Okay, well...thoughts on the Biden administration's just-announced-today student loan cancellation and forgiveness?
*Risky gamble for midterms, there was some real solid Sound Policy Momentum building for a minute there.
*Exacerbates deficit and also inflation(?), largely regressive, most college-educated already lean D (including didn't-graduates in this category).
*Fairness principles, as with all means-tested benefits; plus larger concerns of those who already paid back loans, and thus partially helped pay for this bailout.
*Time discount: yes, covid + Ukraine has been an outsize destabilizing double punch, but all Future Student Debtors are left out of this jubilee. Proposed ongoing Dept. of Ed reforms will eventually equal the magnitude of front-loaded relief, but it's quite a disproportionate situation for anyone currently existing who isn't *quite* college-age yet.
*Money is fungible, this could have been spent on many more higher-ROI things. Like pandemic preparedness or international relations/foreign policy improvement to help avert getting into future situations like the current one.
As someone suddenly relatively much richer and having a positive net worth for the first time ever, I'll take the money self-interestedly, but...even most-charitably framed as some sort of unique economic-justice response (recompense for Millennials who graduated into the Great Recession), it still seems like an inefficient option within that possibility-space. There are so many worse-off and more-deserving than student debtors...and a majority of Americans never went to college, still. I don't know. This does not feel like The Way. I could be wrong though.
A very nice little history of polyhedral dice, but it also caused me to think about redundancy and roughage ratings for videos and texts.
Redundancy has an obvious meaning-- roughage is things like small talk between the hosts.
Some videos are very efficient and I appreciate it, some have a little personalization-- Answers with Joe is at a good level of that for me. Some have what I consider to be agonizing levels of repetition.
And then there's repetition between videos. Is there anyone who talks *about* the Fermi Paradox without explaining it yet again?
For people who follow AI more closely, why is speech generation progressing so much slower than image generation? It feels like it should be much easier to put together a synthetic voice that sounds natural than generate a realistic photo, but at least on the public-facing side of things this doesn't seem to be the case. What's the reason behind this, assuming that it's actually the case?
What if Trump removed the classified documents from the White House on January 20, 2021 and took them to Mar-a-Lago, the National Archives informed him that the action was illegal shortly thereafter, and then Trump promptly returned them to the Archives so they were only in his possession for, say, two months? Would he still be in trouble right now?
I am glad to announce the second of a continuing series of Orange County ACX/LW meetups. Meeting this Saturday and most Saturdays. Note this week we will meet at 3:30 not 2. The first meeting was great, and I hope to see many of you at this one. Based on the first meeting, I chose two popular topics to prompt future conversation and activities.
This week it will be at 3:30 (usually 2:00) to avoid a conflict with an online LW meetup
Activities (all activities are optional)
A) Two conversation starter topics this week will be. (readings at the end)
1) Forecasting and predicting the future
2) Psychedelics.
B) We will also have the card game Predictably Irrational. Feel free to bring your own favorite games or distractions.
C) There will be opportunities to go for a walk and talk about an hour after the meeting starts and use some gas barbeques if anyone wants to grill something. There are two easy-access mini-malls nearby with takeout hot food available.
D) Share a surprise! Tell the group about something that happened that was unexpected or changed the way you look at the universe.
E) Make a prediction and give a probability and end condition.
F) Contribute ideas to the future direction of the group. Topics, types of meetings, activities, etc…
Conversation Starter Readings:
Suggested readings for this week are these summaries. These readings are optional, but if you do them, think about what you find interesting, surprising, useful, questionable, vexing, or exciting.
1) Prediction
Superforcasting is a review of experiments done about how well various types of experts do in predicting the future. Generalists tend to do better than specialists in prediction, but why? Groups tend to do better than individuals, but are there ways to improve the performance of groups even further? How can you train yourself to be better at prediction? How can you help others?
The tale of two receptors is an interesting speculation as to the underlying pharmacology of psychedelics. Hypothesizing that serotonin can both help us cope with the distress of a bad situation and help us look for creative ideas out of the old mental habits that can keep us trapped in a bad situation. Conventional antidepressants and atypical antipsychotics operate on the acceptance system, while psychedelics operate on the later, lateral thinking. Is this perspective useful? Oversimplified? Can it be made into a rigorous scientific idea, or is it just another evolutionary just-so story? What possible uses and hazards does this suggest for psychedelics?
If you want a more general introduction to psychedelics, here is a book summary of the recent popular review of psychedelics. “ How to change your mind” by Michael Pollen
I've been hunting around for datasets regarding stock-prices and commodities prices, but this is a very new space for me. Does anyone have any suggestions? It seems like there's nothing available that's open source.
On Thursday 8/25 night starting at 8:00pm eastern, 7:00pm central The St. Louis Rationality Group will have an online discussion of the ACX Book Review Contest in GatherTown.
We are inviting all other ACX/Rationality readers as well.
If you haven't read all the book reviews, still come.
If you are nervous about talking to people because you find old age makes you less social, still come. If you will be late due to time zones or other obligations, still come.
If you feel awkward, still come.
It'll be fun.
Here is the link to the event, click it for more information:
Let's talk ACX Book Reviews. You need not have read every book review to be part of this event, as long as you have read three or so you are good. The purpose here is to meet and learn from each other and share thoughts about this year's book reviews.
Wander from room to room and group to group and talk! No rules about group size or how long you have to stay within a group.
Here are some questions to ask each other:
Do you like or dislike the book review contest?
What makes a good book review?
Are there any books that you have read because of a great book review?
Which book reviews from this year's contest were your favorite?
Did you learn anything useful or insightful from any of the book reviews?
Is book reviewing an art?
What's the best book you've read in the past two years?
I think there's too much focus on trying to design utopian cities, and there should be more focus on trying to design utopian towns first. Get things right at a small scale first, before scaling up to a megacity. If you can take a greenfield site and a budget in the single-digit billions and build a medium-sized town of twenty thousand people, and persuade twenty thousand people that they want to live there because your urban design is so wonderful, then I might believe you can do it with a larger city.
Here's my idea for a town of 20,000. I've tried to cut a middle path between standard urban design and the anti-car fundamentalism of so many "urbanist" type thinkers.
- The whole design looks like a scaled-down Adelaide (check it out) with a downtown core surrounded by a ring of parks surrounded by a ring of suburbs.
- Downtown core is dense and highly walkable. A few major streets and a bunch of narrow lanes. In general there's no parking within downtown itself, but there's space for delivery trucks etc to keep the shops supplied. A narrow river running through the centre of town would be nice.
- Next, the ring of parks, which hide underground car parking beneath. If you want to go downtown, you can park underground and walk. Supermarkets etc can be placed at the outer edges of downtown close to the carpark entrances to make grocery shopping convenient.
- Immediately outside the ring of parks, medium-density apartments and terrace housing, gradually turning into quarter-acre blocks as you go further out. Some major arterial roads head into the residential areas, but there's also tendrils of parkland with bicycle paths in them. A bunch of corner stores and cafes are sprinkled throughout the suburbs. Street layout isn't a grid, nor the cul-de-sac maze of most modern suburban developments, but some kind of haphazard mishmash of the two.
All the comment threads in this post are hidden behind “Continue Thread” links, which are wildly inconvenient to read on a mobile device: is it one uninteresting reply, or a major subthread of interesting opinions from the most prominent bloggers who follow scott? No idea, lets click on 543 slow-loading links one by one to find out!
I miss static html, and dearly wish for substack and similar single-page-application monstrosities to all shutdown, their leaders discredited, their investors bankrupted, and their developers unhireable.
I've been re-reading Meditations on Moloch. The first time I read it I remember being very struck by how profound it seemed, but looking over it again, I think its main thesis is just that "coordination problems can be very harmful and are difficult to overcome", which seems quite obvious in retrospect.
I'm pretty sure I already had a good understanding of the trouble coordination problems can cause and probably most other readers did as well. Even thinking that though I still feel like that post has some huge insights that aren't immediately obvious and maybe are difficult to articulate, and I'm wondering if anyone else has any ideas about what made it so special.
I've severely fallen behind in my poetry reading, so I'm looking for recommendations of decent modern (writing in the last twenty years) poets.
What I am not looking for:
(1) the standard "chopped-up prose" where it's
Because I write
The lines
Like this, that
Makes this
A
Poem.
(No, it doesn't. e.e. cummings could get away with it, but you, modern poet person, are *not* e.e. cummings).
(2) Em, this is going to sound critical, but also not whatever it was that girl poet produced for Biden's inauguration (granted, all official poets/poet laureates produce crappy stuff for the Big Official Occasions). If we're talking spoken word poetry, I'm afraid I'm stuck on John Cooper Clarke as my most recent exemplar of same (so, yeah, the 90s).
What I am looking for:
Something modern and good. Is this an impossible request? Hit me with your favourites!
(That reminds me, I've got to go re-read "The Four Quartets").
Has there been any attempt to use technology to decrease the number of teachers needed in K-12 education? I.e. we could imagine a model where a high school or even entire school district only needs to hire one history teacher per grade (or 2-3 if we want honors/AP level history too) to teach all the students in that grade history via Zoom. Then, you could have part-time teaching assistants, who could range from local college students to stay-at-home parents to retirees help “tutor” kids outside of the main lessons being taught via Zoom.
Essentially, instead of needing all teachers to hold undergraduate degrees, as is the case in most school districts (although many districts are dropping this requirement already due to a shortage of qualified applicants) I’m picturing a model where 80-90% of the teaching is done by numerically fewer but “higher quality” (whether that means more credentialed like a MA or PhD or more years of experience) teachers and then the last 10-20% of tutoring being done by less credentialed tutors but in a more individualized, in-person style.
I haven’t worked out logistically exactly how this model would work - but I feel like the advent of video calls / YouTube etc. is made for this sort of more centralized, but possibly higher quality teaching + individualized tutoring system.
I know I’m not the only one who has thought of a model like this, (I think I’ve heard it proposed by a few different people on different podcasts) but I’m wondering if anyone/anywhere is actually trying to put this into place?
As far as I can tell the people heavily invested in a manned mission to Mars don't expend a lot of effort trying to predict when that will happen. They may discuss the issue, argue about when it is likely to happen, but they don't try to predict. In contrast, the people heavily invested in AGI, whether or AGI itself or out of fear about AI going rogue, these people do expend a lot of effort trying to predict when AGI will happen. Why?
That is, we have two (possibly overlapping) groups of people heavily invested in two different kinds of future technology. One group tries to predict when their favored technology will emerge, the other group does not. Why the difference?
I found myself unsubscibed from this substack. Any idea why that might have happened? I thought I'd seen nothing from ACX for a while and came to this manually, that's how I knew I'd been unsubscribed. Normally, I receive an email when a new post appears.
@Meetups: why not publish the available list around Aug 24th as planned, and then invite a second round of applications for organizers ... and publish an extended list say two weeks later?
I might organize sth. if there isn't an organizer in my place (I guess there is), but I simply don't have the time *now* to sign up.
Scott wrote repeatedly about Prospera, the startup city in Honduras.
I visited it for 5 weeks, and I love the vision & the team behind it!
I decided to start a VC fund focused on startup cities, because I think we can build great startups there enabled by better regulations (e.g. peer country regulation, 3D on-chain property rights).
If you're an entrepreneur or innovator, I'd like to show it to you:
I recently heard an interview with Ye Tao, founder of the MEER project which aims to reflect sunlight back into space to help cool the climate.
I was surprised by his claim that at least 1.5°C of additional global warming is already baked in, irrespective of decarbonisation, on account of a net energy influx of 1.5W per square metre. Also that if all coal plants shut now, it would actually temporarily exacerbate the problem as reflective matter produced by burning coal would dissipate from the atmosphere much faster than the cooling effect from the decreased CO2. (Note that he was not shilling for coal – he very much agrees that we need to stop burning fossil fuels, he just thinks reducing the energy influx is even more urgent.)
Does anyone know any papers supporting these claims, and more generally whether whether they are generally accepted or in conflict with the mainstream predictions for warming?
I often hear about how American land was taken from indigenous peoples, but I don't know what the goal of those conversations are. When I say "hear", it's usually either in the context of a land acknowledgement ("The land we stand on belonged/belongs to the XYZ nation.") or something more general. It's hard to tell whether most of these aim to just inform people, or come with the hidden call-to-action that we should return the land to the given nation. It always falls short of being meaningful for the former, but I'm more curious about the second option.
Most of this land has been developed and inhabited by people *other* than those who initially took it. That doesn't make it right to *not* give it back, but taking it away from the current inhabitants doesn't feel right either. I was born and have been inside the USA's borders for all but one week of my life. Both of my parents (and, I believe, all of my grandparents) were natural-born citizens. Any connections I have to a "home country" in Europe are distant enough that I don't even know any relatives there. If I "returned" my land to indigenous peoples, I wouldn't have anywhere at all.
I've never seen anyone explicitly argue for returning land, but it feels like it's being heavily hinted at. Do I have that right? Has anyone written on the topic before? Are there other ideas that are out there that don't run into this issue?
REQUEST: Mathematician who knows about measure theory or Bayes's Theorem for mixed variables.
Setting: Millikan's oil drop experiment.
Millikan announces a surprising discovery that all charges are integer multiples of some smallest charge ! What must his priors have been for this discovery to have been consistent with Bayes's Theorem?
For an experimental measurement, you might expect that the result could be any real number. In this case, your prior would be the Lebesgue measure. But then your prior would assign zero probability to the set of integers: there are infinitely many real numbers for each integer. A Bayesian Millikan would not have believed that charges are integers after any finite amount of experiment data, if his prior was the Lebesgue measure or any other continuous probability density. The interval 1±0.1 contains one integer and infinitely many real numbers, but so does the interval 1±0.00001. Increasing the precision of your experiment doesn't get you to an integer.
What would Millikan's priors have to have been for this announcement to have been rational (in a Bayesian sense)?
Perhaps his priors were a combination between the Lebesgue measure and a sum of Dirac deltas, each of which is centered at an integer.
I'm worried that this doesn't work. Dirac deltas are notoriously ill-defined. Does the continuous Bayes Theorem still apply in a function space that includes Dirac deltas? You can't just multiply and divide by Dirac deltas and expect that everything still works. Maybe there's something that can be done with Bayes's Theorem with mixed random variables instead?
As an extension of this problem, what if we wanted to prefer the rationals instead of just the integers? The prior distribution might involve Dirichlet's or Thomae's function, or it might involve a weighted infinite sum of Dirac deltas, centered at all the rational numbers. This seems like an even worse thing to try to plug into Bayes's Theorem.
I'm looking for organizations that might be willing to donate IFAKs or their components to our organization, so we can use them to train and equip Ukrainian soldiers. Any advice about how to go about this search?
At the risk of shameless self-promotion, I'd like to point out that I've been lackadaisically posting to http://reluctantentrepreneur.substack.com about medicine, technology, and dharma.
Admittedly, there is not much in the way of Dharma yet, but some suff on entrepreneurialism.
More subscribers would, I feel sure, inspire me to write more. All welcome!
1) Why does methylphenidate work immediately contrary to SSRIs despite that both serotonergic and dopaminergic neurons have autoreceptors that inhibit the recapture of their respective neurotransmitters?
2) Why hasn't anybody seriously tried to combine MAOIs with drugs that would prevent these side-effects from happening?
3) Why don't we prescribe exocytosis-promoting molecules such as MDMA when initiating SSRI treatments?
4) Why don't we prescribe autoreceptor antagonists such as pindolol when initiating treatment to make patients respond faster to the treatment and augment SSRIs?
After more than 15 years of living overseas (mostly in African countries) I am thinking about moving to Silicon Valley to teach in public schools. I am getting a lot of family pressure to move there (my Dad is in Palo Alto).
Does anyone think this is a good place to raise my ten year old daughters? Are the cultural opportunities and prospects of being part of a community really great?
I like the area, but I have not spent much time there and I feel like it might be a hyper competitive, rat race environment for my kids. I would get help buying a house, but probably only enough to be in East Palo Alto, or San Jose.
Just to mention that I did a podcast about nuclear war with two ACX stalwarts, Battleship Bean and John Schilling. The basic idea was to discuss nuclear war taking Dr Strangelove as a jumping off point. They talked about things like how powerful are modern nuclear weapons? Would they knock out electrical systems world wide? Would such a war result in nuclear winter?
I wrote a paper about one possible mechanism through which social media fuels intolerance towards other points of view (aka "culture wars"). I sum it up in this blog post: https://www.michelecoscia.com/?p=2179
The even shorter TL;DR is:
- We make reasonable assumptions about the fundamental characteristics of the system: echo chambers, confirmation bias, etc.
- We assume that people want to convince others to behave according to their values and would apply whatever strategy that can lead to that result.
- The main action they have at their disposal is punishing content they disagree with, depending on their level of tolerance.
The result is that both sides learn quickly that there is a (low) level of tolerance that represents an inadequate equilibrium (to use a term from Yud). In this equilibrium, everyone is a jerk because, if they weren't, the other side would nudge content producers to go to their side under the threat of online punishment.
This is true even if they both originally started with high levels of tolerance.
Would love to hear some thoughts. Pardon the self promotion, but I think this is a topic people here are interested in.
Food sanity check request: so, sushi is great, but also pretty expensive. On the other hand, it's ultimately just...raw fish. So is there some red flag risk for simply...buying raw fish, slicing it myself, and eating it? Safety caveats:
*Farm raised, for minimized heavy metal/toxin load.
*Industrially frozen 24hrs minimum, preferably flash frozen on boat, to minimize pathogen load.
This still feels like cheating, somehow, so I suspect I'm overlooking something. Otherwise lotsa people would attempt this...
Anyone else feel like they suffer from a certain kind of intellectual masochism?
By this I mean I really enjoy engaging with technical and non-technical content written by very smart people of past and present, but the more of their content I read the more I realise how inferior I am intellectually.
The way they engage with problems and process information is mostly unattainable for me, but I can't resist the lure of their wonderful ideas, leading to a feedback loop of lowered self-esteem.
On the one hand, I know I shouldn't be upset, and should be (and am) thankful that there are such brilliant people in the world. On the other hand, these brilliant people are often the kinds of people I have the most admiration for, though by that same metric I can never admire myself all that much.
New "Game of Thrones" show premiere is good. I expected it to look fantastic, but the most heartening thing is that the casting is excellent all-around. Bodes well for the show.
Does anyone know what actual findings of the Webb telescope challenge our current understanding of the evolution of our universe? There are lots of youtube videos but i have no idea what to believe. There is post-truth, post-belief and I assume soon post-knowledge for you. Thank you.
Does anyone else share my feeling that humanity has a very rough decade ahead of it?
There is a number of crises brewing at the same time:
- we are just getting past Corona, which has exhausted the resources and patience of many countries, and is still disrupting supply lines (mainly due to China's policies)
- climate change is really making itself felt, with droughts and floods disrupting harvests
- Russia is tearing down the international order and causing further disruption to the trade with food, fertilizer, and energy
- America's politics are torn between a vile narcissistic crybully on one side, and a bunch of loony activists who want to eradicate Western civilization, no, sorry, "whiteness", on the other
- China has essentially become a techno-fascist dictatorship, and is in the first stages of the bursting of the mother of all real estate bubbles
- as a consequence, many developing countries face hunger, state bankruptcy, and a host of other problems all at the same time, with no one willing and able to help
At the very least, I expect a number of really bad civil wars (accompanied by Ruanda-style genocide, famine, disease, and refugee crises), but world war does not seem off the table.
Can anyone convince me I'm wrong, or point out what should be done to prevent the worst?
Im interested in learning more about homelessness in the U.S. —it’s causes and solutions to it. Anyone have any books (or podcasts/articles/etc) on the matter that they would recommend to someone like me?
Does anyone know if a company that came into being, created and sold a product, and then stopped (having maxed out their core competency), and either dissolved or just went into maintenance mode and switched to paying dividends?
Under the ideal model of shareholder capitalism this should happen fairly often - competence is transferable but not *that* transferable. However the closest I can think of to an example is Craigslist (which really is arguably the most successful tech company, as measured by average employee productivity).
I've noticed that most people who try to explain why we sleep, say it's got to do with the brain–to prune useless memories, consolidate useful ones, or <insert technobabble here> and thus grant us creativity.
But the qualia that directs us to try to sleep is tiredness. Hunger, the qualia which directs us to eat, arises when we haven't eaten recently. It seems probable that the qualia that directs us towards sleep is likewise caused by some deficiency which sleep remedies. But what causes tiredness is not thinking, but acting, with our muscles.
Almost no mental activity makes me tired. I get a qualia which I might call fatigue when I play an intense game of chess, but it isn't tiredness. It doesn't make me want to sleep; it just makes me want to stop playing chess. And chess is the only mental activity I can think of which gives me this feeling, maybe because it's the only difficult mental activity I engage in which doesn't interest me much. I've been in long, mentally brutal mathematics competitions, like the Putnam, and they rarely if ever fatigued me. I, and I think other people, fall asleep easily after a day of physical labor; but I must discipline myself to go to bed after a day of hard mental labor, or I'll keep on working until 3AM.
You might object that some people get to sleep by reading or listening to audio recordings. But they don't get to sleep by reading or listening to exciting stories; they read something familiar, or (like my mother) listen to recordings of sermons by monotonous Northern Baptists (also known as "the frozen chosen"). I nearly fall asleep whenever I try reading in the bathtub, but that isn't because it's mentally challenging; it's because the bathtub relaxes my muscles.
So I have a strong prior expectation that the primary function of sleep should have more to do with re-energizing our muscles than our brains. Could it be that all this focus on sleep as refreshing our brains is a result of a bias to value the mind above the body?
Does anyone know the origin(s?) of wedding arches in non-Jewish weddings? I tried googling around for it and only found vague and unsubstantiated claims.
We're basically a small team of remote mercenary IT engineers, and we're looking for a new member or two. Stuff we work on covers a broad range; think everything on the spectrum from a high-performance network block device implementation, to Yet Another iPhone App, to CI/CD infrastructure deployment and management.
The base office is located in Tokyo, Japan, but we're fully remote. As long as there's reasonable overlap with JST, we're mostly agnostic about your physical location.
Members have considerable flexibility in how they design their work schedule as well as moderate flexibility in choosing what to work on. We try to shoot for a 3 day work week, plus as much socializing and leisure hacking together as desired. Even with this schedule, pay falls around the 70th percentile for Japan. It's a pretty sweet setup.
If two or more of the following blurbs fit your autobiography synopsis, then we'd definitely like to hear from you:
- Hacker at heart
- Actively works on "soft skills"
- Mad Japanese skillz
If you're interested, please contact me at this address specification:
Copy pasted from my own question on Economics Stack Exchange, I wonder if anyone here can give me an answer.
"Title: Why do stock returns seem to be uncorrelated with interest rate?
Since expected return of stock is risk-free rate plus risk premium, intuitively they should be correlated. Of course the size of risk premium is not constant, but it's hard to imagine why risk premium would move in a way that almost exactly cancels out the change in interest rate.
Questions:
1. Is the data correct(are they really uncorrelated)? Searching google scholar suggests so, but this is pretty hard to believe so I wouldn't be surprised if I was missing something important.
2. Are there any consensus, or at least a good theory, on why this happens?
3. Real life implications - as a retail investor with pretty strong faith in EMH[Efficient Market Hypothesis], is it rational for me to move my money from stock market to bank account because the interest rate went up?"
The below is my perspective on H1Bs after having been involved in screening, interviewing and hiring software engineering candidates since 2014. Also, I advise other companies, and I have personal & professional connections with managers at other firms in the US tech industry. The dirty secret about H1Bs is that the vast majority of them are….. completely subpar engineers- like unhireable not only by a VC-backed startup, but unhireable by a completely boring ‘normal’ company like a bank or marketing agency. They almost always bomb technical interviews. They’re simply not ‘skilled labor’, as their visa would suggest!
They almost all have at least an impressive-sounding Master’s degree in Computer Science, but if you look more carefully all of the bad ones earned it from a college you’ve never heard of (North Dakota State, West Virginia State, etc.)- which makes me think that they’re basically running diploma mills, charging poor families in 3rd world countries very high tuition for a shot at the US H1B lottery. No, an advanced degree does not make an engineer, but they are much less able to pass interviews than a native with a Bachelor’s from even a subpar school.
All of the H1Bs who are low-quality engineers have the same job- working for a contracting firm that does some kind of outsourced development for gigantic, Fortune 200 companies. That’s it. All of their resumes follow the same format and look exactly the same. A small minority of H1Bs are actually good engineers- these are the ones who go directly to Amazon, FB, Google etc. (Strangely, you can easily tell them because they don’t follow the same resume format as the bad ones! I don’t get this at all).
Contrary to very widespread belief, I don’t find that their wages are particularly low either.
I don’t take any pleasure in reporting this, because I have vanilla center-left politics, am not a nativist, and strongly want more skilled immigration here to the US. I am just here to report that as someone in the industry, under our present system, 90+% of H1Bs are just not that- they are not ‘skilled’. I’m not 100% sure how we ended up here, or what a solution might be. Maybe visas could be parceled out via a bidding system based on what companies offer the candidate, so we use market pricing to determine who is actually worth being awarded a skilled immigrant visa. Just something to keep in mind when you hear libertarians hyperventilating about ‘we should staple a Green Card to the diploma of everyone who gets a STEM degree here’. I don’t personally have a huge issue with that, but it just encourages diploma mills (of course a middle class family in India will spend their life savings to get their child to a US college), and we are just not getting the world’s best & brightest under the current system
I remember a post, 85% certain it was by Scott, about a choice of two pills. One essentially wireheads you, you'll have a long life and you'll experience blissful dreams of whatever you wish the entire time, but in the real world you're in a stupor that you cannot wake up from. The other makes your emotions muted and fleeting, putting you in a permanent state of equanimity, but causes you to be undeterred by any fear, doubt, guilt or anxiety you may have had so you can pursue your goals with maximum effectiveness. Anyone know which one I mean?
I started a "Read The Sequences" series of meetups with my local ACX group. Has anyone else done this and if so do you have any resources to share? In particular, I'm thinking of discussion questions/topics and updated information. If anyone has written about how the cited research stands up post-replication crisis (tangent: are we "post" the replication crisis?), I would love to see it. I know there were a lot of critiques of Thinking Fast and Slow, and this seems to cite similar research in some cases. Is there anything that experts consider outdated that I should be aware of?
It occurred to me that different types of AI research mirror different ethical systems that philosophers have studied. Attempts at building AI back in the 70's were largely rule/logic based, which corresponds to deontology. Current AI, trained to minimize loss or maximize reward, corresponds to consequentialism. Attempts at AI safety seem like an effort to move in the direction of virtue ethics, whereby an AI should learn to act fairly, truthfully, reliably, etc. Is this a fruitful analogy? And if so, has this been noted before? Continuing in this vein, what does 'ideal' virtue ethical 'training' look like, both for humans and speculatively for machines?
I read a great blog post about optimization culture (or a similar term) that talked about, for example, how all cars at similar price points drive the same now, as the availability of tons of customer preference data pushes the manufacturers to be the same. Similarly with consumer electronics etc, maybe even movies if I remember correctly. The idea is that all this data makes everything the same. I'm fairly sure that the post was on the Substack platform. Repeated googling has failed me. Ring any bells for anybody?
I wrote an article [1] on Scott's attitudes towards mysticism. The reddit community has already helped me find a couple errors - would love to get this group's feedback as well.
I've lately been thinking about the anthropic principle. There are multiple variants of the principle, but consider just 3:
(1) Given that humans exist, the laws of physics must allow for the existence of human life. This is an uncontroversial observation, and doesn't really bother me.
(2) Given our current understanding, some of the fundamental constants of physics seem 'fine tuned' to allow for life. We can explain why they take the values they do by first positing the existence of multiple universes with different values of the fundamental parameters. We then, using Bayes' rule, condition on the fact that we are sentient observers who exist and know of our own existence. Conditional on this fact, the fundamental parameters we observe will be a draw from the posterior distribution: random in the range which would allow sentient, self-aware life to emerge. This then 'explains' fine-tuning.
(3) Given that fundamental constants seem fine-tuned, it is likely that the universe was created in order to give rise to life. So one could consider that God specified the laws of physics allowing this to happen. A more secular version of this is Nick Bostrom's simulation hypothesis-- if any sufficiently advanced civilization will create simulations of life, then it is likely that we exist in a simulation (or a simulation of a simulation of a ...) that was created for the express purpose of giving rise to life.
My question is: given that one is inclined to believe anthropic arguments, what, if anything, makes variant (2) more credible than (3)? Many notable physicists seem to believe or at least accept the plausibility of (2) or some variant thereof, while as far as I know, only a minority of physicists believe in (3).
I understand why (2) might be more tempting at a mathematical level, as one does not need to invoke a creator who makes particular choices. Yet, in return, one has to postulate a multitude of unobservable universes, and implicitly accept that one 'could' have counterfactually been born in a different universe (in order to justify the application of Bayes' theorem). These seem in some ways more excessive than postulating the existence of one creator.
Am I missing something? Any thoughts are welcome. Thanks.
How do we explain rationalist thinkers in Middle Ages who claimed inspiration and insight from dreams, visions, and angels (though the insights themselves stand alone without appealing to the supernatural or non rational)? Is there anything to it? How Should we consider the suprarational claims of otherwise very rational thinkers?
I once saw someone who denies the hard problem of consciousness pull out what they thought was a trump card for their side, namely, an experiment where scientists figured out how to read the contents of dreams. They even wrote down what they thought was a step by step breakdown of how the hard problem had been dissolved in that experiment. I don't remember it fully, but there was a part where the scientists asked the subject what they had been dreaming about, so they could establish which neuronal activity correlates to it. It's baffling to me how the person didn't notice the hard problem staring right at them: the scientists did not have access to the subjects experience, otherwise they wouldn't have needed to ask to begin with! The trump card seemed like a very articulate enunciation of the hard problem in the end.
Would you prefer to live in a multi-polar world order organized by distributed consensus protocols or a unipolar surveillance order dictated by a single superintelligence?
Have always been pretty sure of the former until reading Scott’s Meditations on Moloch. Now, not as confident.
I wonder what the demographics look like for how long people have been reading these blog(s). Most of the readers I would assume have been around for quite some time, but I for instance have only been around a bit (been aware for much longer, but jumped on during the substack transition). Would be easy I think to measure using age of subscriptions to the email lists and inferring how many came over from the old. Not sure
I’ve seen in a few places (e.g. Huberman lab podcast) that our brains seem to have two separate “maps”, one for “meaning” and one for other more mundane stuff. Is that understanding correct? Does anyone have any sources on what that distinction is, exactly?
Any interest in the next round of the annual SSC-diaspora Diplomacy game? We're trying to get a seven together. You can sign up here
https://www.datasecretslox.com/index.php/topic,7225.0.html
At least some conservatives are taking Scott up on his "Against Classism" suggestion:
https://www.nationalreview.com/2022/08/bidens-student-debt-bonfire-is-a-classist-message-to-the-uncredentialed-screw-em/
Would there be any benefit to Mexico joining NORAD?
Okay, well...thoughts on the Biden administration's just-announced-today student loan cancellation and forgiveness?
*Risky gamble for midterms, there was some real solid Sound Policy Momentum building for a minute there.
*Exacerbates deficit and also inflation(?), largely regressive, most college-educated already lean D (including didn't-graduates in this category).
*Fairness principles, as with all means-tested benefits; plus larger concerns of those who already paid back loans, and thus partially helped pay for this bailout.
*Time discount: yes, covid + Ukraine has been an outsize destabilizing double punch, but all Future Student Debtors are left out of this jubilee. Proposed ongoing Dept. of Ed reforms will eventually equal the magnitude of front-loaded relief, but it's quite a disproportionate situation for anyone currently existing who isn't *quite* college-age yet.
*Money is fungible, this could have been spent on many more higher-ROI things. Like pandemic preparedness or international relations/foreign policy improvement to help avert getting into future situations like the current one.
As someone suddenly relatively much richer and having a positive net worth for the first time ever, I'll take the money self-interestedly, but...even most-charitably framed as some sort of unique economic-justice response (recompense for Millennials who graduated into the Great Recession), it still seems like an inefficient option within that possibility-space. There are so many worse-off and more-deserving than student debtors...and a majority of Americans never went to college, still. I don't know. This does not feel like The Way. I could be wrong though.
https://www.youtube.com/watch?v=d96dgTAfUro
A very nice little history of polyhedral dice, but it also caused me to think about redundancy and roughage ratings for videos and texts.
Redundancy has an obvious meaning-- roughage is things like small talk between the hosts.
Some videos are very efficient and I appreciate it, some have a little personalization-- Answers with Joe is at a good level of that for me. Some have what I consider to be agonizing levels of repetition.
And then there's repetition between videos. Is there anyone who talks *about* the Fermi Paradox without explaining it yet again?
For people who follow AI more closely, why is speech generation progressing so much slower than image generation? It feels like it should be much easier to put together a synthetic voice that sounds natural than generate a realistic photo, but at least on the public-facing side of things this doesn't seem to be the case. What's the reason behind this, assuming that it's actually the case?
What if Trump removed the classified documents from the White House on January 20, 2021 and took them to Mar-a-Lago, the National Archives informed him that the action was illegal shortly thereafter, and then Trump promptly returned them to the Archives so they were only in his possession for, say, two months? Would he still be in trouble right now?
Hello folks!
I am glad to announce the second of a continuing series of Orange County ACX/LW meetups. Meeting this Saturday and most Saturdays. Note this week we will meet at 3:30 not 2. The first meeting was great, and I hope to see many of you at this one. Based on the first meeting, I chose two popular topics to prompt future conversation and activities.
Saturday, 8/27/22, 3:30 pm
1900 Port Carlow Place, Newport Beach, 92660
The Picnic tables outside the community clubhouse
33.6173166789459, -117.85885652037152
https://goo.gl/maps/WmzxQhBM2vdpJvz39
Plus code 8554J48R+WFJ
Contact me, Michael, at michaelmichalchik+acxlw@gmail.com with questions or requests.
This week it will be at 3:30 (usually 2:00) to avoid a conflict with an online LW meetup
Activities (all activities are optional)
A) Two conversation starter topics this week will be. (readings at the end)
1) Forecasting and predicting the future
2) Psychedelics.
B) We will also have the card game Predictably Irrational. Feel free to bring your own favorite games or distractions.
C) There will be opportunities to go for a walk and talk about an hour after the meeting starts and use some gas barbeques if anyone wants to grill something. There are two easy-access mini-malls nearby with takeout hot food available.
D) Share a surprise! Tell the group about something that happened that was unexpected or changed the way you look at the universe.
E) Make a prediction and give a probability and end condition.
F) Contribute ideas to the future direction of the group. Topics, types of meetings, activities, etc…
Conversation Starter Readings:
Suggested readings for this week are these summaries. These readings are optional, but if you do them, think about what you find interesting, surprising, useful, questionable, vexing, or exciting.
1) Prediction
Superforcasting is a review of experiments done about how well various types of experts do in predicting the future. Generalists tend to do better than specialists in prediction, but why? Groups tend to do better than individuals, but are there ways to improve the performance of groups even further? How can you train yourself to be better at prediction? How can you help others?
https://howdo.com/book-summaries/superforecasting-summary-and-review/
Or The Harvard business review application to business
https://hbr.org/2016/05/superforecasting-how-to-upgrade-your-companys-judgment
Or the ACX Review
https://slatestarcodex.com/2016/02/04/book-review-superforecasting/
And this excerpt from Future Babble is a more critical look at prediction science.
https://www.nytimes.com/2011/03/27/books/review/excerpt-future-babble-by-dan-gardner.html
For psychedelics:
The tale of two receptors is an interesting speculation as to the underlying pharmacology of psychedelics. Hypothesizing that serotonin can both help us cope with the distress of a bad situation and help us look for creative ideas out of the old mental habits that can keep us trapped in a bad situation. Conventional antidepressants and atypical antipsychotics operate on the acceptance system, while psychedelics operate on the later, lateral thinking. Is this perspective useful? Oversimplified? Can it be made into a rigorous scientific idea, or is it just another evolutionary just-so story? What possible uses and hazards does this suggest for psychedelics?
https://slatestarcodex.com/2017/10/10/ssc-journal-club-serotonin-receptors/ or https://journals.sagepub.com/doi/full/10.1177/0269881117725915
If you want a more general introduction to psychedelics, here is a book summary of the recent popular review of psychedelics. “ How to change your mind” by Michael Pollen
https://www.hustleescape.com/book-summary-how-to-change-your-mind-by-michael-pollan/
I've been hunting around for datasets regarding stock-prices and commodities prices, but this is a very new space for me. Does anyone have any suggestions? It seems like there's nothing available that's open source.
On Thursday 8/25 night starting at 8:00pm eastern, 7:00pm central The St. Louis Rationality Group will have an online discussion of the ACX Book Review Contest in GatherTown.
We are inviting all other ACX/Rationality readers as well.
If you haven't read all the book reviews, still come.
If you are nervous about talking to people because you find old age makes you less social, still come. If you will be late due to time zones or other obligations, still come.
If you feel awkward, still come.
It'll be fun.
Here is the link to the event, click it for more information:
https://app.gather.town/events/pXVcEMSts1dcxnOc7rqU
Let's talk ACX Book Reviews. You need not have read every book review to be part of this event, as long as you have read three or so you are good. The purpose here is to meet and learn from each other and share thoughts about this year's book reviews.
Wander from room to room and group to group and talk! No rules about group size or how long you have to stay within a group.
Here are some questions to ask each other:
Do you like or dislike the book review contest?
What makes a good book review?
Are there any books that you have read because of a great book review?
Which book reviews from this year's contest were your favorite?
Did you learn anything useful or insightful from any of the book reviews?
Is book reviewing an art?
What's the best book you've read in the past two years?
I think there's too much focus on trying to design utopian cities, and there should be more focus on trying to design utopian towns first. Get things right at a small scale first, before scaling up to a megacity. If you can take a greenfield site and a budget in the single-digit billions and build a medium-sized town of twenty thousand people, and persuade twenty thousand people that they want to live there because your urban design is so wonderful, then I might believe you can do it with a larger city.
Here's my idea for a town of 20,000. I've tried to cut a middle path between standard urban design and the anti-car fundamentalism of so many "urbanist" type thinkers.
- The whole design looks like a scaled-down Adelaide (check it out) with a downtown core surrounded by a ring of parks surrounded by a ring of suburbs.
- Downtown core is dense and highly walkable. A few major streets and a bunch of narrow lanes. In general there's no parking within downtown itself, but there's space for delivery trucks etc to keep the shops supplied. A narrow river running through the centre of town would be nice.
- Next, the ring of parks, which hide underground car parking beneath. If you want to go downtown, you can park underground and walk. Supermarkets etc can be placed at the outer edges of downtown close to the carpark entrances to make grocery shopping convenient.
- Immediately outside the ring of parks, medium-density apartments and terrace housing, gradually turning into quarter-acre blocks as you go further out. Some major arterial roads head into the residential areas, but there's also tendrils of parkland with bicycle paths in them. A bunch of corner stores and cafes are sprinkled throughout the suburbs. Street layout isn't a grid, nor the cul-de-sac maze of most modern suburban developments, but some kind of haphazard mishmash of the two.
All the comment threads in this post are hidden behind “Continue Thread” links, which are wildly inconvenient to read on a mobile device: is it one uninteresting reply, or a major subthread of interesting opinions from the most prominent bloggers who follow scott? No idea, lets click on 543 slow-loading links one by one to find out!
I miss static html, and dearly wish for substack and similar single-page-application monstrosities to all shutdown, their leaders discredited, their investors bankrupted, and their developers unhireable.
This link has received some valuable feedback from a few particpants in the Model Monday thread for August 1.
Rodes.pub/LineLoop
This is a serious attempt at imagining a linear city. It was inspired by Neom but is different in many respects.
Our linear city:
... BUILDING BUILDING BUILDING ...
... PARK ...
... PARK ...
... PARK ...
... PARK ...
... BURIED HYPERLOOP ...
... PARK
... PARK ...
... PARK ...
... PARK ...
... BUILDING BUILDING BUILDING ...
The buildings resemble the UN building.
100 meters high, 30 stories
10 residental stories
90 meters long
25 meters thick
There are 2000 of them in the completed 100-KM city with a population of 1.5 million.
The city is 200 meters wide. The park is 150 meters wide.
I would love to see others participate in the design.
Peter
I've been re-reading Meditations on Moloch. The first time I read it I remember being very struck by how profound it seemed, but looking over it again, I think its main thesis is just that "coordination problems can be very harmful and are difficult to overcome", which seems quite obvious in retrospect.
I'm pretty sure I already had a good understanding of the trouble coordination problems can cause and probably most other readers did as well. Even thinking that though I still feel like that post has some huge insights that aren't immediately obvious and maybe are difficult to articulate, and I'm wondering if anyone else has any ideas about what made it so special.
I've severely fallen behind in my poetry reading, so I'm looking for recommendations of decent modern (writing in the last twenty years) poets.
What I am not looking for:
(1) the standard "chopped-up prose" where it's
Because I write
The lines
Like this, that
Makes this
A
Poem.
(No, it doesn't. e.e. cummings could get away with it, but you, modern poet person, are *not* e.e. cummings).
(2) Em, this is going to sound critical, but also not whatever it was that girl poet produced for Biden's inauguration (granted, all official poets/poet laureates produce crappy stuff for the Big Official Occasions). If we're talking spoken word poetry, I'm afraid I'm stuck on John Cooper Clarke as my most recent exemplar of same (so, yeah, the 90s).
What I am looking for:
Something modern and good. Is this an impossible request? Hit me with your favourites!
(That reminds me, I've got to go re-read "The Four Quartets").
Has there been any attempt to use technology to decrease the number of teachers needed in K-12 education? I.e. we could imagine a model where a high school or even entire school district only needs to hire one history teacher per grade (or 2-3 if we want honors/AP level history too) to teach all the students in that grade history via Zoom. Then, you could have part-time teaching assistants, who could range from local college students to stay-at-home parents to retirees help “tutor” kids outside of the main lessons being taught via Zoom.
Essentially, instead of needing all teachers to hold undergraduate degrees, as is the case in most school districts (although many districts are dropping this requirement already due to a shortage of qualified applicants) I’m picturing a model where 80-90% of the teaching is done by numerically fewer but “higher quality” (whether that means more credentialed like a MA or PhD or more years of experience) teachers and then the last 10-20% of tutoring being done by less credentialed tutors but in a more individualized, in-person style.
I haven’t worked out logistically exactly how this model would work - but I feel like the advent of video calls / YouTube etc. is made for this sort of more centralized, but possibly higher quality teaching + individualized tutoring system.
I know I’m not the only one who has thought of a model like this, (I think I’ve heard it proposed by a few different people on different podcasts) but I’m wondering if anyone/anywhere is actually trying to put this into place?
As far as I can tell the people heavily invested in a manned mission to Mars don't expend a lot of effort trying to predict when that will happen. They may discuss the issue, argue about when it is likely to happen, but they don't try to predict. In contrast, the people heavily invested in AGI, whether or AGI itself or out of fear about AI going rogue, these people do expend a lot of effort trying to predict when AGI will happen. Why?
That is, we have two (possibly overlapping) groups of people heavily invested in two different kinds of future technology. One group tries to predict when their favored technology will emerge, the other group does not. Why the difference?
I found myself unsubscibed from this substack. Any idea why that might have happened? I thought I'd seen nothing from ACX for a while and came to this manually, that's how I knew I'd been unsubscribed. Normally, I receive an email when a new post appears.
@Meetups: why not publish the available list around Aug 24th as planned, and then invite a second round of applications for organizers ... and publish an extended list say two weeks later?
I might organize sth. if there isn't an organizer in my place (I guess there is), but I simply don't have the time *now* to sign up.
Scott wrote repeatedly about Prospera, the startup city in Honduras.
I visited it for 5 weeks, and I love the vision & the team behind it!
I decided to start a VC fund focused on startup cities, because I think we can build great startups there enabled by better regulations (e.g. peer country regulation, 3D on-chain property rights).
If you're an entrepreneur or innovator, I'd like to show it to you:
Prospera Healthtech Summit, September 23-25, 2022: https://infinitafund.com/healthtech2022
Prospera Edtech Summit, October 28-30, 2022: https://infinitafund.com/edtech2022
Prospera Fintech Summit, November 18-20, 2022: https://infinitafund.com/fintech2022
I recently heard an interview with Ye Tao, founder of the MEER project which aims to reflect sunlight back into space to help cool the climate.
I was surprised by his claim that at least 1.5°C of additional global warming is already baked in, irrespective of decarbonisation, on account of a net energy influx of 1.5W per square metre. Also that if all coal plants shut now, it would actually temporarily exacerbate the problem as reflective matter produced by burning coal would dissipate from the atmosphere much faster than the cooling effect from the decreased CO2. (Note that he was not shilling for coal – he very much agrees that we need to stop burning fossil fuels, he just thinks reducing the energy influx is even more urgent.)
Does anyone know any papers supporting these claims, and more generally whether whether they are generally accepted or in conflict with the mainstream predictions for warming?
Interview here:
https://www.volts.wtf/p/volts-podcast-dr-ye-tao-on-a-grand#details
MEER details here:
https://www.meerreflection.com
I often hear about how American land was taken from indigenous peoples, but I don't know what the goal of those conversations are. When I say "hear", it's usually either in the context of a land acknowledgement ("The land we stand on belonged/belongs to the XYZ nation.") or something more general. It's hard to tell whether most of these aim to just inform people, or come with the hidden call-to-action that we should return the land to the given nation. It always falls short of being meaningful for the former, but I'm more curious about the second option.
Most of this land has been developed and inhabited by people *other* than those who initially took it. That doesn't make it right to *not* give it back, but taking it away from the current inhabitants doesn't feel right either. I was born and have been inside the USA's borders for all but one week of my life. Both of my parents (and, I believe, all of my grandparents) were natural-born citizens. Any connections I have to a "home country" in Europe are distant enough that I don't even know any relatives there. If I "returned" my land to indigenous peoples, I wouldn't have anywhere at all.
I've never seen anyone explicitly argue for returning land, but it feels like it's being heavily hinted at. Do I have that right? Has anyone written on the topic before? Are there other ideas that are out there that don't run into this issue?
REQUEST: Mathematician who knows about measure theory or Bayes's Theorem for mixed variables.
Setting: Millikan's oil drop experiment.
Millikan announces a surprising discovery that all charges are integer multiples of some smallest charge ! What must his priors have been for this discovery to have been consistent with Bayes's Theorem?
For an experimental measurement, you might expect that the result could be any real number. In this case, your prior would be the Lebesgue measure. But then your prior would assign zero probability to the set of integers: there are infinitely many real numbers for each integer. A Bayesian Millikan would not have believed that charges are integers after any finite amount of experiment data, if his prior was the Lebesgue measure or any other continuous probability density. The interval 1±0.1 contains one integer and infinitely many real numbers, but so does the interval 1±0.00001. Increasing the precision of your experiment doesn't get you to an integer.
What would Millikan's priors have to have been for this announcement to have been rational (in a Bayesian sense)?
Perhaps his priors were a combination between the Lebesgue measure and a sum of Dirac deltas, each of which is centered at an integer.
I'm worried that this doesn't work. Dirac deltas are notoriously ill-defined. Does the continuous Bayes Theorem still apply in a function space that includes Dirac deltas? You can't just multiply and divide by Dirac deltas and expect that everything still works. Maybe there's something that can be done with Bayes's Theorem with mixed random variables instead?
As an extension of this problem, what if we wanted to prefer the rationals instead of just the integers? The prior distribution might involve Dirichlet's or Thomae's function, or it might involve a weighted infinite sum of Dirac deltas, centered at all the rational numbers. This seems like an even worse thing to try to plug into Bayes's Theorem.
I'm looking for organizations that might be willing to donate IFAKs or their components to our organization, so we can use them to train and equip Ukrainian soldiers. Any advice about how to go about this search?
Anyone else in Lviv?
At the risk of shameless self-promotion, I'd like to point out that I've been lackadaisically posting to http://reluctantentrepreneur.substack.com about medicine, technology, and dharma.
Admittedly, there is not much in the way of Dharma yet, but some suff on entrepreneurialism.
More subscribers would, I feel sure, inspire me to write more. All welcome!
Many thanks
DJ
I'm looking for feedback on my ADHD/depression pharmacology questions and hypotheses:
https://docs.google.com/document/d/1rJW5TInpW2rvd6CFkKCdlPxv-IMefBJ4PdQuff1uo6U/edit?usp=sharing
Thanks!
CC Scott Alexander
—————————
Outline:
1) Why does methylphenidate work immediately contrary to SSRIs despite that both serotonergic and dopaminergic neurons have autoreceptors that inhibit the recapture of their respective neurotransmitters?
2) Why hasn't anybody seriously tried to combine MAOIs with drugs that would prevent these side-effects from happening?
3) Why don't we prescribe exocytosis-promoting molecules such as MDMA when initiating SSRI treatments?
4) Why don't we prescribe autoreceptor antagonists such as pindolol when initiating treatment to make patients respond faster to the treatment and augment SSRIs?
After more than 15 years of living overseas (mostly in African countries) I am thinking about moving to Silicon Valley to teach in public schools. I am getting a lot of family pressure to move there (my Dad is in Palo Alto).
Does anyone think this is a good place to raise my ten year old daughters? Are the cultural opportunities and prospects of being part of a community really great?
I like the area, but I have not spent much time there and I feel like it might be a hyper competitive, rat race environment for my kids. I would get help buying a house, but probably only enough to be in East Palo Alto, or San Jose.
What should I do in Milan today?
Can anyone recommend smaller EA/rationality-aligned blogs? I think I've located all the bigger ones. I'm really curious about the underdogs, though!
Just to mention that I did a podcast about nuclear war with two ACX stalwarts, Battleship Bean and John Schilling. The basic idea was to discuss nuclear war taking Dr Strangelove as a jumping off point. They talked about things like how powerful are modern nuclear weapons? Would they knock out electrical systems world wide? Would such a war result in nuclear winter?
https://pod.link/1436447503/episode/f5853da56114f46a20f255089b965e3a
I wrote a paper about one possible mechanism through which social media fuels intolerance towards other points of view (aka "culture wars"). I sum it up in this blog post: https://www.michelecoscia.com/?p=2179
The even shorter TL;DR is:
- We make reasonable assumptions about the fundamental characteristics of the system: echo chambers, confirmation bias, etc.
- We assume that people want to convince others to behave according to their values and would apply whatever strategy that can lead to that result.
- The main action they have at their disposal is punishing content they disagree with, depending on their level of tolerance.
The result is that both sides learn quickly that there is a (low) level of tolerance that represents an inadequate equilibrium (to use a term from Yud). In this equilibrium, everyone is a jerk because, if they weren't, the other side would nudge content producers to go to their side under the threat of online punishment.
This is true even if they both originally started with high levels of tolerance.
Would love to hear some thoughts. Pardon the self promotion, but I think this is a topic people here are interested in.
Food sanity check request: so, sushi is great, but also pretty expensive. On the other hand, it's ultimately just...raw fish. So is there some red flag risk for simply...buying raw fish, slicing it myself, and eating it? Safety caveats:
*Farm raised, for minimized heavy metal/toxin load.
*Industrially frozen 24hrs minimum, preferably flash frozen on boat, to minimize pathogen load.
This still feels like cheating, somehow, so I suspect I'm overlooking something. Otherwise lotsa people would attempt this...
Anyone else feel like they suffer from a certain kind of intellectual masochism?
By this I mean I really enjoy engaging with technical and non-technical content written by very smart people of past and present, but the more of their content I read the more I realise how inferior I am intellectually.
The way they engage with problems and process information is mostly unattainable for me, but I can't resist the lure of their wonderful ideas, leading to a feedback loop of lowered self-esteem.
On the one hand, I know I shouldn't be upset, and should be (and am) thankful that there are such brilliant people in the world. On the other hand, these brilliant people are often the kinds of people I have the most admiration for, though by that same metric I can never admire myself all that much.
Anyone know what I mean?
New "Game of Thrones" show premiere is good. I expected it to look fantastic, but the most heartening thing is that the casting is excellent all-around. Bodes well for the show.
Does anyone know what actual findings of the Webb telescope challenge our current understanding of the evolution of our universe? There are lots of youtube videos but i have no idea what to believe. There is post-truth, post-belief and I assume soon post-knowledge for you. Thank you.
Does anyone else share my feeling that humanity has a very rough decade ahead of it?
There is a number of crises brewing at the same time:
- we are just getting past Corona, which has exhausted the resources and patience of many countries, and is still disrupting supply lines (mainly due to China's policies)
- climate change is really making itself felt, with droughts and floods disrupting harvests
- Russia is tearing down the international order and causing further disruption to the trade with food, fertilizer, and energy
- America's politics are torn between a vile narcissistic crybully on one side, and a bunch of loony activists who want to eradicate Western civilization, no, sorry, "whiteness", on the other
- China has essentially become a techno-fascist dictatorship, and is in the first stages of the bursting of the mother of all real estate bubbles
- as a consequence, many developing countries face hunger, state bankruptcy, and a host of other problems all at the same time, with no one willing and able to help
At the very least, I expect a number of really bad civil wars (accompanied by Ruanda-style genocide, famine, disease, and refugee crises), but world war does not seem off the table.
Can anyone convince me I'm wrong, or point out what should be done to prevent the worst?
any ACXers in Siem Reap, Cambodia?
Im interested in learning more about homelessness in the U.S. —it’s causes and solutions to it. Anyone have any books (or podcasts/articles/etc) on the matter that they would recommend to someone like me?
Does anyone know if a company that came into being, created and sold a product, and then stopped (having maxed out their core competency), and either dissolved or just went into maintenance mode and switched to paying dividends?
Under the ideal model of shareholder capitalism this should happen fairly often - competence is transferable but not *that* transferable. However the closest I can think of to an example is Craigslist (which really is arguably the most successful tech company, as measured by average employee productivity).
I've noticed that most people who try to explain why we sleep, say it's got to do with the brain–to prune useless memories, consolidate useful ones, or <insert technobabble here> and thus grant us creativity.
But the qualia that directs us to try to sleep is tiredness. Hunger, the qualia which directs us to eat, arises when we haven't eaten recently. It seems probable that the qualia that directs us towards sleep is likewise caused by some deficiency which sleep remedies. But what causes tiredness is not thinking, but acting, with our muscles.
Almost no mental activity makes me tired. I get a qualia which I might call fatigue when I play an intense game of chess, but it isn't tiredness. It doesn't make me want to sleep; it just makes me want to stop playing chess. And chess is the only mental activity I can think of which gives me this feeling, maybe because it's the only difficult mental activity I engage in which doesn't interest me much. I've been in long, mentally brutal mathematics competitions, like the Putnam, and they rarely if ever fatigued me. I, and I think other people, fall asleep easily after a day of physical labor; but I must discipline myself to go to bed after a day of hard mental labor, or I'll keep on working until 3AM.
You might object that some people get to sleep by reading or listening to audio recordings. But they don't get to sleep by reading or listening to exciting stories; they read something familiar, or (like my mother) listen to recordings of sermons by monotonous Northern Baptists (also known as "the frozen chosen"). I nearly fall asleep whenever I try reading in the bathtub, but that isn't because it's mentally challenging; it's because the bathtub relaxes my muscles.
So I have a strong prior expectation that the primary function of sleep should have more to do with re-energizing our muscles than our brains. Could it be that all this focus on sleep as refreshing our brains is a result of a bias to value the mind above the body?
Does anyone know the origin(s?) of wedding arches in non-Jewish weddings? I tried googling around for it and only found vague and unsubstantiated claims.
Not sure if it's kosher to post hiring posts here, so please delete if I'm breaking The Code.
Homepage: https://rainlab.co.jp/
We're basically a small team of remote mercenary IT engineers, and we're looking for a new member or two. Stuff we work on covers a broad range; think everything on the spectrum from a high-performance network block device implementation, to Yet Another iPhone App, to CI/CD infrastructure deployment and management.
The base office is located in Tokyo, Japan, but we're fully remote. As long as there's reasonable overlap with JST, we're mostly agnostic about your physical location.
Members have considerable flexibility in how they design their work schedule as well as moderate flexibility in choosing what to work on. We try to shoot for a 3 day work week, plus as much socializing and leisure hacking together as desired. Even with this schedule, pay falls around the 70th percentile for Japan. It's a pretty sweet setup.
If two or more of the following blurbs fit your autobiography synopsis, then we'd definitely like to hear from you:
- Hacker at heart
- Actively works on "soft skills"
- Mad Japanese skillz
If you're interested, please contact me at this address specification:
local-part: wilson
domain: rainlab.co.jp
Copy pasted from my own question on Economics Stack Exchange, I wonder if anyone here can give me an answer.
"Title: Why do stock returns seem to be uncorrelated with interest rate?
Since expected return of stock is risk-free rate plus risk premium, intuitively they should be correlated. Of course the size of risk premium is not constant, but it's hard to imagine why risk premium would move in a way that almost exactly cancels out the change in interest rate.
Questions:
1. Is the data correct(are they really uncorrelated)? Searching google scholar suggests so, but this is pretty hard to believe so I wouldn't be surprised if I was missing something important.
2. Are there any consensus, or at least a good theory, on why this happens?
3. Real life implications - as a retail investor with pretty strong faith in EMH[Efficient Market Hypothesis], is it rational for me to move my money from stock market to bank account because the interest rate went up?"
How do people who believe in a mind-body division explain the effect of alcohol?
The below is my perspective on H1Bs after having been involved in screening, interviewing and hiring software engineering candidates since 2014. Also, I advise other companies, and I have personal & professional connections with managers at other firms in the US tech industry. The dirty secret about H1Bs is that the vast majority of them are….. completely subpar engineers- like unhireable not only by a VC-backed startup, but unhireable by a completely boring ‘normal’ company like a bank or marketing agency. They almost always bomb technical interviews. They’re simply not ‘skilled labor’, as their visa would suggest!
They almost all have at least an impressive-sounding Master’s degree in Computer Science, but if you look more carefully all of the bad ones earned it from a college you’ve never heard of (North Dakota State, West Virginia State, etc.)- which makes me think that they’re basically running diploma mills, charging poor families in 3rd world countries very high tuition for a shot at the US H1B lottery. No, an advanced degree does not make an engineer, but they are much less able to pass interviews than a native with a Bachelor’s from even a subpar school.
All of the H1Bs who are low-quality engineers have the same job- working for a contracting firm that does some kind of outsourced development for gigantic, Fortune 200 companies. That’s it. All of their resumes follow the same format and look exactly the same. A small minority of H1Bs are actually good engineers- these are the ones who go directly to Amazon, FB, Google etc. (Strangely, you can easily tell them because they don’t follow the same resume format as the bad ones! I don’t get this at all).
Contrary to very widespread belief, I don’t find that their wages are particularly low either.
I don’t take any pleasure in reporting this, because I have vanilla center-left politics, am not a nativist, and strongly want more skilled immigration here to the US. I am just here to report that as someone in the industry, under our present system, 90+% of H1Bs are just not that- they are not ‘skilled’. I’m not 100% sure how we ended up here, or what a solution might be. Maybe visas could be parceled out via a bidding system based on what companies offer the candidate, so we use market pricing to determine who is actually worth being awarded a skilled immigrant visa. Just something to keep in mind when you hear libertarians hyperventilating about ‘we should staple a Green Card to the diploma of everyone who gets a STEM degree here’. I don’t personally have a huge issue with that, but it just encourages diploma mills (of course a middle class family in India will spend their life savings to get their child to a US college), and we are just not getting the world’s best & brightest under the current system
I’ve been wondering if AI could offload the GI part to people and become an AGI in a non trivial way.
i have a small 'based' but rat friendly dinner club here in austin. drop me a line if you are interested
I remember a post, 85% certain it was by Scott, about a choice of two pills. One essentially wireheads you, you'll have a long life and you'll experience blissful dreams of whatever you wish the entire time, but in the real world you're in a stupor that you cannot wake up from. The other makes your emotions muted and fleeting, putting you in a permanent state of equanimity, but causes you to be undeterred by any fear, doubt, guilt or anxiety you may have had so you can pursue your goals with maximum effectiveness. Anyone know which one I mean?
I started a "Read The Sequences" series of meetups with my local ACX group. Has anyone else done this and if so do you have any resources to share? In particular, I'm thinking of discussion questions/topics and updated information. If anyone has written about how the cited research stands up post-replication crisis (tangent: are we "post" the replication crisis?), I would love to see it. I know there were a lot of critiques of Thinking Fast and Slow, and this seems to cite similar research in some cases. Is there anything that experts consider outdated that I should be aware of?
It occurred to me that different types of AI research mirror different ethical systems that philosophers have studied. Attempts at building AI back in the 70's were largely rule/logic based, which corresponds to deontology. Current AI, trained to minimize loss or maximize reward, corresponds to consequentialism. Attempts at AI safety seem like an effort to move in the direction of virtue ethics, whereby an AI should learn to act fairly, truthfully, reliably, etc. Is this a fruitful analogy? And if so, has this been noted before? Continuing in this vein, what does 'ideal' virtue ethical 'training' look like, both for humans and speculatively for machines?
Does anyone give Gladwell’s 10000 hour "rule" the time of day any more?
https://howaboutthis.substack.com/p/curious-realizer-10000-hours-to-what
I read a great blog post about optimization culture (or a similar term) that talked about, for example, how all cars at similar price points drive the same now, as the availability of tons of customer preference data pushes the manufacturers to be the same. Similarly with consumer electronics etc, maybe even movies if I remember correctly. The idea is that all this data makes everything the same. I'm fairly sure that the post was on the Substack platform. Repeated googling has failed me. Ring any bells for anybody?
I wrote an article [1] on Scott's attitudes towards mysticism. The reddit community has already helped me find a couple errors - would love to get this group's feedback as well.
[1] https://superbowl.substack.com/p/the-mysticism-of-scott-alexander
I've lately been thinking about the anthropic principle. There are multiple variants of the principle, but consider just 3:
(1) Given that humans exist, the laws of physics must allow for the existence of human life. This is an uncontroversial observation, and doesn't really bother me.
(2) Given our current understanding, some of the fundamental constants of physics seem 'fine tuned' to allow for life. We can explain why they take the values they do by first positing the existence of multiple universes with different values of the fundamental parameters. We then, using Bayes' rule, condition on the fact that we are sentient observers who exist and know of our own existence. Conditional on this fact, the fundamental parameters we observe will be a draw from the posterior distribution: random in the range which would allow sentient, self-aware life to emerge. This then 'explains' fine-tuning.
(3) Given that fundamental constants seem fine-tuned, it is likely that the universe was created in order to give rise to life. So one could consider that God specified the laws of physics allowing this to happen. A more secular version of this is Nick Bostrom's simulation hypothesis-- if any sufficiently advanced civilization will create simulations of life, then it is likely that we exist in a simulation (or a simulation of a simulation of a ...) that was created for the express purpose of giving rise to life.
My question is: given that one is inclined to believe anthropic arguments, what, if anything, makes variant (2) more credible than (3)? Many notable physicists seem to believe or at least accept the plausibility of (2) or some variant thereof, while as far as I know, only a minority of physicists believe in (3).
I understand why (2) might be more tempting at a mathematical level, as one does not need to invoke a creator who makes particular choices. Yet, in return, one has to postulate a multitude of unobservable universes, and implicitly accept that one 'could' have counterfactually been born in a different universe (in order to justify the application of Bayes' theorem). These seem in some ways more excessive than postulating the existence of one creator.
Am I missing something? Any thoughts are welcome. Thanks.
How do we explain rationalist thinkers in Middle Ages who claimed inspiration and insight from dreams, visions, and angels (though the insights themselves stand alone without appealing to the supernatural or non rational)? Is there anything to it? How Should we consider the suprarational claims of otherwise very rational thinkers?
I once saw someone who denies the hard problem of consciousness pull out what they thought was a trump card for their side, namely, an experiment where scientists figured out how to read the contents of dreams. They even wrote down what they thought was a step by step breakdown of how the hard problem had been dissolved in that experiment. I don't remember it fully, but there was a part where the scientists asked the subject what they had been dreaming about, so they could establish which neuronal activity correlates to it. It's baffling to me how the person didn't notice the hard problem staring right at them: the scientists did not have access to the subjects experience, otherwise they wouldn't have needed to ask to begin with! The trump card seemed like a very articulate enunciation of the hard problem in the end.
This is the research in question: https://www.bbc.com/news/science-environment-22031074
To those who deny there is a hard problem, can you see it now?
Would you prefer to live in a multi-polar world order organized by distributed consensus protocols or a unipolar surveillance order dictated by a single superintelligence?
Have always been pretty sure of the former until reading Scott’s Meditations on Moloch. Now, not as confident.
I wonder what the demographics look like for how long people have been reading these blog(s). Most of the readers I would assume have been around for quite some time, but I for instance have only been around a bit (been aware for much longer, but jumped on during the substack transition). Would be easy I think to measure using age of subscriptions to the email lists and inferring how many came over from the old. Not sure
I’ve seen in a few places (e.g. Huberman lab podcast) that our brains seem to have two separate “maps”, one for “meaning” and one for other more mundane stuff. Is that understanding correct? Does anyone have any sources on what that distinction is, exactly?
Why do new movies suck? Is the medium still relevant?