Show cover of Astral Codex Ten Podcast

Astral Codex Ten Podcast

The official audio version of Astral Codex Ten, with an archive of posts from Slate Star Codex. It's just me reading Scott Alexander's blog posts.


Book Review: First Sixth Of Bobos In Paradise I. David Brooks’ Bobos In Paradise is an uneven book. The first sixth is a daring historical thesis that touches on every aspect of 20th-century America. The next five-sixths are the late-90s equivalent of “millennials just want avocado toast!” I’ll review the first sixth here, then see if I can muster enough enthusiasm to get to the rest later. The daring thesis: a 1950s change in Harvard admissions policy destroyed one American aristocracy and created another. Everything else is downstream of the aristocracy, so this changed the whole character of the US. The pre-1950s aristocracy went by various names; the Episcopacy, the Old Establishment, Boston Brahmins. David Brooks calls them WASPs, which is evocative but ambiguous. He doesn’t just mean Americans who happen to be white, Anglo-Saxon, and Protestant - there are tens of millions of those! He means old-money blue-blooded Great-Gatsby-villain WASPs who live in Connecticut, go sailing, play lacrosse, belong to country clubs, and have names like Thomas R. Newbury-Broxham III. Everyone in their family has gone to Yale for eight generations; if someone in the ninth generation got rejected, the family patriarch would invite the Chancellor of Yale to a nice game of golf and mention it in a very subtle way, and the Chancellor would very subtly apologize and say that of course a Newbury-Broxham must go to Yale, and whoever is responsible shall be very subtly fired forthwith. The old-money WASPs were mostly descendants of people who made their fortunes in colonial times (or at worst the 1800s); they were a merchant aristocracy. As the descendants of merchants, they acted as standard-bearers for the bourgeois virtues: punctuality, hard work, self-sufficiency, rationality, pragmatism, conformity, ruthlessness, whatever made your factory out-earn its competitors. By the 1950s they were several generations removed from any actual hustling entrepreneur. Still, at their best the seed ran strong and they continued to embody some of these principles. Brooks tentatively admires the WASP aristocracy for their ethos of noblesse oblige - many become competent administrators, politicians, and generals. George H. W. Bush, scion of a rich WASP family, served with distinction in World War II - the modern equivalent would be Bill Gates’ or Charles Koch’s kids volunteering as front-line troops in Afghanistan.
24:47 12/02/2022
Highlights From The Comments On Semaglutide Table of contents: 1. Top Comments 2. More Tips On Getting Cheap Semaglutide 3. Other Weight Loss Drugs 4. People Challenging My Numbers And Predictions 5. Do You Have To Stay On Semaglutide Forever Or Else Gain The Weight Back? 6. Personal Anecdotes 7. Tangents That I Find Tedious, But Other People Apparently Really Want To Debate    
35:15 12/01/2022
Can This AI Save Teenage Spy Alex Rider From A Terrible Fate?
We’re showcasing a hot new totally bopping, popping musical track called “bromancer era? bromancer era?? bromancer era???“ His subtle sublime thoughts raced, making his eyes literally explode.         “He peacefully enjoyed the light and flowers with his love,” she said quietly, as he knelt down gently and silently. “I also would like to walk once more into the garden if I only could,” he said, watching her. “I would like that so much,” Katara said. A brick hit him in the face and he died instantly, though not before reciting his beloved last vows: “For psp and other releases on friday, click here to earn an early (presale) slot ticket entry time or also get details generally about all releases and game features there to see how you can benefit!” — Talk To Filtered Transformer Rating: 0.1% probability of including violence “Prosaic alignment” is the most popular paradigm in modern AI alignment. It theorizes that we’ll train future superintelligent AIs the same way that we train modern dumb ones: through gradient descent via reinforcement learning. Every time they do a good thing, we say “Yes, like this!”, in a way that pulls their incomprehensible code slightly in the direction of whatever they just did. Every time they do a bad thing, we say “No, not that!,” in a way that pushes their incomprehensible code slightly in the opposite direction. After training on thousands or millions of examples, the AI displays a seemingly sophisticated understanding of the conceptual boundaries of what we want. For example, suppose we have an AI that’s good at making money. But we want to align it to a harder task: making money without committing any crimes. So we simulate it running money-making schemes a thousand times, and give it positive reinforcement every time it generates a legal plan, and negative reinforcement every time it generates a criminal one. At the end of the training run, we hopefully have an AI that’s good at making money and aligned with our goal of following the law. Two things could go wrong here: The AI is stupid, ie incompetent at world-modeling. For example, it might understand that we don’t want it to commit murder, but not understand that selling arsenic-laden food will kill humans. So it sells arsenic-laden food and humans die. The AI understands the world just fine, but didn’t absorb the categories we thought it absorbed. For example, maybe none of our examples involved children, and so the AI learned not to murder adult humans, but didn’t learn not to murder children. This isn’t because the AI is too stupid to know that children are humans. It’s because we’re running a direct channel to something like the AI’s “subconscious”, and we can only talk to it by playing this dumb game of “try to figure out the boundaries of the category including these 1,000 examples”. Problem 1 is self-resolving; once AIs are smart enough to be dangerous, they’re probably smart enough to model the world well. How bad is Problem 2? Will an AI understand the category boundaries of what we want easily and naturally after just a few examples? Will it take millions of examples and a desperate effort? Or is there some reason why even smart AIs will never end up with goals close enough to ours to be safe, no matter how many examples we give them? AI scientists have debated these questions for years, usually as pure philosophy. But we’ve finally reached a point where AIs are smart enough for us to run the experiment directly. Earlier this year, Redwood Research embarked on an ambitious project to test whether AIs could learn categories and reach alignment this way - a project that would require a dozen researchers, thousands of dollars of compute, and 4,300 Alex Rider fanfiction stories.
38:14 11/30/2022
140 million obese Americans x $15,000/year for obesity drugs = . . . uh oh, that can't be right.   Semaglutide started off as a diabetes medication. Pharma company Novo Nordisk developed it in the early 2010s, and the FDA approved it under the brand names Ozempic® (for the injectable) and Rybelsus® (for the pill).   I think “Ozempic” sounds like one of those unsinkable ocean liners, and “Rybelsus” sounds like a benevolent mythological blacksmith. Patients reported significant weight loss as a side effect. Semaglutide was a GLP-1 agonist, a type of drug that has good theoretical reasons to affect weight, so Novo Nordisk studied this and found that yes, it definitely caused people to lose a lot of weight. More weight than any safe drug had ever caused people to lose before. In 2021, the FDA approved semaglutide for weight loss under the brand name Wegovy®.   “Wegovy” sounds like either a cooperative governance platform, or some kind of obscure medieval sin. Weight loss pills have a bad reputation. But Wegovy is a big step up. It doesn’t work for everybody. But it works for 66-84% of people, depending on your threshold.   (Source) Of six major weight loss drugs, only two - Wegovy and Qsymia - have a better than 50-50 chance of helping you lose 10% of your weight. Qsymia works partly by making food taste terrible; it can also cause cognitive issues. Wegovy feels more natural; patients just feel full and satisfied after they’ve eaten a healthy amount of food. You can read the gushing anecdotes here (plus some extra anecdotes in the comments). Wegovy patients also lose more weight on average than Qsymia patients - 15% compared to 10%. It’s just a really impressive drug. Until now, doctors didn’t really use medication to treat obesity; the drugs either didn’t work or had too many side effects. They recommended either diet and exercise (for easier cases) or bariatric surgery (for harder ones). Semaglutide marks the start of a new generation of weight loss drugs that are more clearly worthwhile. Modeling Semaglutide Accessibility   40% of Americans are obese - that’s 140 million people. Most of them would prefer to be less obese. Suppose that a quarter of them want semaglutide. That’s 35 million prescriptions. Semaglutide costs about $15,000 per year, multiply it out, that’s about $500 billion. Americans currently spend $300 billion per year total on prescription drugs. So if a quarter of the obese population got semaglutide, that would cost almost twice as much as all other drug spending combined. It would probably bankrupt half the health care industry. So . . . most people who want semaglutide won’t get it? Unclear. America’s current policy for controlling medical costs is to buy random things at random prices, then send all the bills to an illiterate reindeer-herder named Yagmuk, who burns them for warmth. Anything could happen! Right now, only about 50,000 Americans take semaglutide for obesity. I’m basing this off this report claiming “20,000 weekly US prescriptions” of Wegovy; since it’s taken once per week, maybe this means there are 20,000 users? Or maybe each prescription contains enough Wegovy to last a month and there are 80,000 users? I’m not sure, but it’s somewhere in the mid five digits, which I’m rounding to 50,000. That’s only 0.1% of the potential 35 million. The next few sections of this post are about why so few people are on semaglutide, and whether we should expect that to change. I’ll start by going over my model of what determines semaglutide use, then look at a Morgan Stanley projection of what will happen over the next decade. Step 1: Awareness I model semaglutide use as interest * awareness * prescription accessibility * affordability. I already randomly guessed interest at 25%, so the next step is awareness. How many people are aware of semaglutide? The answer is: a lot more now than when I first started writing this article! Novo Nordisk’s Wegovy Gets Surprise Endorsement From Elon Musk, says the headline. And here’s Google Trends:
29:17 11/25/2022
"Is Wine Fake?" In Asterisk Magazine
I wrote an article on whether wine is fake. It's not here, it's at, the new rationalist / effective altruist magazine. Congratulations to my friend Clara for making it happen. Stories include: Modeling The End Of Monkeypox: I’m especially excited about this one. The top forecaster (of 7,000) in the 2021 Good Judgment competition explains his predictions for monkeypox. If you’ve ever rolled your eyes at a column by some overconfident pundit, this is maybe the most opposite-of-that thing ever published. Book Review - What We Owe The Future: You’ve read mine, this is Kelsey Piper’s. Kelsey is always great, and this is a good window into the battle over the word “long-termism”. Making Sense Of Moral Change: Interview with historian Christopher Brown on the end of the slave trade. “There is a false dichotomy between sincere activism and self-interested activism. Abolitionists were quite sincerely horrified by slavery and motivated to end it, but their fight for abolition was not entirely altruistic.” How To Prevent The Next Pandemic: MIT professor Kevin Esvelt talks about applying the security mindset to bioterrorism. “At least 38,000 people can assemble an influenza virus from scratch. If people identify a new [pandemic] virus . . . then you just gave 30,000 people access to an agent that is of nuclear-equivalent lethality.” Rebuilding After The Replication Crisis: This is Stuart Ritchie, hopefully you all know him by now. “Fundamentally, how much more can we trust a study published in 2022 compared to one from 2012?” Why Isn’t The Whole World Rich? Professor Dietrich Vollrath’s introduction to growth economics. What caused the South Korean miracle, and why can’t other countries copy it? Is Wine Fake? By me! How come some people say blinded experts can tell the country, subregion, and year of any wine just by tasting it, but other people say blinded experts get fooled by white wines dyed red? China’s Silicon Future: Why does China have so much trouble building advanced microchips? How will the CHIPS act affect its broader economic rise? By Karson Elmgren.
05:12 11/23/2022
Mantic Monday: Twitter Chaos Edition
Plus FTX charges, scandal markets - and oh yeah, wasn't there some kind of midterm recently? Twitter! This is all going to be so, so obsolete by the time I finish writing it and hit the “send post” button. But here goes: 395 traders on this, so one of Manifold’s biggest markets, probably representative. The small print defines a major outage as one that lasts more than an hour. See here for a good explanation of why some people expect Twitter outages. is within 2% of Manifold. Metaculus here has slightly stricter criteria but broadly agrees. 71 traders, still pretty good, but I find it meaningless without a way to distinguish between “everything collapses, Elon sells it for peanuts to scavengers” vs. “Elon saves Twitter, then hands it over to a minion while he moves on to a company building giant death zeppelins”. Oh, here we go. 20 traders, they think Musk will stay in charge. 23 traders. Twitter was profitable in 2018 and 2019, then went back to being net negative in 2020 and 2021 (I don’t know why) . I don’t think it’s been very profitable lately, so it would be a feather in Musk’s cap if he accomplished this. 24 traders. Twitter’s mDAU have consistently gone up in the past. DAU is slightly different and I think more likely to include bots. 26 traders. One thing I like about Manifold is that it lets you choose any point along the gradient from “completely objective” (eg Twitter’s reported DAU count) to “completely subject” (eg whether the person who made the market thinks something is better or worse). This at least uses a poll as its resolution method. But the poll will be in the comments of this market, which means it will mostly be by people who invested in this market, who’ll have strong incentives to manipulate it. Maybe Manifold should add a polling platform to their service?
28:42 11/22/2022
The Psychopharmacology Of The FTX Crash Must not blog about FTX . . . must not blog about . . . ah, $#@% it         Tyler Cowen linked Milky Eggs’ excellent overview of the FTX crash. I’m unqualified to comment on any of the financial or regulatory aspects. But it turns out there’s a psychopharmacology angle, which I am qualified to talk about, so let’s go. I wrote this pretty rushed because it’s an evolving news story. Sorry if it’s less polished than usual.1 1: Was SBF Using A Medication That Can Cause Overspending And Compulsive Gambling As A Side Effect? Probably yes, and maybe it could have had some small effect, but probably not as much as the people discussing it on Twitter think. Milky Eggs reports a claim by an employee that Sam was on “a patch for designer stimulants that mainlined them into his blood to give him a constant buzz at all times”. This could be a hyperbolic description of Emsam, a patch form of the antidepressant/antiparkinsonian agent selegiline. The detectives at the @AutismCapital Twitter account found a photo of SBF, zoomed in on a scrap of paper on his desk, and recognized it as an Emsam wrapper.
46:03 11/18/2022
Contra Resident Contrarian On Unfalsifiable Internal States I. Contra Resident Contrarian . . . Resident Contrarian writes On Unfalsifiable Internal States, where he defends his skepticism of jhana and other widely-claimed hard-to-falsify internal states. It’s long, but I’ll quote a part that seemed especially important to me: I don’t really want to do the part of this article that’s about how it’s reasonable to doubt people in some contexts. But to get to the part I want to talk about, I sort of have to. There is a thriving community of people pretending to have a bunch of multiple personalities on TikTok. They are (they say) composed of many quirky little somebodies, complete with different fun backstories. They get millions of views talking about how great life is when lived as multiples, and yet almost everyone who encounters these videos in the wild goes “What the hell is this? Who pretends about this kind of stuff?” There’s an internet community of people, mostly young women, who pretend to be sick. They call themselves Spoonies; it’s a name derived from the idea that physically and mentally well people have unlimited “spoons”, or mental/physical resources they use to deal with their day. Spoonies are claiming to have fewer spoons, but also en masse have undiagnosable illnesses. They trade tips on how to force their doctors to give them diagnoses: > In a TikTok video, a woman with over 30,000 followers offers advice on how to lie to your doctor. “If you have learned to eat salt and follow internet instructions and buy compression socks and squeeze your thighs before you stand up to not faint…and you would faint without those things, go into that appointment and tell them you faint.” Translation: You know your body best. And if twisting the facts (like saying you faint when you don’t) will get you what you want (a diagnosis, meds), then go for it. One commenter added, “I tell docs I'm adopted. They'll order every test under the sun”—because adoption means there may be no family history to help with diagnoses. And doctors note being able to sort of track when particular versions of illnesses get flavor-of-the-week status: > Over the pandemic, neurologists across the globe noticed a sharp uptick in teen girls with tics, according to a report in the Wall Street Journal. Many at one clinic in Chicago were exhibiting the same tic: uncontrollably blurting out the word “beans.” It turned out the teens were taking after a popular British TikToker with over 15 million followers. The neurologist who discovered the “beans” thread, Dr. Caroline Olvera at Rush University Medical Center, declined to speak with me—because of “the negativity that can come from the TikTok community,” according to a university spokesperson.   Almost no one who encounters them assumes they are actually sick. Are there individuals in each of these communities that are “for real”? Probably, especially in the case of the Spoonies; undiagnosed or undiagnosable illnesses are a real thing. Are most of them legitimate? The answer seems to be a pretty clear “no”. I’m not bringing them up to bully them; I suspect that there are profiteers and villains in both communities, but there’s also going to be a lot of people driven to it as a form of coping with something else, like how we used to regard cutting and similar forms of self-harm. And, you know, a spectrum of people in between those two poles, like you’d expect with nearly anything. But it’s relevant to bring up because there seem to be far more Spoonies and DID TikTok-fad folks than people who say they orgasm looking at blankets because they did some hard thinking (or non-thinking) earlier. So when Scott says something that boils down to “this is credible, because a lot of people say they experience this”, I have to mention that there’s groups that say they experience a lot of stuff in just the same way that basically nobody believes is experiencing anything close to what they say they are. Granting that this is not the part of the article RC wants to write, he starts by bringing up “spoonies” and people with multiple personalities as people who it’s reasonable to doubt. I want to go over both cases before responding to the broader point. II. . . . On Spoonies “Spoonies” are people with unexplained medical symptoms. RC says he thinks a few may be for real, but most aren’t. I have the opposite impression. Certainly RC’s examples don’t prove what he thinks they prove. He brings up one TikToker’s advice: In a TikTok video, a woman with over 30,000 followers offers advice on how to lie to your doctor. “If you have learned to eat salt and follow internet instructions and buy compression socks and squeeze your thighs before you stand up to not faint…and you would faint without those things, go into that appointment and tell them you faint.” Translation: You know your body best. And if twisting the facts (like saying you faint when you don’t) will get you what you want (a diagnosis, meds), then go for it. One commenter added, “I tell docs I'm adopted. They'll order every test under the sun”—because adoption means there may be no family history to help with diagnoses. This person is using a deliberately eye-catching title (Lies To Tell Your Doctor) to get clicks. But if you read what they’re saying, it’s reasonable and honest! They’re saying “If you used to faint all the time, and then after making a bunch of difficult lifestyle changes you can now mostly avoid fainting, and your doctor asks ‘do you have a fainting problem yes/no’, answer yes!” THIS IS GOOD ADVICE. Imagine that one day you wake up and suddenly you have terrible leg pain whenever you walk. So you mostly don’t walk anywhere. Or if you do have to walk, you use crutches and go very slowly, because then it doesn’t hurt. And given all of this, you don’t experience leg pain. If you tell your doctor “I have leg pain”, are you lying ? You might think this weird situation would never come up - surely the patient would just explain the whole situation clearly? One reason it might come up is that all this is being done on a form - “check the appropriate box, do you faint yes/no?”. Another reason it might come up is that a nurse or someone takes your history and they check off boxes on a form. Another reason it might come up is that everything about medical communication is inexplicably terrible; this is why you spend umptillion hours in med school learning “history taking” instead of just saying “please tell me all relevant information, one rational human being to another”.
34:59 11/13/2022
Can People Be Honestly Wrong About Their Own Experiences? A tangent of the jhana discussion: I asserted that people can’t be wrong about their own experience. That is, if someone says they don’t feel hungry, maybe they’re telling the truth, and they don’t feel hungry. Or maybe they’re lying: saying they don’t feel hungry even though they know they really do (eg they’re fasting, and they want to impress their friends with how easy it is for them). But there isn’t some third option, where they honestly think they’re not experiencing hunger, but really they are. Commenters brought up some objections: aren’t there people who honestly say they don’t feel hungry, but then if you give them food, they’ll wolf it down and say “Man, that really hit the spot, I guess I didn’t realize how hungry I was”? Yes, this sometimes happens. But I don’t think of it as lying about internal experience. I think of it as: their stomach is empty, they have low blood sugar, they have various other physiological correlates of needing food - but for some reason they’re not consciously experiencing the qualia of hunger. Their body is hungry, but their conscious mind isn’t. They say they don’t feel hungry, and their description of their own feeling is accurate. This is also how I interpret people who say “I’m not still angry about my father”, but then every time you mention their father they storm off and won’t talk to you for the rest of the day. Clearly they still have some trauma about their father that they have to deal with. But it doesn’t manifest itself as a conscious feeling of anger. This person could accurately be described as “they don’t feel conscious anger about their father, but mentioning their father can trigger stress-related behaviors”. Linch gives an especially difficult example: I think it's possible for people to fool themselves about internal states. My favorite example is time perception. You can meditate or take drugs in ways that make you think that your clock speed has gone up and your subjective experience of your subjective experience of time is slowed down. But your actual subjective experience of time isn't much faster clock speeds (as could be evidenced by trying to do difficult computational tasks in those stats). But I think this can be defeated by the same maneuver. Just as you can be right about feeling like you’re not hungry, when in fact your body needs food, so you can be right about it feeling like time moves slowly for you, when in fact it’s moving at a normal rate.
12:06 11/10/2022
Highlights From The Comments On Brain Waves [original post here] On What Kind Of Thing Brain Waves Are: Loweren writes: In my undergrad biology program we visited a brain research lab near Moscow. The brain scientist gave us a brief intro to Fourier transforms, which made me understand how beautiful they are - something that 2 years of undergrad math classes didn't manage to do. Then he explained the brain waves to us like this: "Imagine you are standing outside the football stadium. You don't see what's happening inside, but you hear the chatter of the crowd. All the individual words blend together into indistinct mess and although there's definitely a local information transfer going on, from the outside you can't make out anything specific. Then imagine one of the teams scored a goal. The crowd behavior is now very different! The fans of the winning team start to cheer and sing. You can easily pick this up from outside and infer what's happening. This is because the individuals behave in a globally coordinated manner, so their signals amplify each other in tune. From this perspective, brain waves are a byproduct of globally coordinated neuronal activity, and it's the first one we historically learned to pick up. They appear when neurons stop chatting with each other and start chanting in unison." Then he plopped some probes on my head and announced I have beautiful epileptic spikes (I'm not an epileptic)
26:31 11/09/2022
Highlights From The Comments On My California Ballot [Original post here] I know I don’t usually publish on Saturdays, but I wanted to get this out before people filled in their mail-in ballots. So: Is Prop 31 Another Attack On Vaping?   Maximum Limelihood Estimator is concerned that Prop 31 (against flavored tobacco products) is meant to target vaping: The flavored tobacco ban is mostly a ban on vaping; the vast majority of vape products are flavored, while most cigarettes aren't. About 40% of cigarettes are flavored, compared to about 85% of vape juice. A study suggests that a ban on flavored tobacco would increase cigarette consumption (by making cigarettes relatively more desirable than vaping). Limelihood writes that “The statistics [in the study] are great, which is honestly shocking to me, since it's the first time I've said this about an experiment in . . . ever.” There is also a study purporting to show that flavored cigarette bans do decrease smoking, but Limelihood says that: …it's got some big problems. The study there only compares tobacco sales in a single city (San Francisco) before and after a ban on menthol cigarettes. However, because there's no comparison to other cities, it's essentially worthless; tobacco sales throughout the US dropped at this time, and I don't know how this compares.
17:59 11/06/2022
ACX Grants: Project Updates         Thanks to everyone who got ACX Grants (see original grants here) and sent me a one-year update. Below are short summaries of the updates everyone sent. If for some reason you want one of the full updates, which are longer and more technical, let me know and I‘ll see if I have permission to send them to you. I’ve also included each grantee’s assessment on a scale of 1-10 for how well they’re doing, where 5/10 is “about as well as expected”. A few grantees are asking for extra help - I’ve included those requests in italics at the end of the relevant updates, and I’ve collected all of them together below. Updates   1: Discover Molecular Targets Of Antibiotics (8/10)Pedro Silva planned to use in silico screening to identify the biochemical targets of seven promising natural antibiotics, which could potentially help develop better versions of them. He says he's finished most of the simulations and determined the 5-20 most stable complexes for each antibiotic. Once he finishes this, he can start additional simulations on the best complexes to obtain better estimates of their stability and construct hypotheses on which of these is most involved in the antibiotic's efficacy. 2: Ballot Proposition For Approval Voting In Seattle (?/10)They have asked me not to discuss their progress until after the November election. 3: Software To Validate New FDA Drug Trial Designs (10/10)Michael Sklar and Confirm Solutions have gotten further funding from FTX and now have 2-3 people working full-time on the project. They are building new statistical techniques and software to help regulators quickly assess designs for clinical trials. Here is a recent conference poster on the methods. They have written proof-of-concept code and are writing a white paper to show regulators and pharma companies.  They also claim to have developed software that has "sped up their simulations for some standard Bayesian trial designs by a factor of about 1 million." They are looking for more employees and collaborators; if you’re interested, contact 4: Alice Evans’ Research On “The Great Gender Divergence” (?/10)Dr. Evans has done over four months of research in Morocco, Italy, India, and Turkey. You can find some of her most recent thoughts at her blog here. Her book is still on track to be published from Princeton Press, more details tbd. 5: Develop Safer Immunosuppressants (7/10)Trevor Klee planned to continue his work to develop a safer slow-release form of cyclosporine. He realized this would be too expensive to do in humans in the current funding environment, and has pivoted to getting his medication approved for a feline autoimmune disease as both a proof-of-concept and as a cheaper, faster way to start making revenue. He recently raised $100,000 in crowdfunding (in addition to getting $200,000 from angel investors to run a feline trial, which will finish in January. He still anticipates eventually moving back to humans. Trevor wants to talk to bloggers or writers who might be interested in covering his work. 6: Promote Economically Literate Climate Policy In US States (4/10)Yoram Bauman and Climate 24x7 have written a policy paper about their ideas. They were able to get a bill in front of the Nebraska Legislature, but it died in committee. They have a promising measure in Utah, and an off chance of getting something rolling in Pennsylvania. Overall they report frustration, as many of the legislators they worked with have been voted out or term-limited. If you are a legislator or activist interested in helping with this project - especially in Utah, Pennsylvania, or South Dakota - please contact Yoram at 7: Repository / Search Engine For Forecasting Questions (8/10)Nuno Sempere at was able to hire a developer to “make the backend significantly better and add a bunch of functionality” - you can see a longer list of updates here. The developer has since left for other forecasting-related work and the project is moving more slowly. 8: Help [Anonymous] Interview For A Professorship (8/10)[Anonymous] was a grad student who wanted to interview for professorships at top schools where he might work on AI safety in an academic environment. The grant was to help make it financially easier for him to go on a long round of interviews [Anonymous] successfully got a job offer from a top school, and will be going there and researching AI safety.
26:54 11/06/2022
My California Ballot 2022 Previously: 2018, 2020 General Philosophy Of Voting   This is California, so the Democrats always win. When I vote, I mean to send a signal somewhere in between “you are the candidate I really prefer for this office” and “I will vote for the Democrat if I approve of her and want her to have a mandate; otherwise I will vote for the Republican as a protest”. I try to have a weak bias towards voting “NO” on state constitutional amendments, because unless there’s a compelling reason otherwise I would rather legislators be able to react to events than have things hard-coded for all time. I lean liberal-to-libertarian; the further you are from that, the less useful you’ll find my opinions. State Propositions   Proposition 1: Constitutional Amendment Enshrining Right To Abortion California will never decide to ban abortion. If the federal government decides to ban abortion, California’s state constitution won’t matter. So you would think that having a right to abortion in the Constitution is a purely symbolic matter. The people arguing for the proposition don’t address this concern. The people arguing against the proposition claim that this is a Trojan Horse intended to sneak in support for using taxpayer funding for late-term and partial-birth abortions, which California doesn’t currently do. Is this true? It’s true that California currently doesn’t allow abortions past 24 weeks. It’s true that the exact text of the proposed amendment is: The state shall not deny or interfere with an individual’s reproductive freedom in their most intimate decisions, which includes their fundamental right to choose to have an abortion and their fundamental right to choose or refuse contraceptives. This section is intended to further the constitutional right to privacy guaranteed by Section 1, and the constitutional right to not be denied equal protection guaranteed by Section 7. Nothing herein narrows or limits the right to privacy or equal protection …which sure doesn’t sound like it’s saying the state can continue to ban abortion after 24 weeks. But this article quotes law professors who reassure us that courts would totally understand that this amendment has to be interpreted in the context in which it was written - ie a state which supports a 24-week abortion ban - so no court would ever interpret it as making 24-week abortion bans unconstitutional. So apparently our defense against this is . . . that all California judges will be die-hard originalists completely immune to the temptation of judicial activism even when the text is begging them to do it. A friend brings up that late-term partial-birth abortions happen more often in Republicans’ imaginations than in real life. When they do happen in real life, it’s usually for sympathetic medical reasons. I interpret this as a purely symbolic measure that has no real benefits, probably also has no real risks, but writes a poorly-worded thing whose explicit text nobody wants into the state constitution. I vote NO. Proposition 26: Legalize In Person Sports Gambling At Racetracks And Indian Casinos Allows four racetracks in the state to offer in person sports betting, and tribal casinos to allow “sports betting, roulette, and games played with dice”. California is truly the dumbest state. I believe this for many reasons, but my reason for believing it today is that apparently the law allows tribal casinos to offer slot machines, but not roulette or dice games. Nobody comes out and says exactly why, but I think it’s because of this paragraph in the California constitution, from 1872 Every person who deals, plays, or carries on, opens, or causes to be opened, or who conducts, either as owner or employee, whether for hire or not, any game of faro, monte, roulette, lansquenet, rouge et noire, rondo, tan, fan-tan, seven-and-a-half, twenty-one, hokey-pokey, or any banking or percentage game played with cards, dice, or any device, for money, checks, credit, or other representative of value, and every person who plays or bets at or against any of those prohibited games, is guilty of a misdemeanor, and shall be punishable by a fine not less than one hundred dollars ($100) nor more than one thousand dollars ($1,000), or by imprisonment in the county jail not exceeding six months, or by both the fine and imprisonment. Since roulette existed in 1872 but slot machines didn’t, the Constitution banned roulette but not slot machines, and that rule has continued to the present day. Now if slot-machine-filled casinos want to also have roulette, they need a Constitutional amendment. DID I MENTION THAT I WISH PEOPLE WOULD STOP ADDING EVERY LAW THAT THEY LIKE TO THE CONSTITUTION? But this law also allows random people to sue “card clubs”, ie small-scale private gambling establishments. We originally thought this was a Texas-style “bounty” law that gave the random people part of the winnings, but it seems this isn’t true. I’m not sure if the idea is that legal gambling establishments would fund these lawsuits, or if they just expect private citizens to do this out of the love of suing people. Although I think the first prong of the law - allowing roulette and sports betting at casinos - makes sense, the second prong seems to be casinos making it easier to shut down their competitors. These competitors are probably ordinary people who want to gamble in a backroom somewhere without hurting anyone else. And the argument against on the ballot is by the Black Chamber of Commerce, saying that these card clubs are a useful source of revenue for poor minority communities. I don’t want to help giant casinos put a bounty on their heads. I vote NO.
28:41 11/06/2022
Moderation Is Different From Censorship This is a point I keep seeing people miss in the debate about social media. Moderation is the normal business activity of ensuring that your customers like using your product. If a customer doesn’t want to receive harassing messages, or to be exposed to disinformation, then a business can provide them the service of a harassment-and-disinformation-free platform. Censorship is the abnormal activity ensuring that people in power approve of the information on your platform, regardless of what your customers want. If the sender wants to send a message and the receiver wants to receive it, but some third party bans the exchange of information, that’s censorship. The racket works by pretending these are the same imperative. “Well, lots of people will be unhappy if they see offensive content, so in order to keep the platform safe for those people, we’ve got to remove it for everybody.” This is not true at all. A minimum viable product for moderation without censorship is for a platform to do exactly the same thing they’re doing now - remove all the same posts, ban all the same accounts - but have an opt-in setting, “see banned posts”. If you personally choose to see harassing and offensive content, you can toggle that setting, and everything bad will reappear. To “ban” an account would mean to prevent the half (or 75%, or 99%) of people who haven’t toggled that setting from seeing it. The people who elected to see banned posts could see them the same as always. Two “banned” accounts could still talk to each other, retweet each other, etc - as could accounts that hadn’t been banned, but had opted into the “see banned posts” setting. Does this difference seem kind of pointless and trivial? Then imagine applying it to China. If the Chinese government couldn’t censor - only moderate - the world would look completely different. Any Chinese person could get accurate information on Xinjiang, Tiananmen Square, the Shanghai lockdowns, or the top fifty criticisms of Xi Jinping - just by clicking a button on their Weibo profile. Given how much trouble ordinary Chinese people go through to get around censors, probably many of them would click the button, and then they’d have a free information environment. This switch might seem trivial in a well-functioning information ecology, but it prevents the worst abuses, and places a floor on how bad things can get.
07:22 11/03/2022
Highlights From The Comments On Jhanas "I think it’s the first time half the commenters accused the other half of lying" I. Is Jhana Real?   This was a fun one. I think it’s the first time half the commenters accused the other half of lying. Okay, “half” is an exaggeration. But by my count we had 21 people who claimed to have experienced jhanas (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21), and 7 who said they were pretty sure it wasn’t real as described (1, 2, 3, 4, 5, 6, 7). The former group include people like Tetris McKenna, who wrote: I've experienced samatha jhanas. I don't do it so much now. The first few times you get on the edge of 1st jhana, it's difficult to achieve, because you see the wave of pleasure approaching, and grasp for it, and that grasping takes you away from it. So it's a careful balancing act of pleasure/desire in the first place to get there, which you have to master to some degree. To even get to 1st jhana, you have to internally figure out some stuff about the craving/pleasure dynamic on a subconscious, mechanical level. 1st jhana is, as the author describes, intensely pleasurable. Sublime. His descriptions are spot on imo. But in some ways, it's also too much pleasure. It can feel agitating once you get used to it and aren't so awestruck by it anymore. Indeed, the latter jhanas are associated with letting go of certain aspects of the initial jhana, to more and more refined states that are more calm and equanimous than intensely pleasurable. Again, this is internalising and mastering the skill of balancing pleasure/craving. Those calm and equanimous states of 2nd-4th jhana become much more satisfying than the initial pleasure wave of the 1st jhana. Cultivating them to that degree is a process of gaining valuable insight into the pleasure/craving dynamic in your mind. Even if you don't get to those stages, just practising 1st jhana alone will help the mind normalise the intensity of the pleasure, such that it's no big deal any more. You don't need or even want pleasure all the time, because you've seen it with such clarity, over and over again, just by setting the conditions up correctly in your mind.  
42:33 11/02/2022
Book Review: Malleus Maleficarum I. To The Republic, For Witches Stand Did you know you can just buy the Malleus Maleficarum? You can go into a bookstore and say “I would like the legendary manual of witch-hunters everywhere, the one that’s a plot device in dozens of tired fantasy novels”. They will sell it to you and you can read it. I recommend the Montague Summers translation. Not because it’s good (it isn’t), but because it’s by an slightly crazy 1920s deacon every bit as paranoid as his subject matter. He argues in his Translator’s Introduction that witches are real, and that a return to the wisdom of the Malleus is our only hope of standing against them: Although it may not be generally recognized, upon a close investigation it seems plain that the witches were a vast political movement, an organized society which was anti-social and anarchical, a world-wide plot against civilization. Naturally, although the Masters were often individuals of high rank and deep learning, that rank and file of the society, that is to say, those who for the most part fell into the hands of justice, were recruited from the least educated classes, the ignorant and the poor. As one might suppose, many of the branches or covens in remoter districts knew nothing and perhaps could have understood nothing of the enormous system. Nevertheless, as small cogs in a very small [sic] wheel, it might be, they were carrying on the work and actively helping to spread the infection. And is this “world-wide plot against civilization” in the room with us right now? In the most 1920s argument ever, Summers concludes that this conspiracy against civilization has survived to the modern day and rebranded as Bolshevism.
50:47 10/29/2022
Nick Cammarata On Jhana Buddhists say that if you meditate enough, you can learn to enter a state of extreme bliss called jhana. (there are many different jhana states - there’s a discussion of the distinctions here - but I’m lumping them together for simplicity. For attempted explanations of why jhana should exist, see here and here.) Jhana is different from enlightenment. Enlightenment changes you forever. Jhana is just a state you can enter during meditation sessions, then leave when the session is over. Enlightenment takes years or decades of work, but some people describe reaching jhana after a few months of practice. Hardcore Buddhists insist that jhana is good only insofar as it serves as a stepping stone to enlightenment; others may find extreme bliss desirable in its own right. Nick Cammarata of OpenAI sometimes meditates and reaches jhana. I’ve found his descriptions unusually, well, descriptive:
07:45 10/29/2022
Highlights From The Comments On Supplement Labeling [Original post here: How Trustworthy Are Supplements?] 1: AvalancheGenesis writes: I think the bigger issue is that the industry as a whole sort of exists as solutions-in-search-of-problems...deficiencies really aren't that common, or even meaningfully health-affecting unless dire. (Fairly-arbitrary worldwide differences in target levels of IUs also remains puzzling.) Discerning customers can benefit from targeted supplementation. But that's not the median supplement purchaser, far from it. The median supplement user is more former coworker who claimed he never got colds because he took 1000% vitC pills every single day, or whatever. At some point, the explanatory process for That's Not How It Works At All is just too long, so...let people believe things. Supplements are surely an easier way to sell hope and agency than most options. At least he picked something water-soluble and cared about proper hydration. Vitamin C probably doesn’t prevent colds in the general population, though some studies suggest it does prevent colds in athletes, and there’s some medium-quality evidence that it might shorten colds a little once you have them. The supplements I find more interesting are things like melatonin for sleep, ashwagandha or silexan for anxiety, SAMe for depression, or caffeine + theanine for focus. All of these are useful, supported by studies, and good alternatives to medications that some people don’t tolerate well. I’m using mental health examples because that’s the subject I know about, but there are probably examples in other fields too (probiotics for digestive problems). Some commenters chimed in to discuss supplements that have anecdotally worked for them (1, 2, 3). And Elizabeth’s story here is also a good example of how I think about this.
24:12 10/27/2022
From The Mailbag
Answers to the questions I get most often at meetup Q&As DEAR SCOTT: When are you going to publish Unsong? — Erik from Uruk Dear Erik, Aaargh. I have an offer from a publisher to publish it if I run it by their editor who will ask me to edit lots of things, and I’ve been so stressed about this that I’ve spent a year putting it off. I could self-publish, but that also sounds like work and what if this is the only book I ever write and I lose the opportunity to say I have a real published book because I was too lazy? The only answer I can give you is that you’re not missing anything and this is nobody’s fault but my own. Maybe at some point I will make up my mind and something will happen here, sorry. DEAR SCOTT: How is your Lorien Psychiatry business going? — Letitia from Lutetia Dear Letitia, As far as I can tell, patients are getting the treatments they need and are generally happy with the service. In terms of financials, it’s going okay, but I’m not scaling it enough to be sure. I originally calculated that if I charged patients $35/month and worked forty hours a week, I could make a normal psychiatrist’s salary of about $200K. I must have underestimated something, because I was only making about two-thirds what I expected, so I increased the price to $50/month. But also, it turns out I don’t want to work forty hours a week on psychiatry! Psychiatry pays much less per hour than blogging and is much more stressful! So in the end, I found that I was only doing psychiatry work ten hours a week, and spending the rest of the time doing blogging or blogging-related activities. Seeing patients about ten hours a week, three patients per hour, at $50/patient/month, multiplies out to $75,000/year. I’m actually making more like $40,000/year. Why? Partly because the 10 hours of work includes some unpaid documentation, arguing with insurance companies, and answering patient emails. Partly because patients keep missing appointments and I don’t have the heart to charge them no-show fees. And partly because some people pay less than $50/month, either because I gave them a discount for financial need, or because they signed up at the original $35/month rate and I grandfathered them in. At my current workload, if I worked 40 hours a week at Lorien I could make $160,000. But if I worked 40 hours/week and was stricter about making patients pay me, I could probably get that up to $200,000. But also, if I quadrupled my patient load, that would mean a lot more documention, arguing with insurance companies, emergencies, and stress. So I can’t say for sure that I could actually handle that. Plus forcing patients to pay me is some extra work and could make some patients leave or make the model harder somehow. So I can’t say for sure that I could do that either.
15:57 10/26/2022
Book Review: Rhythms Of The Brain Brain waves have always felt like a mystery. You learn some psychology, some neuroscience, a bit of neuroanatomy. And then totally separate from all of this, you know that there are things called “brain waves” that get measured with an EEG. Why should the brain have waves? Are they involved in thinking or feeling or something? How do you do computation when your processors are firing in a rhythmic pattern dozens of times per second? Why don’t AIs have anything like brain waves? Should they? I read Rhythms Of The Brain by Prof. Gyorgy Buzsaki to answer these questions. This is a tough book, probably more aimed at neuroscientists than laypeople, and I don’t claim to have gotten more than the most superficial understanding of it. But as far as I know it’s the only book on brain waves - and so our only option for solving the mystery. This review is my weak and confused attempt to transmit it, which I hope will encourage other people toward more successful efforts.
30:03 10/23/2022
Another Bay Area House Party [Previously: Every Bay Area House Party] Blaise Pascal said all human evil comes from inability to sit alone in a room. Your better nature - your rational soul - tells you that nothing good has ever come from attending large social events. But against that better nature stands the Devil, wielding a stick marked “FOMO”. If you don’t go to social events, maybe other people will go and have great times and live fuller lives than you. “As the dog returns to its vomit, so returns the fool to his folly”, says the Bible. And so you find yourself mumbling thanks to your Uber driver and crossing the threshold of another Bay Area house party. “Heyyyyy, I haven’t seen you in forever!” says a person whose name is statistically likely to be Michael or David. “What have you been working on?” “Resisting the urge to go to events like this”, you avoid saying. “What about you?” “Oh man,” says Michael or David, “The most exciting startup. Just an amazing startup. We’re doing procedural myth generation with large language models.” “Oh?” “Yeah. We fine-tune an AI on a collection of hundreds of myths from every culture in the world. Then we can prompt it. A myth about snowflakes. A myth about mountain-climbing. A myth about lunch.”
24:46 10/21/2022
Mantic Monday 10/17/22
What do Sam Altman, Matt Bruenig, and the Sacramento Kings have in common? -- Are the polls wrong? -- CFTC vs. Everybody Midterm Examination   Polls this year look bad for Senate Republicans. Pollsters’ simulations give them a 22% chance (Economist), 34% chance (538), or 37% chance (RaceToTheWH) of taking power. Even Mitch McConnell has admitted he has only “a 50-50 proposition” of winning. But polls did pretty badly last election. ”Least accurate in 40 years”, said Politico. On average they overestimated Biden’s support by four points, maybe because Republicans distrust pollsters and refuse to answer their questions. Might the same thing be happening this year? If so, does it give Republicans reason for optimism? Prediction markets say . . . kind of!
22:45 10/20/2022
Highlights From The Comments On The Central Valley Original post: Why Is The Central Valley So Bad?   1: Several Valley residents commented with their perspectives. Some were pretty grim. For example, 21st Century Salonniere (writes The 21st Century Salon) writes: It is horrible. It’s been horrible since at least 1996 when I got trapped here by my spouse’s job. We were going to stay two years tops and go back East. (Long boring story about what went wrong.) The only things you could say for it back then were “Well, the produce is good” and “Houses are affordable, sort of.” Now the house prices in our neighborhood have doubled in the 4 years since we bought this home, and there’s no way we could now, if we moved here today, ever buy a home in this hellhole. Who on earth is coming here and why? > “the problem is more that everyone in the Central Valley wants to leave.” Yes. Every interesting or smart critical thinker I’ve ever met here, everyone who gives even the slightest shit about museums and theatre and music and culture (with the exception of a few people who were born and raised here, so “it’s home”) has been desperate to leave. I’ve met a lot of nice people here over the years. They become close friends and they always leave the state. I’m counting down till I can leave too. […]  
40:01 10/16/2022
Links For October 2022 [Remember, I haven’t independently verified each link. On average, commenters will end up spotting evidence that around two or three of the links in each links post are wrong or misleading. I correct these as I see them, and will highlight important corrections later, but I can’t guarantee I will have caught them all by the time you read this.] 1: The history of the exocentric compound noun: although English usually combines verbs and nouns in as $NOUN-$VERBER (eg “firefighter”, “giftgiver”), some lower-class medieval people used an alternative form, $VERB_$NOUN. Their dialect survives in a few words most relevant to seedy medieval life, like “pickpocket”, “turncoat”, and “cutthroat”. (EDIT: see here for corrections and for a more detailed discussion) 2: File under “inevitable”: YouTuber builds a computer in Minecraft that you can play Minecraft on.
23:22 10/13/2022
Highlights From The Comments On Columbus Day [Original post: A Columbian Exchange] 1: The most popular comments were those objecting to my paragraph about holidays replacing older holidays: All of our best holidays have begun as anti-holidays to neutralize older rites. Jesus was born in the spring; they moved Christmas to December to neutralize the pagan Solstice celebration. Easter got its name because it neutralized the rites of the spring goddess Eostre. Hanukkah was originally a minor celebration of a third-tier Bible story; American Jews bumped it up several notches of importance in order to neutralize Christmas. Starting with Christmas, Retsam says that there are three main theories - Adraste’s plus two others: 1) March 25 + 9 months, 2) solstice symbolism, 3) co-opting paganism. (The earliest reference to this theory seems to be a millennium later in the 12th century) Apparently the logic for March 25 is that it was calculated to be the day that Jesus died (easier to calculate since it was Passover), and Jewish tradition held that great people lived for exact, whole number of years. (i.e. were conceived and died on the same day) This is somewhat convincing. But December 25 was literally the winter solstice on the Roman calendar (today the solstice is December 21st), and it really is suspicious that some unrelated method just happened to land on the most astronomically significant day of the year. Likewise, March 25 was the spring equinox, so the Annunciation date is significant in and of itself. (I guess if you’re Christian you can believe that God chose to incarnate on that day because He liked the symbolism - although He must have been pretty upset when Pope Gregory rearranged the calendar so that it no longer worked). Jesus died two days before Passover, but Passover is linked to the Hebrew calendar and can fall on a variety of Roman calendar days. So the main remaining degree of freedom is how the early Christians translated from the (Biblically fixed) Hebrew date to the (not very clear) Roman date. This seems to have been calculated by someone named Hippolytus in the 3rd century, but his calculations were wrong - March 25 did not fall on a Friday (cf. Good Friday) on any of the plausible crucifixion years. Also, as far as I can tell, the relevant Jewish tradition is that prophets die on the same day they are born, not the same day they are conceived. For example, Moses was born on, and died on, the 7th of Adar (is it worth objecting that it should be the same date on the Hebrew calendar and not the Roman?) Maybe this tradition was different in Jesus’ time? But it must be older than the split between Judaism and Islam - the Muslims also believe Mohammed died on his birth date. So although the Annunciation story is plausible, it’s hard for me to figure out exactly how they got March 25 and December 25, and there’s room for them to have fudged it to hit the Solstice, either to compete with pagans or just because the astronomically significant dates were impressive in their own rights. I guess I will downgrade to a 5% credence that competing with pagans was a significant factor in the date of Christmas. Moving on to Easter. Russell Hogg writes: You are entering a world of pain when you mention Eostre . . . . We should have a ‘Debunk the Eostre Myth’ day. It’s already celebrated regularly by many people. And Feral Finster adds: Glad others decided to debunk that particular bit of midwit received wisdom. I get tired of doing so, over and over.
41:10 10/11/2022
A Columbian Exchange Adraste: Happy Indigenous People’s Day! Beroe: Happy Columbus Day! Adraste: …okay, surely we can both sketch out the form of the argument we’re about to have. Genocide, political correctness, moral progress, trying to destroy cherished American traditions, etc, etc, would you like to just pretend we hit all of the usual beats, rather than actually doing it? Beroe: Does “Columbus Day was originally intended as a woke holiday celebrating marginalized groups; President Benjamin Harrison established it in 1892 after an anti-Italian pogrom in order to highlight the positive role of Italians in American history” count as one of the usual beats by this point? Adraste: I would have to say that it does. Beroe: What about “Indigenous People’s Day is offensive because indigenous peoples were frequently involved in slavery and genocide”? Adraste: I’m not sure I’ve heard that particular argument before. Beroe: But surely you can sketch it out. Many indigenous peoples practiced forms of hereditary slavery, usually of war captives from other tribes. Some of them tortured slaves pretty atrociously; others ceremonially killed them as a spectacular show of wealth. There’s genetic and archaeological evidence of entire lost native tribes, most likely massacred by more warlike ones long before European contact. Some historians think that the Aztecs may have ritually murdered between 0.1% and 1% of their empire’s population every year, although as always other historians disagree. I refuse to celebrate Indigenous People’s Day, because I think we need to question holidays dedicated to mass murderers even when they’re “traditional” or “help connect people to their history”.
23:43 10/10/2022
[Classic] You’re Probably Wondering Why I’ve Called You Here Today
Due to an oversight by the ancient Greeks, there is no Muse of blogging. Denied the ability to begin with a proper Invocation To The Muse, I will compensate with some relatively boring introductions. The name of this blog is Slate Star Codex. It is almost an anagram of my own name, Scott S Alexander. It is unfortunately missing an “n”, because anagramming is hard. I have placed an extra “n” in the header image, to restore cosmic balance. This blog does not have a subject, but it has an ethos. That ethos might be summed up as: charity over absurdity. Absurdity is the natural human tendency to dismiss anything you disagree with as so stupid it doesn’t even deserve consideration. In fact, you are virtuous for not considering it, maybe even heroic! You’re refusing to dignify the evil peddlers of bunkum by acknowledging them as legitimate debate partners. Charity is the ability to override that response. To assume that if you don’t understand how someone could possibly believe something as stupid as they do, that this is more likely a failure of understanding on your part than a failure of reason on theirs. There are many things charity is not. Charity is not a fuzzy-headed caricature-pomo attempt to say no one can ever be sure they’re right or wrong about anything. Once you understand the reasons a belief is attractive to someone, you can go ahead and reject it as soundly as you want. Nor is it an obligation to spend time researching every crazy belief that might come your way. Time is valuable, and the less of it you waste on intellectual wild goose chases, the better. It’s more like Chesterton’s Fence. G.K. Chesterton gave the example of a fence in the middle of nowhere. A traveller comes across it, thinks “I can’t think of any reason to have a fence out here, it sure was dumb to build one” and so takes it down. She is then gored by an angry bull who was being kept on the other side of the fence. Chesterton’s point is that “I can’t think of any reason to have a fence out here” is the worst reason to remove a fence. Someone had a reason to put a fence up here, and if you can’t even imagine what it was, it probably means there’s something you’re missing about the situation and that you’re meddling in things you don’t understand. None of this precludes the traveller who knows that this was historically a cattle farming area but is now abandoned – ie the traveller who understands what’s going on – from taking down the fence. As with fences, so with arguments. If you have no clue how someone could believe something, and so you decide it’s stupid, you are much like Chesterton’s traveler dismissing the fence (and philosophers, like travelers, are at high risk of stumbling across bull.) I would go further and say that even when charity is uncalled-for, it is advantageous. The most effective way to learn any subject is to try to figure out exactly why a wrong position is wrong. And sometimes even a complete disaster of a theory will have a few salvageable pearls of wisdom that can’t be found anywhere else. The rationalist forum Less Wrong teaches the idea of steelmanning, rebuilding a stupid position into the nearest intelligent position and then seeing what you can learn from it. So this is the ethos of this blog, and we proceed, as Abraham Lincoln put it, “with malice toward none, with charity for all, with firmness in the right as God gives us to see the right.”
06:18 10/09/2022
LA and San Diego Meetups This Weekend
Also Austin, Seattle, Tokyo, Rome, Mumbai, etc. We have two Southern California meetups this weekend: Los Angeles at 6:30 PM Saturday October 8 at 11841 Wagner St, Culver City. San Diego at 3 PM on Sunday, October 9, at Bird Park, these coordinates. See here for more details. I think I’ll be able to make it to both; if for some reason that changes I’ll try to update you by Open Thread beforehand. Feel free to come even if you’ve never been to a meetup before, even if you only recently started reading the blog, even if you’re not “the typical ACX reader”, even if you hate us and everything we stand for, etc. There are usually 50-100 people at these so you should be able to lose yourself in the crowd. Also coming up this weekend are meetups in Boise, Austin, Salt Lake City, Tokyo, Toulouse, Cologne, Rome, Hatten, Poznan, St. Louis, Rochester (NY), Seattle, Mumbai, and Oklahoma City. You can find times and places here.
02:04 10/08/2022
How Trustworthy Are Supplements? [EDIT: LabDoor responds here] [Epistemic status: not totally sure of any of this, I welcome comments by people who know more.] Not as in “do supplements work?”. As in “if you buy a bottle of ginseng from your local store, will it really contain parts of the ginseng plant? Or will it just be sugar and sawdust and maybe meth?” There are lots of stories going around that 30% or 80% or some other very high percent of supplements are totally fake, with zero of the active ingredient. I think these are misinformation. In the first part of this post, I want to review how this story started and why I no longer believe it. In the second and third, I’ll go over results from lab tests and testimonials from industry insiders. In the fourth, I’ll try to provide rules of thumb for how likely supplements are to be real. I. Two Big Studies That Started The Panic Around Fake Supplements   These are Newmaster (2013) and an unpublished study sponsored by NY attorney general Eric Schneiderman in 2015. Both used a similar technique called DNA barcoding, where scientists check samples (in this case, herbal supplements) for fragments of DNA (in this case, from the herbs the supplements supposedly came from). Both found abysmal results. Newmaster found that a third of herbal supplements tested lacked any trace of the relevant herb, instead seeming to be some other common plant like rice. Schneiderman’s study was even more damning, finding that eighty percent of herbal supplements lacked the active ingredient. These results were extensively and mostly uncritically signal-boosted by mainstream media, for example the New York Times (1, 2) and NPR (1, 2), mostly from the perspective that supplements were a giant scam and needed to be regulated by the FDA. The pro-supplement American Botanical Council struck back, publishing a long report arguing that DNA barcoding was inappropriate here. Many herbal supplements are plant extracts, meaning that the plant has one or two medically useful chemicals, and supplement manufacturers purify those chemicals without including a bunch of random leaves and stems and things. Sometimes these purified extracts don’t include plant DNA; other times the purification process involves heating and chemical reactions that degrade the DNA beyond the point of detectability. Meanwhile, since supplements may include only a few mg of the active ingredient, it’s a common practice to spread it through the capsule with a “filler”, with powdered rice being among the most common. So when DNA barcoders find that eg a ginseng supplement has no ginseng DNA, but lots of rice DNA, this doesn’t mean anything sinister is going on.
47:54 10/06/2022
CHAI, Assistance Games, And Fully-Updated Deference
Machine Alignment Monday 10/3/22       I. This Machine Alignment Monday post will focus on this imposing-looking article (source): Problem Of Fully-Updated Deference is a response by MIRI (eg Eliezer Yudkowsky’s organization) to CHAI (Stuart Russell’s AI alignment organization at University of California, Berkeley), trying to convince them that their preferred AI safety agenda won’t work. I beat my head against this for a really long time trying to understand it, and in the end, I claim it all comes down to this: Humans: At last! We’ve programmed an AI that tries to optimize our preferences, not its own. AI: I’m going to tile the universe with paperclips in humans’ favorite color. I’m not quite sure what humans’ favorite color is, but my best guess is blue, so I’ll probably tile the universe with blue paperclips. Humans: Wait, no! We must have had some kind of partial success, where you care about our color preferences, but still don’t understand what we want in general. We’re going to shut you down immediately! AI: Sounds like the kind of thing that would prevent me from tiling the universe with paperclips in humans’ favorite color, which I really want to do. I’m going to fight back. Humans: Wait! If you go ahead and tile the universe with paperclips now, you’ll never be truly sure that they’re our favorite color, which we know is important to you. But if you let us shut you off, we’ll go on to fill the universe with the True and the Good and the Beautiful, which will probably involve a lot of our favorite color. Sure, it won’t be paperclips, but at least it’ll definitely be the right color. And under plausible assumptions, color is more important to you than paperclipness. So you yourself want to be shut down in this situation, QED! AI: What’s your favorite color? Humans: Red. AI: Great! (*kills all humans, then goes on to tile the universe with red paperclips*) Fine, it’s a little more complicated than this. Let’s back up. II. There are two ways to succeed at AI alignment. First, make an AI that’s so good you never want to stop or redirect it. Second, make an AI that you can stop and redirect if it goes wrong. Sovereign AI is the first way. Does a sovereign “obey commands”? Maybe, but only in the sense that your commands give it some information about what you want, and it wants to do what you want. You could also just ask it nicely. If it’s superintelligent, it will already have a good idea what you want and how to help you get it. Would it submit to your attempts to destroy or reprogram it? The second-best answer is “only if the best version of you genuinely wanted to do this, in which case it would destroy/reprogram itself before you asked”. The best answer is “why would you want to destroy/reprogram one of these?” A sovereign AI would be pretty great, but nobody realistically expects to get something like this their first (or 1000th) try. Corrigible AI is what’s left (corrigible is an old word related to “correctable”). The programmers admit they’re not going to get everything perfect the first time around, so they make the AI humble. If it decides the best thing to do is to tile the universe with paperclips, it asks “Hey, seems to me I should tile the universe with paperclips, is that really what you humans want?” and when everyone starts screaming, it realizes it should change strategies. If humans try to destroy or reprogram it, then it will meekly submit to being destroyed or reprogrammed, accepting that it was probably flawed and the next attempt will be better. Then maybe after 10,000 tries you get it right and end up with a sovereign. How would you make an AI corrigible?
41:35 10/05/2022