Greta Gerwig the director and the hIstory of Barbie the doll. It seems like there is more to the whole thing than I ever would have guessed.
I doubt Mrs Gunflint and I will make a dinner and a movie thing out of it - that kinda came to halt with the rise of the MCU - but I’m interested enough to watch for it on cable.
This is the most interesting thing I've heard about the Barbie movie, although I suspect it's a story older than Barbie, and perhaps even time itself.
I think it’d be A. The steady state might settle lower than the initial shock, but overall it would be better.
As an example consider the sense of involuntary repulsion an outlier person can cause in you. If these people were all within the Overton window of beauty, you’d never have that experience.
Although it’s possible perception of beauty would shift towards something else, like outfits or some more nuanced physical attribute.
Dear ACX readers.
I write about philosophy, psychology, economics, technology and health. \
Recently I have been exploring intersections between Alexander Technique/Expanding Awareness and other psychiatric/psychological themes.
On Acceptance and Commitment Therapy: https://zantafakari.substack.com/p/how-can-alexander-technique-boost
On modulating conscientiousness: https://zantafakari.substack.com/p/16-refine-your-conscientiousness
The themes might be of interest to some of you - if so please feel free to read, share and subscribe. As always, I would love to hear your thoughts and any differing views.
Did Scott Alexander ever review the Alexander technique?
Not directly - he did put it in the "woo" category here:
https://astralcodexten.substack.com/p/are-woo-non-responders-defective
And he's written a lot on Nick Cammarata's experiences - and nick does AT (though this post is more about Jhana) - https://astralcodexten.substack.com/p/nick-cammarata-on-jhana
I continue hearing reports that 5-MeO-DMT is the most powerful psychedelic and I’m curious to hear more takes from a rationalist perspective. Please anonymously share anecdotes and resources.
"most powerful" doesn't make sense. The dose decided how powerful the effect is. All psychedelics are very powerful at a high dose.
Would you also agree that "more toxic" does not make sense, since water can be just as toxic as botulinus at the correct dosage?
I agree though that for most pharmaceuticals, the required dosage should probably not be the main concern compared to other desirable qualities such as selectivity or side effects.
If someone asked "I continue hearing reports that botulinum is the most powerful toxin and I’m curious to hear more takes from a rationalist perspective", I think the correct response is that "more powerful" doesn't make sense since dosage decides how powerful the effect is.
If the hypothetical poster is interested in LD50 values, or if OP was interested in dosages for psychedelics (LSD has way smaller doses than DMT by the way) I would presume that they look that up on Wikipedia instead of asking here. But that doesn't seem to be what OP is asking for, instead it seems to me that OP is working on some misinformed pop understanding of how psychedelics work.
Fair, I may have put too much burden on the reader to parse my meaning.
Is that an insult?
If you're looking for trip reports, then just go to Erowid like everyone else.
Do you also pedant at someone who says iron is heavier than aluminum?
I have a Substack so obscure that only one of its posts has ever generated any subscriptions. Now it occurred to me I could use the allowed once-per-year unsubtle-plug-outside-classifieds to explore that post's appeal to a wider audience than the Classifieds Thread reached.
Its point is that increased political polarization and dysfunction could simply have to do with the moment in time when the progressive side achieved victory on all its reasonable demands (think of gay marriage) and therefore moved on to unreasonable ones (think of certain trans demands).
https://birdperspectives.substack.com/p/progressive-train
Unrelated to that post, I also had another thought related to polarization. Are there culture-war debate formats on TV like, say, the Soho Forum Debates? If not, why not? They could be a win-win: on ratings, because people like culture-wars battles; and on political literacy, because viewers would, for once, get exposure to both sides of a contested issue.
Perhaps the relative nature of these formats would be unsatisfactory for a TV audience? (Your side will never get full support and might even win by merely going from, say, 30-70 to 40-60 behind.) Or do liberal media realize that on the identity-politics issues dearest to them they might lose once an audience is properly exposed to both sides of the argument? In the case of the US, which has a conservative TV station (Fox News), the latter cannot be a complete answer.
There's a state involvement as a ratchet aspect to this question. Over time the number of people who question if the Leviathan really is the best option to do something has grown in concert with the growth of the Leviathan. So what was progressive in the late 1800s (say the temperance movement) that relied a lot on social pressure to make a change has shifted to a demand that the pressure come directly from government itself, who cares what "society" wants. As the Leviathan gets bigger, it puts the resistors in the position of also having to rely on it to stop/retard change, which of course, just feeds the Leviathan making it easier to use it to beat up on the smaller group. Or something like that.
I've noticed this dynamic as well. My diagnosis is that the concept of progressivism as a movement is too abstract and unmoored from concrete concerns that it will inevitably go off the rails. Progress for progress's sake is unsustainable. There's no built in limiting factor that constrains what is considered a proper target of optimization. You eventually run out of sane dimensions to optimize and the movement is left to latch onto more and more unreasonable targets.
But the concept of progress for its own sake is antithetical to questioning the reasonableness of the current direction; self-reflection is verboten. Social media greases the wheels of this sort of degenerate progress because it incentivizes jumping onto these movements uncritically. The Good and the Moral will always have some intrinsic value for self-promotion. Progressivism as a concept situates itself as the arbiters of the Good and the Moral, which makes its excesses so hard to push back against. One of the biggest challenges this century is giving people the conceptual tools to say no to people who claim the mantle of the Good.
That‘s some good analysis on an obvious question that I chose to sidestep in the post, namely „why would progressives still want to progress further once they have won“. I agree particularly on the importance of the Good, the Moral, and self-promotion. A key concept here, I think, is prestige status, to be distinguished from dominance status.
In the post, the situation gets complicated by the fact that, after having moved on from the topic of trans demands, I end up not actually insisting that the progressive side has won (as I now put it in the mini-summary) „on all its reasonable demands“. There may still be implicit discrimination —- yet I argue (in the post) that a shift from explicit to implicit can already explain increased polarization.
Cynically, you can probably get all the heat (which generates the ratings and money) from a one-sided rant or bringing a strawman on to mock or talk over. Optimizing for political literacy brings the network no additional benefit, and might be a harm if they diversify their audience, reducing ad impact.
Theres a debate program somehow affiliated with npr called intelligence squared that featured "hot button" issues. Iirc one of them was "is Islam a religion of peace?" Of late they seem to do more moderated discussions or less sensitive issues, though.
I believe there is a YouTube channel that does this, but it was more along the lines of putting ordinary people on opposite ends of an issue in a room and having them discuss. I think topics included flat earth and trans rights.
I thought a boost to political literacy could be a nice side effect from a format that would be good for ratings (the way I put it in the original comment wasn‘t ideal). But I guess I understand now what the problem is, thanks to you and to Zan Tafakari, who says in a comment below that Piers Morgan indeed brings opponents on „more with an intent to mock“. Even if I’m right that contests are good for ratings, why risk a loss for your side in a genuine one, if you can instead just mock weak opponents or strawmen.
What makes gay marriage reasonable? For basically all historical cultures across the millennia the idea of two men getting married would be insane, even for those cultures that allowed the open practice of homosexuality, usually as pederasty.
The biggest difference I see is that trans demands are heavy on other people (use of preferred pronouns, access to women's changing rooms and prisons, health insurance coverage for cosmetic procedures) while gay marriage does not impact others much unless gay people deliberately go out of their way for it (wedding cake bakers)
These demands came in conflict with other influential leftist groups like british feminists resulting in some pushback.
I would say that „heavy on other people“ versus „does not impact others much“ has something to do with reasonableness (also, it‘s not only or mainly feminists that push back against trans demands).
But admittedly I never define in the post (because I can’t) what I actually mean by reasonableness. Historical standards don‘t cut it for me, but that‘s only implied, when I write that we should thank progressives for overcoming slavery and inherited absolute power.
Your posts look interesting - I've subscribed! I also write a fairly new and esoteric substack which you might enjoy: https://zantafakari.substack.com/
I mean often times the left leaning or right leaning news outlets will bring on opposing people - but I wouldn't say it's a "debate".
Sometimes there are panel shows (like Good Morning in the UK) where opposed view people are brought on - but the debate is not exactly moving the dialogue forward.
I suppose it's not entirely in media's interest to have a reasoned debate if something more polarizing keeps the eyeballs watching longer. So I think it's probably "unsatisfactory" as you say. I doubt that media are concerned that people would change their minds but have no evidence to back this up. Perhaps its an intrinsic idea I have that it takes A LOT to change an opinion, particularly a political one, and TV debates aren't likely to cut it
Thank you! And I'm looking forward to exploring your Substack (which is much less obscure than mine). Starting with post number 12, for the philosophy and evolution of language. As for the debates in my original comment, they could still be bad-tempered (like the US presidential debates were last time). And if they are good for ratings, the media don't have to have a higher purpose such as promoting political literacy. But you say that already "often times the left leaning or right leaning news outlets will bring on opposing people": I have to admit I didn't even know that they do at least that (I don't watch TV these days).
Yeah they do - Piers Morgan often brings on people he disagrees with - but often more with an intent to mock rather than engage in meaningful debate.
Better that you don't watch TV these days - not worth it
That‘s indeed a good counterargument, a worldwide phenomenon is exactly what the social-media thesis (which I frame as the competing thesis in the post) would predict. I thought there was a specific Western phenomenon, though, exemplified by Trump, Brexit, Le Pen, etc.
Anyone have recommendations (ideally ones that helped them personally) for books/things to read about how to come to terms with things one can't change about oneself? This is basically stoicism, but I'm curious if there's anything that people have found highly effective in actually creating these mindset shifts.
I have a friend dealing with something (going to change it entirely for anonymity), let's say they aren't an amazing basketball player and wish they were. What are techniques people have found effective for coming to peace with qualities about themselves they wish were better? (Since obviously the basketball example is made up, I'm _not_ looking for responses of the form "convince yourself that you never really wanted to be good at basketball anyways").
I'd recommend morning pages.
If the thing your friend can't change is really something of about the size and specificity of playing basketball, there's something a bit funky going on. There are probably dozens of things of about that size he's no good at, right? -- like, say, water skiing, chess, portrait painting, decorating one's home attractively, tennis, etc etc. Why is he so distressed about lack of awesome talent for this particular skill? He must have pinned other hopes on it. I know someone is convinced he can only find a woman if he becomes a millionaire. He's an extremely smart guy getting a degree in a respected profession that is guaranteed to make him upper middle class. His is kind and funny when he's comfortable with someone. His looks are ok. I'd say they're average, but could be better than average is he lost weight and/or worked out some. But he's convinced he has to be millionaire to attract a woman's interest. Is it something like that with your friend -- he thinks being an awesome basketball player is the only thing that might make him worth knowing? If so, that's just not true. If the story is that he's devoted years to becoming excellent at basketball, and is forced to admit he'll never get better then Above Average, that's a different story, and needs to be examined a different way. Now if the skill is something most people would consider of more significance than basketball -- say, being a good conversational partner, or a good writer -- that's still a different story. Is basketball really a good stand in for whatever he's bad at?
It’s not basketball at all. The op says he changed it entirely for anonymity.
Oh yeah, I get that. I meant is it something on the order of basketball, or it he using a basketball as a stand-in for something most people would consider it much more important to be good at.
I'm using basketball for something people consider to be more important, definitely not a perfect analogy.
If it's something about the person's physical appearance it is possible they have a condition called Body Dysmorphic Disorder, which is generally considered a form of OCD.
I'd suggest checking out Core Transformation (Connierae and Tamara Andreas).
IDK about ACT, but I'm concerned it may not engage directly with the underlying structure generating the suffering that stems from "I must be X, but I'm not".
There's *usually* an implicit continuation to that sentence, ", therefore A B and C core needs will not be met/drives will not be fulfilled and expressed, and that's excruciating.", which IMO is a completely coherent stance to have *if that inference and the chain behind IT is valid*.
But it's almost never valid.
And accepting that "I'm not X" can and often is conflated with accepting a bunch of dubious (to say the least!, empirically speaking) implicit inferences that are a lot more excruciating; which is why there's so much resistance to accepting that reality in the first place!
A) Clarifying the 'outcome chain', and B) grounding that in a core state (for this, please read the book/refer to more detailed descriptions) leads naturally to the unwinding of the suffering-causing resistance to reality, and crucially, *to a re-orientation to the system's actual desires and goals and needs and values*. The surface-level thing is not so much 'accepted' in some resigned way as the undergirding delusory meaning structures are dispelled and rendered irrelevant in the clear light of direct insight and pure desire.
(In the relatively rare cases where there is in fact something to grieve, well, this allows for that to happen organically, along with the subsequent re-orientation that the denial was blocking.)
Good luck to your friend. May this situation resolve for him, and may he be fulfilled and happy!
If you want novels that basically have a theme of “well, it’s not so bad as that” I know it seems trite but I love the Discworld books for this.
Almost everything genuinely effective is going to need at least some details, so I expect you'll struggle to find a good response here. Like, advice with respect to sexual orientation is going to look very different from advice with respect to disability, is going to look very different from advice with respect to talent or g-factor.
The only genuinely useful and *universal* trick I can think of is meditation, starting in the mindfulness/samatha end of things and progressing to insight meditation after a few months or years when you feel ready for it. I'm particularly fond of "The Mind Illuminated," but there's no shortage of eager introductions really.
Yeah I figured it was a long shot, although Scott's answer looks promising. I can't really share more though. Unfortunately I think they're averse to a lot of meditation since most meditation practitioners are also selling philosophy and possibly even religion alongside it.
I don't know anything about it, but I understand there's a psychotherapy (Acceptance and Commitment Therapy) which is about this.
Yes I was also going to say Acceptance Commitment Therapy. I've actually written about ACT and it's relationship to Expanding Awareness (a branch of Alexander Technique) - which might also be helpful.
Have a read here:
https://zantafakari.substack.com/p/how-can-alexander-technique-boost
https://zantafakari.substack.com/p/16-refine-your-conscientiousness
I'll suggest they check it out, thanks!
ACT is in many ways similar to CBT (in fact I think it is described as a third wave CBT), except that the emphasis is more on changing one's relationship to one's thoughts instead of changing the content of one's thoughts. I think ACT is meant to be for all the things CBT is for. So if a typical CBT exercise is to write down reasons to challenge a negative thought, ACT focuses more on seeing thoughts as mental events that occur but don't control you (you learn not to be "fused" to them). That is the Acceptance part. The Commitment part is about clarifying one's values and acting on them (the behavioral part). There is a self help book called Get Out of Your Mind and Into Your Life by Steven Hayes, who developed ACT. I thought it was pretty good.
Are medical professionals in training subject to hazing and abuse, and does this lead to worse patient outcomes in hospitals?
This claim of medical interns being abused by requiring them to work extreme hours and crazy shifts is pretty common. Based on a few conversations I've had with medical professionals, it appears to be accurate.
One would expect that medical professionals would be most likely to be familiarzied with literature on the importance of being well rested in order to achieve optimal performance. And yet medical interns are claimed to often be denied this "luxury".
If so many people know about this toxic work environment, and it's bad for both patients and workers, why does it persist and what would it take for the system to change?
Medical residents are worked hard for a few reasons:
-- Hospitals pay a lot of money to train them and want to get useful work out of them.
-- You learn medicine faster/better the more you practice, and a lot of doctors believe you can't really learn medicine in a reasonable amount of time working forty hour weeks.
-- "Handoffs" - ie when a doctor goes home and transfers their patients to another - are notorious for going badly. The leaving doctor forgets to convey important information and then the new doctor does something wrong and causes a catastrophe. Hospitals have tried hard to formalize/ritualize these and make them ironclad, but the best defense is still to handoff patients as little as possible, ie make doctors and nurses work very long shifts.
-- Yeah, I agree there's a sense of "I did this in my day and let's not coddle the new generation by giving them a pass".
The ACGME, the relevant body, has passed various rules - the most important is that interns can't work more than 80 hours/week. These are both overenforced (in the sense that everyone constantly has to sign a lot of paperwork saying they definitely didn't work more than 80 hours/week, waste time listening to the Don't Work More Than 80 Hours/Week Campaign, and so on) and underenforced (in the sense that in real life, hospitals can just assign you X number of patients and say this is the amount a doctor should be able to treat in 80 hours, and then it takes you 120 hours to treat them, and if you complain they'll just say 'git gud')
I assume the “handoff” studies account for the obvious confounded that “more handoffs” usually means “more time in the hospital” ie probably a more serious condition?
The 80 hours/week thing is pretty confusing to me, coming from the IT world. Here we have to log our hours every day, and if we hit above 80 the system won't even let us put in more until our manager gives approval and allows overtime. Plus if we lie and just don't put in our hours that's illegal (in the contracting world anyway. probably it's illegal in normal private sector work as well).
How does it work in hospitals? If you hit 80 hours, but you still have patients to take care of, do you just keep working and then put in your hours and the hospital shrugs and says "ok we'll pay you overtime i guess"? That seems like it would outright violate the ACGME rule. Or are hours just not tracked at all?
"Hours"? I'm guessing doctors, even as residents, are not paid by the hour. Fixed salary, and if you don't complete the work assigned they will eventually convert that to zero work assigned and zero salary.
Ah, yeah that would make sense. I guess in that case the ACGME just made up these rules without having any actual way to enforce or check them. No mandatory time keeping or anything.
The "handoffs" issue seems like it would be equally solved by having people work longer shifts but fewer total shifts, such that the total hours/week was the same?
The handoffs thing seems like a fake reason to me (for example patients are handed off anyway when they go from the ER to the ward), but so does the whole “you can’t learn medicine without working 80 h a week”. I mean, sure, up to a certain point the more you work, the more you learn, but there’s nothing specific about medicine that makes it more true than in other fields.
As far as the hand offs are concerned, you could just as well argue that they are useful because the fresh doctor might come up with something the last one forgot.
The handoffs thing is absolutely fake. Pretty much everyone else who has to deal with continuous critical workload, e.g. the military, or in my field space flight operations, knows how to do handoffs properly. None of them demand 16-hour shifts or 80-hour weeks in anything but the most serious emergencies.
But if you *want* to make your residents work 24 hours straight or whatever, because that's how it was done in your day, these kids today, no respect, etc, etc, then sure, say something about "handoffs" and just make sure you're never officially seen studying proper shift handoff technique.
"I mean, sure, up to a certain point the more you work, the more you learn, but there’s nothing specific about medicine that makes it more true than in other fields." Also, doesn't sleep deprivation impede memory and learning?
Yes. It is also funny that the profession where supposedly people have to work such long hours practicing because it's so complicated to learn is the same profession where "see one, do one, teach one" is an established concept.
Good point!
Handoffs have been pretty extensively studied, long shifts (eg. 12 hours instead of 8) improve patient outcomes overall, but Dweomite's point is important too, that long shifts *should* mean less shifts
A lot of the problem boil down to a "doctor shortage", but a lot of that scarcity is artificial, both in the forms of restrictive quotas on number of doctors trained each year, and in restrictions on what nurse can do and what legally requires an MD. Keeping doctors scarce means higher wages for them, standard guild behaviour.
I think another major factor is that the primary bottleneck creating the doctor shortage is the cap on the number of residency slots, combined with the residency match system. There aren't enough residency slots to go around for all the med school graduates, and the places that hire residents coordinate their hiring so that each prospective resident has either zero or one job offer available to them.
So as an aspiring doctor, once you've been matched to a residency slot, your choices are 1) do the work they want you to do under their terms and conditions, or 2) give up on being a doctor and find some other way to pay off your med school loans. This gives you very little negotiating power, so the employer has very wide latitude to dictate the terms of employment, and you have very little recourse if mid-residency you don't like your working conditions. You have a little leverage during the match process if you look like a desirable candidate to multiple hospitals (since then your preferences between positions make more of a difference in the match process), but I think you're stuck with a residency once you're matched.
Without the binding constraint on the number of residency slots, hospitals would need to compete more to get med students to choose to do their residencies with them, and developing a reputation for having vaguely sane schedules would be one possible axis of competition. And if there are also changes to the match system to make it less collusive (e.g. allowing job offers outside the match process for senior residents looking to transfer), that would further improve the negotiating position of residents.
Yes, it was a good point and yes, there is a lot of gatekeeping in medicine.
12 hours vs 8 h seems understandable to me, but residents routinely work much longer shifts than that.
I found some studies where it said the impact of shift length on patient outcome was unclear.
Would it be a case of "paying your dues" given that understaffing is common in the training hospitals? Thus someone needs to cover the shifts etc, and that in turn falls to the new staff?
And for change, I do hear a lot of chatter on the matter, but I believe the bottleneck at the training level, where there is artificial restriction on the education needed means there is a perennial understaffing
Does anyone have good resources for someone looking to get across the evidence on optimising for healthy pregnancies (both prior to and after conception)? Is there an established rationalist-friendly person or set of posts? Preferably something that says here's what's consensus, here's what isn't, here's how we know X works etc. There's obviously lots on Doctor Google but I don't trust myself to identify who's a credible expert and who is well aware of e.g. publication bias and replication crisis, and has already taken these into account. Thanks in advance.
https://www.lesswrong.com/posts/yT22RcWrxZcXyGjsA/how-to-have-polygenically-screened-children
I'm sure there are good resources, but my impression from having seen a lot of people's take on being pregnant is that most are pretty resistant to advice. People you would think would be excited to see some actual data, and hear some actual new and well-supportedf ideas, often start making clear early on in their exchange with you that they have already made up their mind on what pregnancy and birth approach is best for them: They want a home birth. Or they want a super-conventional hospital birth, because they are convinced that hospitals give you lots of drugs or just knock you out so that's they way to get the best pain relief. Or they've found a nice book about pregnancy that made them feel better about the whole thing, and now they have bonded with it and do not want the relationship interfered with. They do not want to hear anything from you that undercuts their certainty about what ever take on the situation they have bonded with.
I think pregnancy and the prospect of parenthood is so stressful that people kind of regress to clinging tightly to ideas that make them feel less anxious, and do not want more input. I don't know what to do about that, except to be persistent and kind.
What's so great about your advice?
In my experience, of the masses of pregnancy and parenting advice available, not 10% will apply to any given combination of parent and child. Everyone thinks they've got a data-driven approach, but biology and behavior are idiosyncratic and highly variable in every aspect; edge cases comprise a long and meaty tail.
An expectant parent may as well plan the most appealing course, while staying informed and flexible enough to quickly change their approach as circumstances merit. Anyone who eagerly promotes universal advice is best kept at arm's length.
Yes, agree that anyone who eagerly promotes universal advice is best kept at arms length. You got any special reason to think I'm an eager promoter of my favorite pregnancy ideas, which I obnoxiously assume are universal truths? Cuz if you don't, consider that a person who uses somebody's post to build and attack an ugly little straw man is also best kept at arm's length.
I've got the post that you made, a very general diatribe against "people". If you don't care for my interpretation, then you're welcome to clarify.
Listen, when I write a diatribe you will know it: I will openly angry and critical. Somewhere on this thread I wrote a diatribe about GPT4, which begins "I hate GPT4." Here I am just saying people seem much more unwilling to listen to info and ideas than they are about other matters, and it's kind of striking and surprising.. At the end I say that's probably due to the stress of the immense changes in their lives. It might help you to know that I am a psychotherapist, and many of the people this comes up with are psychotherapy patients. The most recent time I got a reminder of the difficulty making recommendations to the pregnant was about a year ago -- so as you can see, this is not a topic I'm spewing advice about 24/7. The situation a year ago was the my patient had a long history of taking terrible care of herself. In her early 20's she had ignored symptoms of appendicitis and continued to be active until her appendix burst. She then became dangerously ill, was hospitalized, and had a bunch of her bowel removed. Ten years later, pregnant, she was again not taking decent care of herself. It wasn't as bad as the appendicitis situation, but really was worrisome. Other times when I have asked whether I can make a suggestion it has been with close friends, people with whom there has been a lot of discussion of private matters on both sides -- how to handle relationship problems, how big a bid to make for a house, etc. So the "people" I am talking about are people who have sometimes asked me for advice, and have been willing to listen to ideas even if they haven't asked for them, and have given me advice and ideas. What I am saying is that when it comes to pregnancy they seem different and it's kind of startling.
There is some truth to this to be sure. There's also just a *vast* amount of inbound advice, much of it poorly-sourced and conflicting.
Emily Oster's "Expecting Better". After you've already read that and optimized for everything important you can optimize for, you can also read my https://astralcodexten.substack.com/p/obscure-pregnancy-interventions-much
I had seen the latter but forgotten it also included not obscure recommendations, thank you very much for both of these
Parts of that book are pretty questionable, especially the greenlighting of alcohol consumption.
I wrote more about it here
https://www.reddit.com/r/BabyBumps/comments/j874co/i_read_expecting_better_here_are_my_thoughts/
Seconding this: Emily Oster is head and shoulders above the other baby books my partner and I have read (we've got one toddler and one on the way) due to things like assessing the quality of studies when citing them and being clear about the level of certainty and absolute risk involved.
Thank you for this comment, I will check it out!
Pretty sure Scott has a very long post summarizing the research on this topic. Every thing from air filters to supplements. It was a while ago but I think it was one of the “way more than you needed to know” themed posts.
https://parentdata.org/ is the closest thing to this that I know of. Generally research is well discussed and its limitations are acknowledged.
Fantastic, thank you very much
I recently read Stephen Omohundro’s oft-referenced papers, The Basic AI Drives https://citeseerx.ist.psu.edu/doc/10.1.1.393.8356 and The Nature of Self Improving Artificial Intelligence https://selfawaresystems.files.wordpress.com/2008/01/nature_of_self_improving_ai.pdf, and don’t understand why they are so influential as they are 100% argument by assertion. The big assertion is that AIs will necessarily, through self-improvement, become perfectly Rational and will therefore create utility functions with which they will operate entirely in accordance. Why will they become perfectly Rational? Well, because. As far as I can tell, almost every AI x-risk argument follows from that moverless-mover assertion. Why believe it?
A one sentence summary of those papers could be: “In the beginning was the Word, and the Word was Rationality.”
If you think I am wrong and Omohundro does explain in a non-circular way why the AIs will become Rational, will you explain it to me?
This feels kind of like arguing that dictatorships can't happen, because a dictator is someone who tries to maximize their political power and politicians are not perfectly rational power-maximizers. They aren't, but you don't have to be perfectly rational to decide that being a dictator is a good strategy, you just need to be capable enough to seize power and amoral enough that you don't care about democracy. So people who theorize about government will often say "imagine that a ruthless power-maximizer wanted to gain power in this system - how do we stop him from overthrowing the government?" even though such people don't literally exist.
Similarly, to pursue Omohundro's basic drives, you don't need to be perfectly rational, you just need to be rational enough to pursue a long-term goal, and be amoral enough that you aren't satisfied at regular human levels of power that we know how to deal with. And it can be worthwhile to ask yourself "Suppose this AI was a ruthless goal-maximizer, how do we give it a goal that doesn't drive it to conquer the world?" even if such an AI doesn't exist yet.
Why should we assume an AI has *any* level of rationality? Omohundro does, but I do not understand why he does. He assumes some modicum of rationality which will increase as the AIs self-improve.
Related question: Do you believe GPT-4 is partially rational, in the sense that it is partially interested in avoiding vulnerabilities?
If an AI has *literally no* rationality, it's not capable of doing anything logical, and there's no point in worrying about it. I'm not sure you could even call such an entity an "AI" rather than "a random noise generator." But any AI worth talking about will have at least some ability for rational thinking.
I'm not sure what you mean by "avoiding vulnerabilities," but the example of AutoGPT shows that GPT-4 is partially rational - it can take a goal, come up with an action that pursues that goal, and execute the action (with appropriate wiring). The example of ChaosGPT shows that it can even be prompted with long-term goals like "try to destroy the world." It's just very bad at it, because GPT isn't that smart and taking over the world is hard.
> If an AI has *literally no* rationality, it's not capable of doing anything logical, and there's no point in worrying about it.
I think you are conflating different senses of "(ir)rational" there. A pocket calculator isn't a rational *agent* but it's impeccably logical.
I'm using the definition of "rational" which Omohundro does. Under that definition, rationality isn't merely logical thinking but includes, fundamentally, "avoiding vulnerabilities", e.g., protecting one's own resources of time, space and free energy. Arguments that an advanced AI won't allow itself to be turned off stem from that notion of rationality.
(I can't help but notice a suspicious similarity between Omohundro's use of "rational" and the Rationalists use of the term.)
It involves, fundamentally acting towards some kind of goal. Avoiding vulnerabilities, that would prevent the goal, is secondary.
AI x-risk in no way follows from assuming rational behaviour - a totally *irrational* powerful AI sounds even more dangerous!
Aligning a rational intelligence is a very very hard problem, but aligning one that's unpredictable and random seems logically impossible.
>AI x-risk in no way follows from assuming rational behaviour - a totally *irrational* powerful AI sounds even more dangerous!
Perhaps, but I want to focus here on the assumptions Omohundro makes.
Yes, perfect rationality doesn't exist for practical purposes. Rationality is effective use of cognitive resources ,and always an approximartion. "Avoiding vulnerabilities" has a cost.
Why should we assume an AI has *any* level of rationality? Omohundro does, but I do not understand why he does. He assumes some modicum of rationality which will increase as the AIs self-improve.
Related question: Do you believe GPT-4 is partially rational, in the sense that it is partially interested in avoiding vulnerabilities?
An AI doesn't have to be a rationalist in some technical sense, such as von Neumann rationality. Confusion between technical and loose senses of rationality is part of the problem.
Heh. I just wrote a coupla paragraphs about why pregnant women do not want to hear any new and interesting data about optimal approaches to pregnancy and birth. Ended by saying "They do not want to hear anything from you that undercuts their certainty about what ever take on the situation they have bonded with. I think pregnancy and the prospect of parenthood is so stressful that people kind of regress to clinging tightly to ideas that make them feel less anxious, and do not want more input. I don't know what to do about that."
Seems like the same explanation applies to people's take on AI and it's possible dangers. I think quite a lot about AI and the future, and find that I'm constantly somewhat anxious about it, and crave to find others who are unsure who AI development will unfold but worry a lot about certain aspects of the process. It is very hard to find a peer group. So many people are wedded to the idea that AI's gonna kill us or else to the view that anyone who thinks AI is gonna kill us is a stoopit asshole.
Sure gets old after a while.
If you think of them as our descendants and their (current) strategy as making our lives so engaging and fun that birthrates for half the human population have fallen below replacement levels the prospect of, ah, an evolutionary step might hurt less.
Naw, Scott, but thanks for trying. Even when I'm thinking of the AI era as a big change with no doom in it, it still hurts like hell. I'm not old, but I am considerably older than you, and also happened to have parents who were in their 40's when I was born, and that gave me an unusually strong tie to life several generations ago. Then in college the literature I loved was 15th-19th century. And of course I became book-oriented, because there was no online in those days. Once read an interview with an elderly person who said "your world dies before you do." And I thought, no, that won't happen to me, I'll keep up with the changes. And up til now I have. I'm computer fluent, etc. But the shift to an AI-oriented world just looks ugly, bleak and weird to me, like a life lived inside a big bright windowless mall, and I can't really tell how much of that is an accurate perception that something vital is being lost and how much is just being unable to connect with such a different future. There is a lot about the internet era that I'm fairly sure truly is bad for people: Twitter. Facebook. And motherfucking Tinder is just *brutal.* If it goes one step further into hideousness, the poor people on there trying to find fun and love are going to have star ratings and reviews next to their names, in addition to being subject to the old left swipe. I always met my friends and honeys through actual activities. That is *so* different, and I'm pretty sure it's much closer to what we are wired for, and brutalizes far fewer people. And then there are all the shy avoidant young males who are gaming addicts. And hardly anyone reads books any more. My book consumption is way down too, because I read so much online, and I can sense that that's not nourishing in the same way books are. There is much less unspoiled natural beauty in the world than there was when I was in my 20's and traveled around Europe. People sneer at the humanities. I *loved* the college professors I had who'd spent their lives studying Renaissance literature. I don't know whether charging people money to take classes from them when they could be learning some great STEM stuff is defensible, but those people were important to me. I thought there would always be scholars like that. There won't be.
All that is bearable, though. But the further step, to a world where AI is important, is just too big for me. I loathe GPT4. I hate its blandness, it's prissiness, its wiki-summary world view, it's heartless helpfulness, and the vast emptiness inside it which is so easy to sense. I heartily dislike most of the AI honchos whose thoughts I've read or heard interviewed. They seem to me so deep into tech bro land that they are not playing with a full deck, and they're singing "Tomorrow belongs to me."
I'm like the older generation in Childhood's End. I can't even fully wish the AI-bonded generation well, because when I imagine them they seem so different from me that I can neither like them nor hate them.
"I'm like the older generation in Childhood's End." Interesting reference! I do tend to see (the admittedly currently purely hypothetical) ASI as like the Overmind. I tend to be sympathetic to Karellen, but tastes vary.
I'm also not happy with chatGPT, as it stands today, but I suspect a large chunk of that is from the wokescolding behavior that's been layered on top of the LLM itself, probably from politically correct RLHF - which competitors may tune differently.
I sympathize, honest; I'm 55. I do try to combat my innate risk/change aversion by deliberately looking for positives, but I do not mesh well with, ah, the current mainstream.
> they are 100% argument by assertion
This is trivially false.
> The big assertion is that AIs will necessarily, through self-improvement, become perfectly Rational and will therefore create utility functions with which they will operate entirely in accordance.
This is also trivially false.
> If you think I am wrong and Omohundro does explain in a non-circular way why the AIs will become Rational, will you explain it to me?
There's no circularity in the original papers, so I don't know why I'd expect to succeed here.
1) Rational AI's are somewhat special fixed points, in that for many of the ways AI can be irrational, it will try to fix itself.
2) It is possible to make an AI that is self consistently irrational. These AI's will do worse at whatever.
3) If the AI is very irrational, ignore it, it's not able to do much useful or dangerous stuff.
4) Perfect rationality is mathematically simple, most designs of self consistent irrationality are more complex.
5) Perfect rationality is easier to predict, and works as a good "average case" when we are unable to predict how the irrationality may influence things.
You might be right but as a neutral observer I’m not seeing an argument. Writing “this is trivially false” isn’t all that convincing, without an explanation as to why.
So here we have an example of the "if you worry about AI you are a stoopit asshole" point of view.
Sorry, are you calling me a stupid asshole, calling Hank a stupid asshole, or some third secret thing?
Needless to say I disagree with the first two possibilities. Pointing out that someone's claims about a source text don't stand up to a 5-second examination doesn't make one an asshole, and neither does making claims about a source text which don't stand up to a 5-second examination. (That doesn't even necessarily make one stupid.)
I am not calling anyone a stupid asshole, as should be evident from an attentive reading of what I said. Believe me, if I think you are a stupid asshole you will know I think so. What I am doing is implying that you are coming close to calling Hank a stupid asshole for expressing doubts about your thesis.
Seems to me that people have become rigid, irritable and intolerant around the issue of what future AI will be like and in what ways, if any, it will be dangerous. Want evidence of that? Your tone with Hank is quite irritable. Gist of what you say to him, in a fuming-with-annoyance tone, is good grief buddy, the answer to your objections is right there in the paper, why the hell don't you get it? People are like that about covid, too, of course, but in the case of covid there really is research to back up some ideas and to prove that others were nonsense.
I don't think that's the case with future AI. There are too many unknowns, so while a few ideas are pretty clearly nonsense, lots of the other ones are still in play, in my opinion.
And I still think you are sneaking in AI's being rational, and the having of private goals, inside the Trojan horse of *really smart.*. "Yes, the argument assumes some minimum level of rationality (as would be necessary for any goal-driven system capable of achieving its goals at all). " So when and how does AI become goal driven? Do you thnk GPT4 is goal-driven? If so, what's the evidence? If not, can you give a convincing argument why the having of goals will be an emergent property in systems trained the way GPT was, only with larger data sets, more determined and persistent reinforcement training, etc.? Will preferences in music also be emergent? Will emotions? Will a sense of humor? Hey, how about a taste for fine wines?
"I don't think that's the case with future AI. There are too many unknowns, so while a few ideas are pretty clearly nonsense, lots of the other ones are still in play, in my opinion."
I agree with you. I think the largest single immediate unknown is how hard the hallucination problem is. AFAIK, it is anyone's guess whether this is
a) "easy", with a decent solution already coded somewhere in OpenAI or Google, awaiting just further debugging, testing, and deployment.
b) "possible but hard" with several candidate solutions, all requiring significant enhancements to data capture during brand new training sessions, and associated high multi-million dollar gambles needed.
c) "excruciatingly hard" with LLMs being fundamentally the wrong approach, and something much closer to human learning with much closer coupling to robotics and much less text needed.
( One admission: Yeah, I'd like to see the end game happen soon enough that I live to see it, so I'm "rooting" for (a) - but of course I have zero control on this, so my preference doesn't matter. )
BTW, I have a wild guess about the "emergent" capabilities from scaling. What if it is from the expansion of the _training sets_ stumbling on "expert" text in the "emergent" fields (or particularly "digestible" text in those areas)? Handwaving argument:
If I understand correctly, LLMs are purely feedforward nets. They have frozen weights (at prompting time), and their only version of short term memory is what they have in the current prompt plus (for those with a large enough context area) as much of the current discussion as fits. It makes perfect sense that they would have the limitation that you saw in your experiments, with an inability to plan.
My handwaving guess is that when their training set is expanded to include transcripts of working sessions with e.g. someone describing their train of logical reasoning, the LLM's predict-the-next-token training can then _ape_ the appearance of the reasoning. In other words, the crystallized intelligence of the expert, reflected in the transcript, then turns into _pattern prediction_ in the LLM. The expert's "Type II" thinking turns into the LLM's "Type I" reasoning - though it _looks_ like "Type II".
If something like this is happening, then the LLM still needs to be augmented with true "Type II" reasoning - but this is essentially what classical symbolic AI was all aimed at. (Vigorous handwaving) It should be possible to take the existing body of work in automating "Type II" thinking (forward and backward chaining, theorem proving, Cyc, etc.) and wrap it around LLMs, to get both modes of thought. Does this sound plausible?
“ BTW, I have a wild guess about the "emergent" capabilities from scaling. What if it is from the expansion of the _training sets_ stumbling on "expert" text in the "emergent" fields (or particularly "digestible" text in those areas)?”
You know that’s very plausible! And I have never heard it suggested. I’m sure the tech bros working in AI are much fonder of the “emergent” view than on the “stumbled on some good info” one, because emergence is kind of gloriously spooky and exciting. The thing I’m not sure of is how much AI could benefit from stumbling on some particularly digestible text. Because it doesn’t digest text, really — not in the way we digest books, or with food — extracting the info and looking for ways to connect it with previously-learned info and filing the whole mass away. It learns the word sequences, and connects them with previously-learned sequences and files *that* mass away. I’m not sure it’s possible to directly teach it anything. (And yet, to contradict myself, if I give it complicated instructions it does seem to grasp them and follow them. So within its encounter with me it does extract the actual info in my instructions, and use the info (not the word sequences) to guide its behavior. So I dunno.)
About its hallucinating. I have a theory about that. I once inadvertently caused it to hallucinate by presenting it with a multiple choice problem with no right answer. Here’s the problem:
Here are 4 generalizations:
-All dogs are fat.
-No pets are thin.
-No thin dogs are pets.
-All thin pets are cats.
Which of the following, if it existed, would contradict all 4 generalizations? A fat cat, a thin dog, a fat pet or a thin pet?
The right answer is, I believe, a thin pet dog. But when I typed the question into GPT I left out “pet” and just typed “thin dog”, and the result of my mistake was that none of the items in the list were the right answer. None of them contradicted all 4 generalizations. So GPT picked one of the 4 answer possibilities and wrote complete illogical nonsense about how that one contradicted all 4 generalizations. That was the hallucination. (Then later I gave it the problem with the correct answer set, and it gave the correct answer and valid explanations for why that answer contradicted all 4 statements.).
But about the episode of hallucinating: It’s like GPT doesn’t have enough “sense of self” to say, “wtf? none of these answers is right.” You can think of its response as eagerness to please, or a failure to trust its own judgment after I have told it that one of the 4 choices will contradict all 4 generalizations. So I was thinking about this and the phrase “As-of Disorder” came to mind. Am I remembering right that you work or worked in psychiatry or psychology? If so, I expect that phrase is familiar. I haven’t heard ir used in a long time, but 25 or 30 years ago it was very stylish to diagnose people with As-if Disorder. And it seems to me that is a good term for the structure of GPT4. It has no opinions and preferences — it learns what to *say,* how to *sound*. Was thinking that GPT rambling on nonsensically was sort of like the moment when the therapist makes an awful discovery about the patient: They do not really grasp the idea of a dialogue, where you tell them what you think and feel and they do the same thing, and an exchange of thoughts and feelings develops. Their conception of the situation is that they are to figure out who you want them to be, and then act as is they are that person. And hallucinating is something they do when they, as usual, have no ideas of their owh, and also can’t see a convincing way to appear to be who you want them to be. So they just spout nonsense that sort of sounds coherent and hope that is good enough to get them through the awkward situation.
Anyhow, if that theory of hallucinations is right, I’d say it’s a very. big task to free GPT of that tendency. Seems like it would have to have a whole different structure.
About your last paragraph: I know very little about Type II thinking, and the attempts to train AI to do it. Can you point me to a good source of info?
> good grief buddy, the answer to your objections is right there in the paper, why the hell don't you get it?
I'm not asking him to "get it", I'm telling readers that they should spend 5 seconds checking the papers to confirm that the claims being made about it are, in fact, trivially false. I agree that I was irritated when I wrote that; I should have omitted the last sentence. Major epistemic sins don't justify minor epistemic sins in return.
> So when and how does AI become goal driven?
Who knows?
> Do you thnk GPT4 is goal-driven?
Not in the sense being discussed.
> If not, can you give a convincing argument why the having of goals will be an emergent property in systems trained the way GPT was, only with larger data sets, more determined and persistent reinforcement training, etc.?
This is not the claim under dispute, nor one that I myself am highly confident in. I _am_ highly confident that people are actively working to make AGI, which explicitly requires goal-driven behavior in that sense (since otherwise it would be unable to perform the desired functions), and have no strong opinions on what technical architectures will lead us there, or how many novel breakthroughs are required. It is obviously possible to build (humans as an existence proof), and a bunch of very smart people are trying very hard to build it. This doesn't guarantee success under any particular timeframe, but we sure do seem to be running out of unsolved capabilities.
"But we sure do seem to be running out of unsolved capabilities." Not if you include having emotions, (and emotion & motivation/goals and are so connected in our species that they are pretty much 2 sides of the same coin), being able to recognize subtle isomorphisms, inventing genuinely novel solutions, performing whatever the hell important process our brains are up to when we sleep and dream, having and learning from role models, & forming attachments. And my guess is that all of these capabilities are required to function at the level that we do.
What is his explanation. I don't see one.
Here is the table of contents for the second section in The Nature of Self Improving Artificial Intelligence:
2 Convergence To Rational Economic Behavior 4
2.1 The five stages of technology . . . . . . . . . . . . . . . . . . . . 5
2.2 Deliberative systems . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Avoiding vulnerabilities leads to rational economic behavior . . . 8
2.4 Time discounting . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.5 Instrumental goals . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.6 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.6.1 Rational approximations and proxy systems . . . . . . . . 14
2.6.2 Systems which lack knowledge . . . . . . . . . . . . . . 16
2.6.3 Reflective utility functions . . . . . . . . . . . . . . . . . 16
The first paragraph of "Avoiding vulnerabilities leads to rational economic behavior" (2.3) includes:
"But it isn’t clear a priori why self-improving agents should necessarily follow any particular set of axioms. The argument is more compelling if we can identify explicit negative consequences for a system if it fails to follow the axioms." The rest of the section then explains why avoiding vulnerabilities leads to rational economic behavior. The argument is structurally similar to https://en.wikipedia.org/wiki/Von_Neumann%E2%80%93Morgenstern_utility_theorem, which is hardly an "argument by assertion".
>The argument is more compelling if we can identify explicit negative consequences for a system if it fails to follow the axioms.
This is assuming rationality. Negative consequences are neither here nor there for an AI that is not rational. "Negative consequences" has no meaning for an AI that is not rational.
Von Neumann-Morgenstern models rational behavior. It assumes rational behavior; it doesn't argue for it. Omohundro argues for it, in AIs.
Negative consequences don't exist for an AI that has no definition of utility at all, but it doesn't follow that every AI is a perfect vNM rationalist...That's a false dichotomy.. You can be rational-ish, and in many ways, you can only be.
I have no idea what structure of argument you're looking for. Yes, the argument assumes some minimum level of rationality (as would be necessary for any goal-driven system capable of achieving its goals at all). This is not circular: the claim is that imperfectly rational agents have incentives to make themselves more rational.
But nobody is really clear whether advanced AI will be goal-driven. Seems like producing one that is what require a paradigm change. Of course we can make it goal-driven by giving it goals. Right now, I give GPT4 goals like "write a limerick." Somebody could give it larger goals, and additionally train it to generate subgoals that are appropriate, and walk itself through them step by step. And of course it is possible that in the future we could decide to give it huge goals, like "cure cancer" or " design and implement a plan for optimizing human contentment." So if you did that AI would indeed have goals and subgoals. But they would still come from us.
You don't get to *assume * that because future AI is going to be really fucking smart it will therefore be rational, in the sense of being able to read a situation and see how compatible it is with one's goals, and then design a plan for maximizing the chance one of one's goals being realized. AI now is a lot smarter than we thought it could be, and it seems to be goal-free. You don't get to sneak in "goals" and "rationality" under the umbrella of *really smart.* Seems entirely possible to me we are about to have another AI winter, as people discover the limits of what can be achieved via the present deep learning approach, even with larger data sets and more training. Could be we end up stuck with something that is 10x as good as predicting the next word, but still can't reason it's way out of a paper bag and has no goal except to sit there and respond to prompts.
Assuming AI is gonna get really smart and therefore will be able t do anything leads to paradox and nonsense, as does the idea that God can do anything: "so can he make a rock so big that even he can't pick it up?"
> Yes, the argument assumes some minimum level of rationality (as would be necessary for any goal-driven system capable of achieving its goals at all).
If by "rational" all Omohundro meant was "logical" I would see no problem here. But that's clearly not what he means. By rational he means a self-interested agent which protects vulnerabilities. He assumes that quality is present a priori.
Fun project. I created a "reverse prompt engineer-er". You can paste in any writing and it will reverse engineer an LLM prompt that could supposedly output. It's free to use. https://home.pickaxeproject.com/prompt-reverse-engineer
Does this just ask a LLM to do the prompt engineering thing? I had assumed from your description it did some sort of search or running the network backwards or something (I don't know enough about LLMs' architecture to know whether that's plausible). The response I got looked more like the sort of thing what would be the prompt used by Prompt Reverse Engineer itself than an actual response. I assume this is a bug. Do you mean it's meant to produce a piece of text which, if you prompted some LLM with it, it would likely produce the output you said?
If you're interested in finding more humanely raised meat products that are independently certified by ASPCA-recommended certifiers, check out the free Find Humane app at https://findhumane.com/
Lots of people are talking about Britain having an unprecedented economic decline, being on the path to a Third World country, etc.
I wanted to write about this, but looking at https://ourworldindata.org/grapher/gdp-per-capita-maddison?tab=chart&stackMode=relative&time=2000..latest&country=GBR~USA~DEU~FRA~ITA~SWE~ESP~CAN~AUS~NZL~JPN , British economic growth since 2000 looks completely normal for developed countries, and British economic growth since 2010 looks pretty good.
Can someone who believes Britain is in the middle of a terrible economic decline explain what they mean and what they think of the GDP stats?
I haven't read this thread in detail, but I dont think anyone posted this article yet? Claims a negative effect due to Brexit of 2-3% of GDP: https://cepr.org/voxeu/columns/impact-brexit-uk-economy-reviewing-evidence
I looked into this relatively recently and found two things at work:
1. The decline under debate is very recent, lined up chronologically with the Brexit process and the Covid recession. Your linked graph stops in 2018; instead, you probably want to look at 2015 (the year before the Brexit referendum happened) through 2022 (most recent full year).
2. There are two different ways of calculating real GDP per capita in an international context, both of which involve converting local currency to dollars and then inflation-adjusting the dollars based on the US's GDP deflator. One uses market exchange rates, while the other uses "Purchasing Power Parity", attempting to optimize the GDP figure as a proxy for standard-of-living by using local prices for equivalent goods and services as the currency conversion factor. For Brexit-related and Covid-related reasons, the relationship between PPP and market exchange rates for Britain have been highlyd unstable in the period in question: exchange rates have been very volatile (ranging from US$1.08 to US$1.40 per £1.00), and tariffs and covid disruption have both radically changed the availability and prices of imported goods.
Looking at either the PPP or market exchange rate numbers, everyone took a big hit in 2020, while Britain appears to have taken a deeper hit than France and the overall OECD average (the two control groups I picked off the top of my head). The big difference is that in market exchange rate terms, the recovery looks proportionate to the decline (i.e. Britain fell more, but also recovered proportionately faster so as to bounce back to approximately 2019 levels in 2022 the same as France and OECD). But in PPP terms, the UK has recovered at the same rate as France and OECD and thus appears to have permanently (so far) lost ground in standard of living relative to other countries. UK was also growing more slowly in PPP terms between 2015 and 2019 than France, but about the same as the OECD average.
Here are the charts I'm looking at:
PPP: https://data.worldbank.org/indicator/NY.GDP.PCAP.PP.CD?end=2022&locations=FR-GB-OE&start=2015
Market exchange rates: https://data.worldbank.org/indicator/NY.GDP.PCAP.KD?end=2022&locations=FR-GB-OE&start=2015
"Unprecedented economic decline" and "path to a Third World country" are ridiculous exaggerations (as the use of the phrase "Third World" should indicate, since that term no longer has any real meaning). But when the best you can say is "at least we're not as bad as Japan's lost decade", that's not really a good sign. As others note, the past few years have been harsher on the UK than the period in this chart. Even in this chart, the only countries as bad as the UK post-financial-crisis are Spain, France, and Italy.
I don't think that "being on the path to a third world country" applies to the economy as a whole. But economy isn't everything, and the decline of the NHS is sudden and shocking. An example where they no longer achieved first-world performance is the ambulance response time:
https://twitter.com/Quality_Watch/status/1458816998431940613
The problem is specific to UK, and is probably due to political decisions like Brexit of the last 10-20 years. That said, the extreme failures were only temporarily, and the system has stabilized a bit, though on a level that would have been unacceptable before 2020.
https://www.nuffieldtrust.org.uk/resource/ambulance-response-times
The belief in Britain's economic decline arises from an understanding of its long-standing, extractive, rentier economy. Post-WWII industrial decline and the subsequent financialisation of the economy have led to a dearth of technical jobs and over-reliance on property ownership, fuelled by planning restrictions that artificially limit supply (and in recent years, mass migration).
Most people's savings are locked in housing, a non-productive sector. This situation, exacerbated by significant migration and pension funds also being invested in this bubble, leads to chronic underinvestment elsewhere. The government is reluctant to rectify this, as any action deflating housing prices would upset voters.
This results in a precarious cycle. Profits divert to housing and legacy industries, inflating property values and concentrating spending on rent, while investment in productive sectors dwindles. The economy hinges on importing cheap labour to sustain GDP and attract foreign capital, which often funnels into real estate or financial services.
The concern, therefore, is that Britain's economy rests on unsustainable pillars: a housing bubble, foreign labour, and capital inflow, rather than robust domestic investment and innovation. There is a strong argument to be made that this situation is the result of an inevitable downward spiral since ~1900, when first the aristocracy, and then the government post-nationalisation, decided to sit on industries rather than refine, grow, invest. And then the big sell off and shut down of industry in the 1980s was chosen rather than trying to protect and modernise what we had left. I suspect the collapse of human capital after WW1 exacerbated this issue.
And not to be too doomerist about the whole situation, but the feel on the ground is certainly one of general malaise. Even in London there is a general decline- property used to be expensive, it is now exorbitant, rents were kinda bad, they are now catching up to property prices. The job market is ok, but is entirely focused on professional services, with a small tech sector and a huge swathe of service jobs. Leave London and the feeling is one of awful decline since the mid 2000s boom, with Manchester perhaps an exception.
Something I think illustrates the situation in Modern Britain well- over the last 5-10 years, many UK cities skylines, previously remarkably low, with the odd office building or council tower block, have filled with swathes of garish student housing blocks. The remarkable growth of the Leeds skyline, now filled with cheap student accommodation, sums up the problems with our current capital allocation- billions spent on building non-permanent, suitable only for 18-21 year olds, modern tenement blocks, with the aim of growing our 'university sector'- dozens of, putting it politely, 'second-rate' institutions, which rely on farming out expensive masters degrees to international students and pile-em-high style courses for British students.
This comment is persuasive and well-written. If you're verified to be right I hope this goes in the "best of" follow-up post. Of course, you could be completely wrong about everything -- I have no way to tell. Sounds spot-on to me, as you Brits might say.
I appreciate the compliment, but yes, I have provided no data or evidence for the claims above, so certainly wouldn't expect to be 'best of-ed'. The core elements- Britain underinvests in R&D/Capital, choosing instead to plough it into a housing bubble which stifles young mobility and talent, is well trodden ground. I hope people who are familiar first-hand with the territory nod as they read it and say 'that's the vibe'.
Lol told you.
1) Some of this is anti Brexit propaganda. I think Brexit is a folly myself, but it’s clear there is a clear attempt to blame it for everything and even to exaggerate it’s affects.
2) that said GDP per capita isn’t everything. Wages gave stagnated in real terms, especially since 2008. They were bouncing back a bit until the recent inflationary period.
3) Britain seems to handled inflation reduction very badly. I might contradict myself here by saying Brexit could have a roll there, but if so why now?
I was thinking of replying until I read this. I agree, except the part about it being a folly. My understanding is it was really a question of sovereignty, not economics, such as that they didn't want to be forced to take immigrants from the rest of Europe they were finding disruptive. https://en.wikipedia.org/wiki/Causes_of_the_vote_in_favour_of_Brexit
As a sovereign nation, economics ought indeed to take a back-seat to governing your own people. Long-term, their economy will recover.
It's not like there aren't ample precedents for mass migration from the continent effectively displacing the culture of the island...
The Welsh can tell you how happy they are about that. I just learned the other day that within living memory they were burning down English vacation homes in Wales in reprisal.
As a Brit I’m not sure I would say we are becoming a third world country, but the last 10/15 years have been pretty bad. You can see this in the gdp per capita stats (link at the bottom, 2000s we were similar to Germany, now we are similar to France), public services (which have gone from being fairly decent to completely dysfunctional, try getting a non life threatening illness dealt with or any response from the police) and house prices now 9x average earnings, leaving many young people to be paying a very high share of their income of rent.
I think particularly for the smart/driven things have changed from a sense of you can achieve your dreams here (in the same way you can in US or maybe Germany and parts of east Asia) to the sense it’s more like much of Europe where you have to leave to make something of yourself (becoming a France or Italy, full of old people on massive pensions with little growth or dynamism). In the 2000s you could make ridiculous amounts of cash in finance if you worked hard, now it’s a pretty bad deal. Most other industries you would be much better off in the US either as an employee or setting up your own company (e.g. McKinsey associate in London earns £95k, in New York earns $185k with bigger bonus). Hard to think of many successful UK companies that have been created in recent years (Deliveroo, Sky, Ocado?). When I try to explain to american friends that £35k is considered an enviable salary when graduating university they all assume I’m joking. I think the average salary of an Oxford undergrad five years after graduating is around £50k.
Personally I’m somewhat optimistic, partly because the last few years we’ve been paying the costs of brexit (now shouldn’t be as much of a drag, and might even be some benefits from being outside EU regulatory nonsense), a lot of the UK industries have some prospects for growth (tech, pharma) and others unlikely to massively decline (law, finance, insurance). Ultimately think British political elite will fix it as they’ve done in the past, can take them a while but typically they ultimately take the difficult decisions.
https://data.worldbank.org/indicator/NY.GDP.PCAP.PP.CD?locations=GB-US-DE-FR&start=2002
That only goes up to 2018, so before Brexit, Covid, and the recent inflation and interest rate hikes. Obviously the last two affected other countries too, but all three have interacted to cause a lot of increases in food prices (as many foods are now more expensive to import) and mortgage repayments (which I think has thrown a lot of people, especially younger borrowers who had never previously seen rates above about 3%).
The NHS is noticeably under strain (I think all my lifetime you could get a same-day doctor appointment, and now it's usually a couple of weeks; I've heard (from friends, not just newspapers) that hospitals are similarly struggling). This may be partly because of European doctors leaving after Brexit.
There are a lot of strikes, which we haven't had significantly since the 70s: NHS strikes, school strikes, rail strikes, postal strikes, etc. This adds to the subjective impression that things are starting to fall apart.
"That only goes up to 2018, so before Brexit, Covid, and the recent inflation and interest rate hikes. Obviously the last two affected other countries too."
For Covid yes, but the UK was hit harder by the 2022 energy crisis than other Western European countries, because they were relying more heavily on gas as energy source.
https://www.theguardian.com/money/2022/sep/01/energy-crisis-uk-households-worst-hit-in-western-europe-finds-imf
And privatised our fossil fuel resources, and got rid of our gas storage facility.
Looking here: https://www.imf.org/en/Countries/GBR#countrydata UK has basically the same GDP it had in 2007, and a significantly lower GDP per capita (toggle 'select indicator' button).
Contrast that with the EU, for example, which apparently at least grew moderately during that period, see e.g. here: https://www.macrotrends.net/countries/EUU/european-union/gdp-gross-domestic-product, let alone the US, which grew quite a bit during that period: https://www.macrotrends.net/countries/USA/united-states/gdp-gross-domestic-product.
Your direct link shows yearly gdp per capita growth, not total. But you can edit it.
But you are right. The per capita gdp seems to have maxed out 15 years ago at around $50k. To put that in perspective the previous decade saw it grow from $30k or so.
If you mean the IMF page, I was referring to the actual GDP there which, as I said, you can see by toggling the "Select an Indicator" button and selecting "GDP, current prices (Billions of U.S. dollars)."
It shows GDP rising until 2007 when it reached 3.09 T and then bouncing around until the present when it is at the nearly identical 3.16 T.
Yea, i think that per capita gdp is more useful as the size of population has increased. Even more useful would be wages. For instance GDP might increase this year (and definitely did last) but wages have not.
Mallard, I’ll be honest with you. I didn’t think my comment largely agreeing with you would cause this level of hostility.
It’s a pity that Substack can’t stop emails from sub threads.
Adam Tooze wrote a piece about the decline:
https://adamtooze.substack.com/p/chartbook-184-nostalgia-for-decline
On most issues to do with the global economy, it would usually be reasonable to trust media institutions like the Economist and the Financial Times; and if they have a settled view about something, it makes sense that that view would come to be seen as representing reality.
The problem is that Brexit was personally, agonisingly painful for them, and the wounds haven't healed after seven years. I listened to an interview with the long-term editor of the Economist shortly after Brexit, and it felt like she was discussing the death of a member of her family. That depth of feeling doesn't just go away; and it hasn't.
So you have otherwise sensible, reliable British journalists credulously spreading nonsense that in other contexts they would see through instantly - one widely used model that purports to demonstrate British decline post-2016 compares the UK to a basket of other countries, 45% of which by weight is made up of the US and New Zealand.
The truth is that Western Europe is in long term relative decline. And that (as you've argued before) you can trust the mainstream media on nearly everything - but you have to know when you absolutely shouldn't.
(1) That graph ends in 2018, which makes it almost entirely useless. (All three articles you posted below are from 2022-2023, ie after COVID and during inflation.) (Better data doesn’t disagree, to be clear.)
(2) Inter-country comparisons aren’t too useful here, since AFAICT, pretty much every developed country currently is having economic issues around periods of declining GDP and inflation. At best, they can establish that other countries are also having similar difficulties.
(3) The US has had among the best developed-country economies during and after COVID; Europe in general was the worst affected by the Russian invasion, and so is worse off than the rest of the world; the UK speaks English, and so is most exposed to the US media, even if it’s not worse off than the rest of Europe.
Britain leaving the EU has somewhat slowed economic growth. While some British nationalists deny this even the conservative UK government admits this and has sought free trade agreements with several East Asian countries to offset this. However, pro-EU types have a political incentive to exaggerate the damage because they want Britain to rejoin the EU. Which probably would genuinely have a positive economic effect but not to the degree either Brexiteers or Europhiles want to pretend it does.
Further complicating this narrative is the fact that Europe in general is experiencing economic stagnation. This is not just a relative loss like the US where the US is roughly standing in place in terms of relative economic strength but competitors like China are catching up. It's declining in importance globally as its economy doesn't grow significantly while East Asia grows a huge amount. The "smart" Brexiteer case is that independence from the stagnant EU is allowing them to pursue trade deals with the more dynamic east and the US which are ultimately better.
As I see it, Britain's bet is that being a junior partner of the United States will both give them more success than chaining themselves to Europe and give them more freedom (since the US has no interest in British domestic politics in the way the EU does). To this end, it's aligning itself with other "individual" US allies like Japan or Thailand and attempting to take a leading role in things like Ukraine or Pacific defense.
I have no idea whether it will work (if I did I would be quite rich). But they are right that the EU has not really dealt with the degree to which they are experiencing economic decline. They really ought to be taking relatively radical action. Though whether Brexit as a specifically radical action is a good idea... well, it's too early to tell.
> since the US has no interest in British domestic politics in the way the EU does)
The EU has no interest in British foreign policy the way the US does. Somehow, we always end up helping with whatever war the US wants to fight.
The idea the EU doesn't conduct or affect foreign policy is laughably wrong.
Likewise the fact the UK has gone out ahead of the US in agreeing to help defend its own East Asian allies or Ukraine points out that Britain has plenty of independence. Of course, the British left is not immune to the siren song of blaming foreigners (in this case Americans) for domestic disagreement.
> (since the US has no interest in British domestic politics in the way the EU does)
The obvious counter example to this is Northern Ireland, where President Biden has shown incredible willingness to intervene. Many people in the UK consider this a significant overreach.
That's because we nefarious Paddies are getting the Yanks who like to claim the Ould Sod heritage to intervene, since they're the 800lb gorilla in the 'special relationship' no matter what the Brits like to think, because we don't trust you Brits.
https://arethebritsatitagain.org/
Now fair enough, I might feel like burning a picture of Leo Varadkar myself, but not in the context of "hey let's go back to the good old days of bombs and bullets":
https://www.bbc.com/news/uk-northern-ireland-66149441
Firstly, the US mediation in the UK-Irish disputes is international (and was requested by Britain and Ireland originally). The EU tried to somewhat take over but then used it as a Brexit chit so now it's back to the US. Secondly, the idea that that's anywhere near comparable to the amount of domestic influence the EU had is ridiculous.
Razib Khan interviewed a youngish British person and they both took the decline as written for the sake of discussing it. My takeaways were that salaries are low and there are few technical jobs, and that smart people can either leave the country or go into banking.
Even banking isn't that great; I have a friend who was an investment banker and moved to London for a job offer (thinking the promotion opportunities would be better there than in Australia), only to realise rather belatedly that the pay was slightly *lower* in London and the cost of living much higher - he was flat sharing with 2 other Investment Bankers, which IMO is rather damning about living costs; how is any young professional meant to afford rent when one of the best paying jobs possible is only enough to afford a flat-share?!?
A friend of mine (in the US) has been in a long distance relationship with an English guy for many years. They’re engaged and he will be moving to the US soon. He recently visited and I met with them for lunch after not seeing him for about 5 years. The couple are your stereotypical progressive lefties who don’t follow politics very closely. It was fascinating to hear him talk about issues England is facing, because he was pretty much confirming things that 5 years ago were considered far right talking points. He specifically seemed concerned about the quality of doctors and nurses which they’ve brought in from other countries, along with general concerns about immigration, wages, etc.
World bank has to 2022 as well: https://data.worldbank.org/indicator/NY.GDP.PCAP.CD?locations=GB-DE-US-FR-SE-IT-CA-NZ-AU-JP
What extraordinary there is US growth since 2007 relative to peer countries (in GDP per capita). Some of this is currency movement though, as it’s measured in dollars. Nevertheless it’s impressive.
There's a button on the sidebar that can change the view to PPP, and yes, it is much better than I expected from general media portrayals (and not just in the US).
That's not the claim I've heard. The claim I've heard is that, since 1914, Britain has declined from being the dominant economic power in the world to a middling European economic power –– which is consistent with the data you cite. Although to a citizen of Great Britain with a historical view stretching back to 1914, this might well seem like "terrible economic decline."
Long slow decline is compatible with short, sharp extra-decline.
See for example https://www.theatlantic.com/newsletters/archive/2022/10/uk-economy-disaster-degrowth-brexit/671847/,
https://foreignpolicy.com/2023/02/03/britain-worse-off-1970s/,
https://mainlymacro.blogspot.com/2022/02/we-are-in-unprecedented-era-of-uk.html
maybe the malaise is pan European, just British malaise is more visible to us because of the common language?
Noah Smith had a recent post arguing that people's opinions about economic decline are better interpreted as opinions about how they are personally being affected by cost of living rises.
I think a lot of it is people being sane about brexit.
>Threads is supposed to be just like Twitter except better in a few aspects
I haven't used Threads, but from what I've heard, there is no desktop website to use it on and your feed is entirely algorithm-controlled with no option to only get the post of people you follow. To me, that's not "better in a few aspects", that's "worse in a few deal-breaking aspects".
I'm pretty sure most of the people leaving twitter, or contemplating leaving twitter, have grievances that go *far* beyond "not liking the designated hitter rule". A fair number of them claim that the "home team" is now run by a Literal Nazi, who is stacking the deck such that his Nazi pals will always win and will violently harass people like themselves. Most are somewhat less histrionic in their terminology, and I think all of them overstate the real threat, but the nature of their objections is the same.
If they stick with twitter, believing that, they're chumps. And some of them will be chumps. But not all of them, and maybe not a Threads' worth of them.
I think the idea here is that those people don't literally believe there are Nazis running Twitter. They just *say* that, the same way people will say "I will not rest until the real killer is found" and still go to bed in the evening. Virtually no one is hiding Jews in their cellar or desperately applying for immigration.
If I had to guess, I'd say the most genuine motivation at actual play here is a ploy for good old fashioned attention.
I thought this post was going to be about reddit and its various floundering clones at first.
I'm skeptical about threads as well, but Zuck has a history of building social networks that obsolete others. Facebook ate Myspace, and Instagram ate Facebook when Facebook become the thing for uncool olds (and Zuck bought it). Now twitter is flailing, and Zuck is trying for a three-in-a-row.
Also, keep in mind that threads isn't a truly new platform that you have to abandon your friends for, it's something that all of your Instagram friends are going to be inevitably pop-up-bombed to join the moment the initial hype and exclusivity is no longer in place.
On Facebook buying Instagram, it doesn't require much talent or skill to exchange a bunch of money for an effective product someone else created.
While there's some truth in that, I think you're underestimating just how successful Instagram has been under Zuck's ownership - specifically with gen Z. Facebook was *the* social network for millennials, and then they turned around and did it a *second time* with Instagram - basically everyone in Gen Z has an Instagram account, without ever having made a Facebook account. This isn't just them transferring network effects, they managed to fully capture a wholly separate demographic.
And they did this all with an app that in 2012 (when they bought it) was mostly a niche photo-sharing app, that they managed to turn it into the *default* social media app for a whole generation.
Go back in time and look at Friendster, followed by MySpace, followed by Facebook. In each case, it just took a bit of technical superiority for the network effect to come to attach to the new network. In the case of Twitter vs Threads, Threads starts with the superior network effect of Instagram, so it might not need much technical superiority.
It's harder to imagine both coexisting for a long time, because network effects will likely drag people to one or the other, unlike with the kinds of markets where long-term competition can be stable.
Threads isn't just an alternative to Twitter with some better features. It's an alternative to Twitter at a time when Twitter is in crisis. Imagine if the NFL had a Bud Light level PR crisis, and the XFL came around at that exact moment. Then, I could see the XFL succeeding.
The only problem that I have with Threads is that the name invokes involuntary memories of the deeply disturbing nuclear war docudrama that the BBC broadcast back in the 1980s.
>Twitter is in crisis
Wasn't there supposed to a mass exodus of users that didn't go through the formality of actually happening?
Aren't the people telling us how Threads will kill Twitter the same ones who told us Mastadon would kill Twitter?
>and added sudden-death overtime<
I'm trying to imagine what Sudden Death Baseball looks like, and I've got the image of the pitchers from both teams pitching to batters from both teams; we turn second base into one team's home plate, and the other team fields the foul section, and then both teams pitch and hit simultaneously.
...This should probably be the standard inning for Baseball 2.
I still don't know why Facebook beat out Myspace. But I'll say, very few people give up baseball because of the people in the audience, whereas that's the biggest complaint I hear about Twitter. One advantage could be that the worst trolls stay on Twitter where they're comfortable, so you don't have to deal with them on Threads.
Here is what I wrote about this topic a couple years back
MySpace is one exemplar, and nearly decade on, the lessons are still not well understood. I’m old enough to remember just how buggy the site was. Error rates were as high as 30 or 40 percent at times. In contrast, commercial operations like Yahoo! or Salesforce had at most 1 percent at the time. Simply put, MySpace wasn't created with the kind of systematic approach to computer engineering that went into Yahoo, eBay, Google, and Facebook. Instead, the site slogged through upgrades to implement flexible data caching and the distributed architecture needed to scale. Here is how one article explained it at the time:
Every profile page view on MySpace has to be created dynamically—that is, stitched together from database lookups. In fact, because each profile page includes links to those of the user's friends, the Web site software has to pull together information from multiple tables in multiple databases on multiple servers. The database workload can be mitigated somewhat by caching data in memory, but this scheme has to account for constant changes to the underlying data…
And although the systems architecture has been relatively stable since the Web site crossed the 7 million account mark in early 2005, MySpace continues to knock up against limits such as the number of simultaneous connections supported by SQL Server, Benedetto says: "We've maxed out pretty much everything."
For those that haven’t designed a system to freely scale, and indeed many commentators in the tech regulatory space haven’t, it’s hard to impress just how difficult of a problem engineers face. This is what unites the biggest players in the space, the ability to innovate to solve problems at scale without interruptions. How each site has done it serves as a case study in its own right. Todd Hoff’s site High Scalability collects these ideas. For example, here is how Twitter solved its issues with scaling. Here is how Facebook did it. Here is how eBay did it. As for Google, well, the comapany’s solutions are legendary in the computing world.
"[...] the Web site software has to pull together information from multiple tables in multiple databases on multiple servers."
How else can it be done? Query only one table on one database on one server? I can't imagine that it would be better.
Thanks for the pointer (Todd Hoff). Could you summarize some of the tricks? This sounds super interesting to me.
Here are the raw links:
* http://highscalability.com/scaling-twitter-making-twitter-10000-percent-faster
* http://web.archive.org/web/20130927072116/https://fbcdn-dragon-a.akamaihd.net/hphotos-ak-ash3/851578_534669043269870_1216439060_n.pdf
* https://www.cs.mcgill.ca/~mahmed26/eBay_Architecture_Study.pdf
* https://www.wired.com/2012/08/google-as-xerox-parc/
Thanks a lot! I had already discovered highscability.com, I can't wait to look at the others.
My idea for Baseball 2.0 is that everyone but the batter, pitcher, and catcher are on horseback. In case the batter hits the ball into play, his horse is waiting for him next to first base.
Horses are fast, but they aren't as maneuverable as people. I think this would make it harder to catch the ball. And it would most certainly make it harder to pick up the ball off of the ground. And what happens when a horse gets hit by the ball? Or a runner's horse slams into a baseman's horse? (Come to think of it, what happens when a polo horse gets hit by the ball?)
"Getting hit by the ball" was a bit part of why I'd excluded the batter and catcher. I thought of also excluding the infielders, too, but then they'd be at a critical disadvantage in both speed and collision-survival ability to the base-riders. I'm not too worried about horse-on-horse collisions: horses tend to be more sensible than humans about barreling into one another at speed (i.e. they'll slow down and try to get out of one another's way if they notice a collision is imminent: a lot of historical anti-cavalry tactics rely on this kind of horse-sense), and polo horses rarely seem to get seriously injured by collisions (with a quick googling just now, I found a 2015 veterinary journal article that concludes that polo horses have a similar injury rate to the general horse population). I'd expect horse-baseball to be somewhat less dangerous in this respect than polo, for similar reasons to why regular baseball has fewer collision injuries than soccer.
Good point about picking the ball off the ground. Maybe give the fielders polo nets? Maneuverability is another good point, at least for infielders, although for outfielders speed is probably more important. Enlarging the playing field somewhat, especially the infield, seems like it'd give the horses' speed to be more of an advantage and their maneuverability less of a disadvantage. Positioning could also change to emphasize speed over maneuverability, with infielders (especially the middle infielders) taking positions further back to get more time to react before the ball passes them, then charging inwards to field pop flies and slow grounders.
Getting hit by the ball doesn't seem to be a big risk to the horses in polo, either, which is surprising since it seems that polo balls go as fast or faster than well-struck baseballs. I think the difference is probably the size and composition of the balls. Outdoor polo balls are 3-3.5 inches in diameter, in between the sizes of baseballs (9 inch circumference -> 2.9 inch diameter) and fastpitch softballs (11 inch circumference -> 3.5 inch diameter) and are lighter than either (~4 ounces for a polo ball, 5 for a baseball, or ~6 for softball). Tweaking the ball design for safety would probably be tricky, especially given my previous suggestion of enlarging the field, since a longer and lighter ball probably wouldn't carry as far from being batted or thrown due to being more susceptible to air resistance. Maybe outfit the horses with safety armor if necessary?
There will be less trolls on Threads?
>You can pray that something (more) catastrophic happens to Twitter and its users don't realize that they've actually been freed and rush to get into another prison, I suppose. Maybe the login wall does it, but that's a risky gamble at best.
Social media companies have catastrophic things happen to them all the time. Reddit was basically "Digg 2" for a while, but then Digg pissed off their user base and suddenly Digg 2 was the best link aggregator in town. If there was some sort of massive scandal in MLB and we discovered that the World Series was as fake as pro wrestling, "Baseball 2" would probably get a lot more viewers.
Twitter has been slowly combusting for a while, but there wasn't an obvious alternative for people to converge on - Mastodon is confusing, Bluesky is still in development, Truth Social is full of nutjobs, etc. Threads has the potential to be that rallying point since it's run by a well-known company that already has a massive user base.
The login wall is the dumbest thing Elon has done. Maybe he thinks we will all sign up to read a tweet, but by hiding tweets twitter moves from the town square to a private club.
The login wall existed before Elon bought Twitter.
The limit on number of readable posts was probably dumber, but it seems to have been ephemeral. (So far.)
Having seen several Twitter-killers over the last year or so, I'll have to see it happen to be convinced that the next one is going to do it. But I wouldn't be surprised if Twitter is in its decline phase whether or not there's a New Twitter to go to. Social networks seem to have their cycles, whether they actually shut down or just go from the hot spot to the place young people only log onto to wish their grandparents happy birthday.
Rather than baseball, the analogy I would use is phones. Imagine someone started a competing telephone network, with the limitation that the new network only lets you call people who are also on the new network.
Like phones, social media is locked in by network effects. It's inherently more valuable the more people who are already there (insofar as you value it at all). In order to compete, you need a major disruptive innovation or some way to quickly acquire a similarly-enormous user base.
If you wanted to have competition in social networks, I think you'd need to do something analogous to phones where the networks are forced to be intercompatible so that someone using phone provider A can call someone using phone provider B. Which would probably be a technical nightmare, and which obviously the current dominant social networks wouldn't go along with willingly even if the technical obstacles were removed. (Currently it's illegal to write your own user interface for someone else's social network.)
This is more-or-less how MCI started. MCI originally stood for "Microwave Communications, Inc": they provided a network of narrow-beam microwave radio relay stations in order to allow trucking dispatchers to communicate beyond the reach of traditional two-way radios. They soon branched out into providing business-to-business voice and data communication between fixed locations via the same network.
They didn't get really big until some time later, when three things happened in succession. They added satellite and fiber optic lines to their ground-based radio network, giving more cost-effective nationwide coverage and better signal quality. They integrated with the telephone switching network so you could use them to make long-distance phone calls through regular Bell System telephones without needing the recipient to be a subscriber (i.e. you'd make a local call to an MCI endpoint, they'd send it over their network to another MCI endpoint, and from there they'd make a local call to whoever you wanted to talk to). And of course, they won a big regulatory and legal struggle with AT&T that concluded with AT&T divesting its control of the local Bell phone companies and those companies being required to let their customers choose whether their outgoing long-distance call would be routed through AT&T's or MCI's networks.
In that case at least, network effects were a big enough deal to keep MCI from getting AT&T-level big until they managed to first route around them and then to be given full access to the Bell local phone networks by regulatory fiat. But even before that, MCI had been able to run a good-sized and profitable business making a parallel competing phone network.
Well, actually, the Powers That Be in international sport governance did invent "Baseball5" a few years ago for more-or-less these reasons.
There was also the infamous XFL of 2001, begun as a competitor to the NFL. That crashed pretty fast. The idea never died, though, and it came back in 2020, which, like other things in 2020, showed the Best Timing Ever for starting a new major venture. It took some major investment from The Rock to save it from a second death.
From what I'm hearing, it's still experiencing what could be plausibly described as "growing pains".
Also the AFL goes in and out of existence, and gave us Kurt Warner
I thought you were joking - surely 'Baseball5' isn't real - but... https://en.wikipedia.org/wiki/Baseball5
> The WBSC invented B5 in part to increase the odds that a sport similar to baseball and softball will be played in the Olympics,[26] with B5 meant to contribute to an ultimate goal of having a billion fans in the baseball-softball community by 2030[27][28][29] and helping to demonstrate the global reach of the games.[30] It also intended to create a game that would be more accessible and cheap,[9] as the only equipment used in B5 is the rubber ball,[11][31] and the field is much smaller than a baseball diamond or softball field,[32] with the game being playable indoors or outdoors,[33] even during the winter.[6]...In addition, the game is meant to be simpler to learn,[40] have more youth appeal,[41] and be more exciting, with players participating more frequently throughout the game.[35] Another claimed benefit is that with B5 being introduced across the entire world at the same time, no one country is likely to dominate the game;[42] the lack of the pitcher role, which is highly specialized, contributes to this.[32]
Banana Ball also seems to be fairly popular as spinoff sports go
https://en.wikipedia.org/wiki/Savannah_Bananas
There’s nothing quite like Finland’s PESÄPALLO though.
https://m.youtube.com/watch?v=Pz19WhYxaxA
Baseball 5 sounds like the kind of thing kids growing up in boring small towns invent after they tire of playing Jump the Weeds.
Jhana twitter has nice overlap with tpot
Tpot has decided to practice being disagreeable so it’s not the best time lol
Nick Cammarata jhana-pilled everyone like two years ago. Scott wrote about it at some point but running the below search in the Twitter UI is maybe a good start:
"jhana (from:nickcammarata)"
As an American currently living in Canada for reasons beyond my control (I'll be back in a few years) the America-bashing I hear from progressive family members drives me crazy. There's a reason why the Canada-->US per capital immigration rate is 10-20x that of the US-->Canada direction. If you want a slower easier life, sure move to Canada. If you want a career, money, technology, excitement, interesting people, network effects, etc, etc, then move to the US.
I have never heard a more appealing advertisement for Canada than that one. Sounds amazing.
Are you American and have a college degree and decent skills? You can move and get permanent residency, no problem.
Or are you, like many Americans, not actually serious about moving to Canada because you do in fact realize that your own country has more to offer?
Certainly the latter.
and Tim Horton's is overrated
Tim Hortons is terrible. The coffee is a crime against humanity.
Agreed on the coffee. The timbits otoh...
Okay, that second one should be catnip to me, though I really would prefer a transcript rather than listening to a lecture.
However, I'm just moments in and already I'm going "Well duh, the reason it was popular among minor clergy was (1) they were the literate ones and this was not a grassroots tradition, this was very heavily based on scholarship needing some knowledge of Latin, Greek and Hebrew and (2) the reason it was so uniform was *because* it was based on scholarship and to an increasing extent, lack of knowledge of original languages - you see this in the corruption of Hebrew terms over time as people are just copying without understanding."
I realise the guy is ten times more educated than I am and has genuine academic degrees, but I'm always very wary when someone starts off with "Gosh, whatever could be the secret hidden reason for Blinkin' Obvious Thing?"
I'll have to wait for further into this to see what he says about exorcism. I'm not holding my breath, though.
EDIT: Yeah, I'm five minutes in, and I'm not impressed. I should definitely refrain from commenting on him being a goyische convert to Judaism, but the winking and smirking at your audience about the Christianity you left behind you is not inducing me to keep watching.
I think there's the guts of an interesting topic there, but I honestly would prefer an atheist scholar to deliver it.