One post I've long wished you would write is "Sugar: Much More Than You Wanted To Know", particularly now that you have children and you will have to decide on their sugar access and consumption.
* What is your confidence that sugar is unhealthy beyond excess calories?
* What limits (daily? monthly?) would you place on children's sugar intake, ignoring practical concerns and complications?
* How much (dietary % ?) do you predict your children will realistically consume during their first decade?
* Anything else you think about sugar, or present American children's diets?
You've shown interest in Jhanas. What about glimpse practices/non-dual awareness. These were made popular mostly by Sam Harris in his app in recent times I believe.
Getting there feels less structured than Jhanas, but the feeling of doing it is almost like activating a passive skill in an RPG. But you can do it whenever you like, and it just makes you feel good about everything and allows you to relax immediately about whatever important crisis you think is super important right now, but inevitably you'll wonder why you cared so much. Quite useful and incredibly interesting.
You've talked before about how your blogging topics have shifted with your interests and with age. What is the most important / most fun-to-write topic that you no longer blog about?
You said recently [1] that you'd "like to be more observant, but the prospect of following Jewish law in its entirety feels overwhelming, and picking and choosing feels too unprincipled."
How come you're interested in observance if you're not a believer? Did you find any practice here that works for you?
Have you ever closely looked at the evidence for big recent Catholic miracles? e.g. human heart tissue in Eucharistic host, the Fatima apparition, healings at Lourdes
(for disclosure, I'm a Christian but not a Catholic)
Even Catholics are not obliged to believe in those miracles, so it's something that could be extremely contentious. Faked relics and miracles have been knocking around forever, more or less; see Chaucer poking fun with the likes of the Pardoner who admits he just produces animal bones, claim they are saints' relics, and gets the sheep to be fleeced to venerate them and pay for the privilege.
Do you have a take on all the recent alien stuff / have you looked into it (the hearings / people in the goverment saying that there have been all these UFO incidents historically)?
when Aaron says "I was doing this. I was making this happen." in Malia Ngo's office. Although technically that's a reference to Sweet Bro and Hella Jeff, which is like a Homestuck spinoff.
Does progress actually make people durably happier? (I feel like it can't. GDP/PPP per capita has increased 20x or 30x over the past 1000 years. If progress increases happiness, then... man, people living 1000 years ago must have been SO depressed!)
Yeah, I also find this confusing. Rich people are on average happier than poor people, and people in rich countries are on average happier than people in poor countries, but these are all smaller effects than I'd expect. I have more about this at https://slatestarcodex.com/2016/07/20/things-probably-matter/ .
But here's a useful thought experiment - would your kids dying make you sad? It seems like it has to, right, this is basically the definition of tragedy. But 50% of kids died before adulthood in the past, and very few do today. So if there hasn't been any happiness change since then, it condemns our belief that kids-not-dying creates happiness as much as our belief that economic growth does.
I think there's got to be something wrong with this - either with the research itself, or with some kind of happiness treadmill that scales back how we feel about most things (not everything) as things improve. See also https://www.astralcodexten.com/p/unpredictable-reward-predictable .
My intuition is that richness leads to happiness through financial security and relative status seeking, and much less so through physical comforts.
Financial security is great and leads people to feel comfortable making babies which enhances community, and on the low end maybe leads to less crime and despair.
Relative status seeking is zero sum and a culture that promotes it will always have a class of people feeling horrible. Social media amplifies. Maybe promotes crime.
Physical comforts: I don't really see how this could make a whole lot of difference in happiness. Having an extra bedroom to store crap in you never use, or having power windows, a bigger TV?
Outside of this, the pleasantness of your lived environment has to factor in somewhere. People pay A LOT, both in money and in physical comfort to live in cute walkable old towns built when you could still build those. Or beach towns or lakefront towns.
It's like the story Sam Spade tells in "The Maltese Falcon" about a missing person case he worked; the guy just disappeared one day, left his wife and family and old life behind, and couldn't be found, until five years later a chance sighting led to his discovery:
"Here's what had happened to him. Going to lunch he passed an office-building that was being put up--just the skeleton. A beam or something fell eight or ten stories down and smacked the sidewalk alongside him. It brushed pretty close to him, but didn't touch him, though a piece of the sidewalk was chipped off and flew up and hit his cheek. It only took a piece of skin off, but he still had the scar when I saw him. He rubbed it with his finger--well, affectionately--when he told me about it. He was scared stiff of course, he said, but he was more shocked than really frightened. He felt like somebody had taken the lid off life and let him look at the works."
Flitcraft had been a good citizen and a good husband and father, not by any outer compulsion, but simply because he was a man who was most comfortable in step with his surroundings. He had been raised that way. The people he knew were like that. The life he knew was a clean orderly sane responsible affair. Now a falling beam had shown him that life was fundamentally none of these things. He, the good citizen-husband-father, could be wiped out between office and restaurant by the accident of a falling beam. He knew then that men died at haphazard like that, and lived only while blind chance spared them.
It was not, primarily, the injustice of it that disturbed him: he accepted that after the first shock. What disturbed him was the discovery that in sensibly ordering his affairs he had got out of step, and not into step, with life. He said he knew before he had gone twenty feet from the fallen beam that he would never know peace again until he had adjusted himself to this new glimpse of life. By the time he had eaten his luncheon he had found his means of adjustment. Life could be ended for him at random by a falling beam: he would change his life at random by simply going away. He loved his family, he said, as much as he supposed was usual, but he knew he was leaving them adequately provided for, and his love for them was not of the sort that would make absence painful.
"He went to Seattle that afternoon," Spade said, "and from there by boat to San Francisco. For a couple of years he wandered around and then drifted back to the Northwest, and settled in Spokane and got married. His second wife didn't look like the first, but they were more alike than they were different. You know, the kind of women that play fair games of golf and bridge and like new salad-recipes. He wasn't sorry for what he had done. It seemed reasonable enough to him. I don't think he even knew he had settled back naturally into the same groove he had jumped out of in Tacoma. But that's the part of it I always liked. He adjusted himself to beams falling, and then no more of them fell, and he adjusted himself to them not falling."
In the past, child mortality was high. People were sad when their children died, but it was a fact of life: they adjusted to the beams falling. Then we got medicine and progress and all the rest of it, and child mortality decreased by a great deal. Now people might be happier about the chance their child would survive and not die. But then we got used to "your kids won't die" and that as a source of happiness faded away: we adjusted to the beams not falling.
Same with everything else in modern life. When it's normal that you have running water and indoor plumbing and heat and food and all the rest of it, that ceases to be a source of especial happiness ("we got a telephone!") but rather becomes part of what is expected and normal. Now we become unhappy about new things ("my phone isn't the latest iPhone!")
I believe that almost all of happiness is relative, not absolute. Which would explain all of these findings.
It's nearly impossible to put ourselves in the shoes of people with different expectations for quality of life than us. In response to your thought experiment: I can *almost* imagine someone 1000 years ago saying, "I was utterly devastated when my baby died. But... only 3 of my 8 children have died. My neighbor had it far worse than me. And we rarely go hungry for more than a day or two. So overall we're doing OK." It's really hard to imagine feeling this way... but the only other alternative is to believe that people were depressed beyond imagination back then, which I'm pretty sure is not true.
It is literally incomprehensible to us that improved health (and other life improvements) don't durably increase happiness in a meaningful way. And yet I'm pretty sure it's true. Which kind of implies that progress is not a very useful goal to try to achieve.
Of course, from an evolutionary POV, this is very easy to believe. Imagine that our quality of life gets 20% better over some period of time. Person A experiences happiness in absolute terms. They become durably happier due to this progress. They become more content, and so they strive less to improve their life. They take it easy.
Person B experiences happiness in relative terms. They see that they are still no better off than their neighbors. They keep striving.
B people will outcompete A people. Evolution ensures that happiness is mostly relative, but it *feels* completely absolute.
Progress causing durable happiness is evolutionarily maladaptive, unfortunately. (But *thinking* that progress will cause durable happiness is extremely adaptive!)
I totally agree. My great-grandparents had something like 8-10 children. One of them was named Maria, and she died in the first year or so. Well, then they named the one after that Maria again. I haven't met them personally, but from what I was told, they considered the second Maria as a replacement for the first one, and were satisfied with that. I don't think that "a child died" was such a terrible tragedy for them. Children were to some extent fungible back then.
And yet it's worth being careful in how we talk about this, because this kind of thinking is so alien to most people. (Myself almost included.)
Telling someone "I believe that children dying wasn't as big of a deal 100 years ago, they were somewhat fungible" will disgust a lot of people, regardless of whether it's true or not.
It is pretty clear that there IS a kind of happiness treadmill. See, for example, the book "Never Enough" by neuroscientist and former addict Judith Grisel. The book explains that we are happy when things seem to be going better now as compared to some kind of average of the recent past. Hence the futility of seeking happiness via drugs. Yes, they can create happiness now, but they also "raise the bar" for future happiness.
Section 7 of your Nonfiction Writing Advice -- https://slatestarcodex.com/2016/02/20/writing-advice/ -- from early 2016, before Trump's election, is fascinating to reread in 2025. I was just chatting with someone about how the Trump example there holds up. The 2016 claim was that Trump was skillfully, via tribal signaling, persuading conservatives about liberal ideas. Do you think, in retrospect, that's what was going on?
Assuming Prospera survived at least 20 years, expanded to 1 million people, and worked at least half as well as advertised, how much would you update toward Anarcho-capitalism?
Not much. Prospera isn't anarcho-. It's just normal, having-a-government capitalism. We already have ample evidence of how well this works and I don't think one more example in an admittedly-weird jurisdiction would update me much.
I can't remember what premodern people I was talking about, but a lot of them seem to wake up at the break of dawn, do twelve hours of hard labor a day, and never (noticeably) have chronic fatigue or get burnout.
I agree that a plausible solution is "if they don't do this, they'll die, and that's a great motivator!", but I do think it's helpful to notice the puzzle.
I just visited Poverty Point in Louisiana and had a similar reaction. "This mound was built by a few hundred to about 1000 people in 3 months, carrying a total of about 15 million 50 pound baskets of dirt." Today they have tram rides to take people who can't or don't feel like walking on tours around the ~2.5 mile loop to see the site.
I think you had said you did something along the lines of what Experimental Fat Loss had done to lose weight prior to your wedding. Can you share any details about that?
I lost about 30 lbs in a year on a relatively simple diet of eating the same few healthy foods every day. I was going to write about discovering the perfect diet, but I figured I should wait to see if I gained it all back first.
I went about a year without gaining it back. Then I had kids, got more stressed, overate more, and - yup - gained it all back.
Then the kids started sleeping through the night, my life stabilized again, and I tried the exact same healthy diet as before - and it had zero effect, barely even lost 1 lb in six months of trying.
Three days ago I gave up and started Zepbound (an Ozempic relative). We'll see how it goes!
The first time I tried the keto diet, it was super effective, I lost weight rapidly, my health was fine. However, I really hate the keto diet, stopped, and gained the weight back.
When I tried the keto diet a second time, I experienced almost no response.
I wonder how common that is. I haven't tried one of the Ozempic drugs, but it's on my to-do list.
Do you think the Rationalist community could be doing more for humanity than just EA? For example, maybe if we did some copy editing on HPMOR, lobbied, and advertised, we could grow more appreciation for epistemics in public schooling.
I've found that people either find rationalist stuff instantly exciting (1%), or bounce off of it bu either finding it boring or having dumb objections that they won't drop (99%). My preferred strategy is to make sure we have enough reach to catch the 1%, then gradually bring the 99% over by exposing them to things that sort of assume it and obviously work (like prediction markets).
I know what you mean. But maybe we just need to target the right people rather than the public. Like imagine getting Niel Degrass Tyson, Bill Nye, or the right bullion gushing publicly about it and trying to get public schooling onboard.
To put it another way, a portion of rationalism won through popularity of science. Whatever the mechanism science succeeded by, with effort we should be able to get the rest of rationalism popular.
If it's through practical application and "winning", we need to find a way to amplify that into something visible.
No doubt from a politics/tribes perspective we are at a disadvantage. I guess I'm imagining spinning it whichever way is convenient (if only THEY would think RATIONALLY like us, we should make rational thinking required!) I also was hopeful people in charge actually could be convinced. Someone out there decided to reach evolution and not creationism...
Thank you so much for keeping this wonderful blog! Really got me through some rough times in my recent struggles with pregnancy and being a new mom. I've been really enjoying re-reading Unsong and some of the posts one the old blog, too.
To what extent did you and your wife follow the obscure pregnancy interventions you mentioned in your post on obscure pregnancy interventions? (I'm a bit curious about one in particular: the post says it's probably bad to use all painkillers including aspirin. I ended up taking baby aspirin every day for the second half of my pregnancy on recommendation by my obgyn (because of advanced maternal age, I think?). My understanding is that it was a trade-off between risk of pre-eclampsia and risk of causing cardiovascular problems for the baby. I was wondering if you also had to make any difficult decisions like that?)
Also you mentioned in one of your posts that your babies had trouble with direct breastfeeding at first. I was wondering to what extent it got better? (Been on more than two months of triple feeding (breastfeeding, then bottle feeding, then pumping every cycle), since that's what the lactation consultants and pediatrician suggested; it's been tough having no end in sight.)
Yeah, we got told the same thing - 325 mg is bad, but 81 mg is net positive because of preventing pre-eclampsia (which we were at extra risk of because of twins). I think my wife still didn't end up taking it because she got too many digestive side effects.
The breastfeeding situation didn't get better. My wife slightly blames me and my parents, because she said to wake her up at night whenever the baby cried, but we felt so bad for her being so exhausted that we bottle-fed the baby instead the first few days without waking her up, and then her milk production kind of died down. But we realistically weren't going to be able to solely breast-feed both twins anyway, so we gave up pretty easily and switched to bottles. The babies liked the bottles and bonded with us in other ways and there was no problem.
I'm working on a crowdfunding platform (link below) to enable systematic chemical testing of consumer products (initially focused on plasticizers and phthalates, inspired by PlasticList).
Given your interest in aggregating distributed knowledge and your writing on FDA approval processes, I'm curious:
Do you see potential in using crowdfunding models to generate publicly-available safety data? Or would such systems inevitably select for sensationalist concerns over legitimate risks?
I don't really know anything about it yet and haven't tried it on anyone. My very weak prior is that everyone is always coming out with new medications branded as "Exactly as good as this addictive medication you're reluctant to use! Switch all your addictive-medication patients to it!" and then it's never really as good.
I remember you stating at one point that you have misophonia. What do you do, if anything, to manage it? (I imagine the standard headphones + music but what else?)
We’re going to the bay area soon for a visit and I was wondering if there are any rationality-related sightseeing activities available. Like, is there a plaque on the tree that dropped an apple on Eliezer Yudkowsky’s head?
Lighthaven is best, but it's not usually open. However, there's a public Sequences reading group Tuesdays at 6:30 PM, see https://www.lesswrong.com/groups/LK6GNnKp8PDCkqcxx for details. I am told that the rule is that Aella slaps anyone who hasn't done the reading beforehand (this is literally true, not a joke, although the exact penalty varies from week to week and that was just a particularly memorable one).
If you were having kids 5 years from now, and wanted to use a new reproductive technology that could offer significant benefits, how many healthy babies from the technology would you need to see before you were comfortable using it yourself?
(And if the answer is zero, how many healthy monkey babies would you want to see?)
Obviously depends a lot on specifics. We're shaping up to be in the first 100 or so to do polygenic selection on traits, but I don't care because there aren't many ways it could go wrong (even if they mess up everything, we just get a random embryo, which is what would have happened without them).
If it's some kind of highly complex gene editing, then I would want - not necessarily more babies, but waiting longer to see if there are things that only show up later. I vaguely remember hearing that the thing where Dolly aged prematurely didn't pan out, or doesn't happen anymore, but that's the kind of thing I'd be worried about. Or maybe not, you could try to convince me that this was an unlikely failure mode.
...at least, that's the responsible answer. If it was something really cool like superbabies, I might be tempted to be early, just to support mad science and be able to say we were one of the first.
The Trump election wasn't great. I don't necessarily think he'll be some kind of extreme perma-dictator or something, but I don't think his administration is likely to be very good at handling the Singularity and I think there's a decent chance it happens before 2029.
What kind of change in policy and/or in the underlying culture would lead you to update your currently negative outlook on Italy? Also, somewhat related, what is your opinion on the value of studying classical languages and literature, both in general and more specifically for someone who identifies as an intellectual?
How many ceteris would have to paribus to have made it worthwhile to have kids earlier, for you? Every Sad Singles' Day that goes by has me thinking of the rising startup costs of reaching the same milestone myself...what's the Pareto optimal point of age-enhanced wistom and resources, vs. the waning ardour of youth?
I got really lucky, in that I married someone younger than myself and got the advantages of a stable career etc without having to compromise on fertility.
I don't know if I could recommend this strategy to other men, because at some point your chances of getting a woman sufficiently younger than yourself for it to work go down, and it's not worth the risk. And I'm not in the right place to recommend/disrecommend it to women at all. But if it works, it works really well.
Would you consider trying to get a collection of your short stories/humorous published in a collection to try to make a splash on the literary scene and perhaps submit for award consideration? Maybe try to write a novel in the tradition of the "novel of ideas" in the future? You've written some lovely short stories in the past that show a good command of emotional modulation, humor, and character development in addition to wrestling with philosophical questions.
Not really. I think blogging is now big enough that the sort of splash I could expect from the literary scene wouldn't be a huge benefit. Exceptions if I got a Pulitzer or something but I think that's out of reach, both because it's not my style (I'm not that serious and sophisticated and in touch) and for the usual reasons that it's crazy hubris to think you could win a Pulitzer.
What do you think about the value of MDMA for everyday non-clinical therapy? I am especially thinking about couple therapy, where in my experience MDMA enables people to talk to the partner in a very connected way about difficult issues between them.
Partly this is out of normal outside view "don't bet your whole future on your idiosyncratic crazy beliefs being true" considerations, and partly it's out of the possibility that a Singularity which doesn't kill all humans will massively multiply capital and obsolete labor, so you'll want to have money saved up if it happens.
For a while you have read biographies of various dictators. I think you have looked for patterns in their strategies for dictatorship, so how they consolidated their power after they came into office.
Would you say that there is a big pattern (like, which checks and balances to dismantle, how much to build up their popularity, how much to control the media etc), or are the persons and countries too different for comparisons?
- Usually this doesn't happen to healthy countries. On a cosmic level, nobody "deserves" dictatorship, but on a mundane level, the countries it happens to have really been screwing their populations over for a long time.
- Usually comes on the back of right-wing anti-elite sentiment, and the elites usually have been extremely out of touch and terrible leaders.
- Often the future dictator builds credibility by being a genuinely good leader beforehand, or at least lucking into a good track record. Often they are more honest than their opponents, or at least differently corrupt, or at least good at seeming this way.
- Usually less about "literally kill every opponent" and more about someone who could naturally get 55% support in an election tweaking things enough to get 65% or 75% support; but they still sometimes lose elections if they really screw up!
In some of the media I follow, a lot of hoopla is being made of late about near-death and other mystical-type experiences challenging materialist scientific presumptions about how things work (eg. Sebastian Junger, Ross Douthat). It seems like such experiences can only be one of 3 things: lies, hallucinations, or veridical in some sense. What probability do you place on the third explanation (veridicality) being correct, ever, and can you please explain why?
Pretty low. I remember reading about a cool study where some operation tended to produce near death experiences. The patients would say they were floating above their body or something. One of the surgeons put some kind of really bizarre thing on top of a bookshelf, such that a normal person couldn't see it, but if you were floating on top of the room you definitely would. They asked the patients who said they saw the whole thing from above what was on top of the bookshelf. Nobody could answer.
This study sounded really interesting so I decided to try and hunt it down. I think this comment is conflating two (similar in concept and similarly-named) studies by the same researcher, Sam Parnia.
The first study, AWARE (https://doi.org/10.1016/j.resuscitation.2014.09.004), used the experimental design described here. The operation in question is CPR following cardiac arrest. Basically what the researchers did is they installed the aforementioned shelves in roughly 1000 hospital rooms in critical care wards where they expected cardiac arrests to take place. The items placed on the shelves were images of "a combination of nationalistic and religious symbols, people, animals, and major newspaper headlines". Unfortunately, most cardiac arrests in the participating hospitals didn't actually happen in these rooms, and of the 2 subjects who survived cardiac arrest and reported out-of-body experiences, neither were in a room with an installed shelf.
The second study, AWARE II (https://doi.org/10.1016/j.resuscitation.2023.109903), used a somewhat different design, wherein researchers wore pagers that alerted them to cardiac arrests, upon which they would bring a tablet into the room showing an image randomly selected from a set of 10 images, and they would also put bluetooth headphones on the patient playing audio that consisted of the words "apple", "pear", and "banana". Among 28 survivors that this was done on, 1 was able to recall the correct fruits, and none were able to identify the image on the tablet.
I was honestly just trying to find out what the object on the shelf was and didn't expect to go down a rabbit hole.
Thanks for the papers! The second study is particularly interesting. To expand a bit:
They had over 500 subjects, but only 53 of them survived and due to health concerns they were only able to interview 28 of them about their experiences. Of those 28 only 11 had experiences. Of those 11, 6 had the classic "out of body experience" thing. Though the specifics varied, all 6 reported 5 similar stages of the experience: Being separated from their body, heading towards a destination, undergoing a meaningful re-evaluation of their life, arriving at a good place (heaven like), and then returning to their bodies again. Notably all 6 reported feeling lucid during the experience, aware of their situation and identity while being able to reason and remember.
As far as the tablet images go, none of the 6 reported noticing the tablet. None of them reported seeing a different or wrong image on the tablet then what was displayed, and none of them said there wasn't a tablet there. So it's not exactly a smoking gun refutation: presumably if the experience was veridical it wouldn't' be unreasonable for someone not to notice a tablet computer on a bookshelf while they're watching themselves on the table and then traveling to heaven or wherever. One of the subjects reported watching the doctors put electrodes on his chest and shock him, and the researchers were able to confirm that he was indeed shocked with electrodes. None of the other subjects provided any information that could be checked like that.
Some fun quotes from the 6 who had the classic experience:
"I was no longer in my body. I floated without weight or physicality. I was above my body and directly below the ceiling of the intensive therapy room. I observed the scene that was taking place below me ... I, who no longer was the body that had belonged to me just a moment prior, found myself in a position which was … more elevated. It was a place that had nothing to do with any kind of … material experience.”
“I remember entering a … tunnel. The feelings I experienced … were much more intense than [usual]. The first feeling was a feeling of intense peace. It was so calm and serene with an incredible amount of tranquility. All of my … worries, thoughts, fears, and opinions were gone. The intensity of the tranquility was so incredible and overwhelming that there was no fear in what I was experiencing. I had no fear about where I was going and what to expect when I arrived there. Then I felt warmth … Then there was the desire to be home.”
“I do remember a being of light … standing near me. It was looming over me like a great tower of strength, yet radiating only warmth and love … I caught glimpses of my life and felt pride, love, joy, and sadness, all pouring into me. Each images was of me, but from the standpoint of a being standing with me or looking on… I was shown the consequences of my life, thousands of people that I'd interacted with and felt what they felt about me, saw their life and how I had impacted them. Next I saw the consequences of my life and the influence of my actions.”
"“I went directly to a place of light. It was calm and immediate … The place where I was I perceived to be analogous in a way to the exterior of an entry way…There was one major being of love and many other beings of love … There was nothing but love, goodness, truth, and all things to do with love. There was no room for fear or evil or anything but this love. It was more wonderful than any of my best hopes or experiences [in this place]. It was beyond perfect and loving, as we in our human state know it. There are no words to describe it. I was so happy to be there.”
“I was asked if I wanted to come home (meaning there) or wanted to come back here. I told them that my two sons needed me and I had to go back. I was suddenly in my body again feeling my achy joints flaring in pain. I really don't remember what was going on around me at that point, just that I hurt.”
As a fellow healthcare provider, I’ve been seeing a lot of folks lately who are very distressed about the state of the world — more than usual. What approach do you use to address these fears? It’s a tricky subject and it has a lot of potential to derail the appointment, but some people seem almost paralyzed with worry.
I think first step is to figure out whether this is a sign of a deeper psych problem or just nonpathologic (albeit often unreasonable) worry. Normal psych screening can help with this.
If it's nonpathologic, then - while validating their concerns enough to let them know I'm not objecting merely out of love for the other political party - I try to figure out the pattern of what's bothering them and what points of leverage we have. Often it's something like the news triggering them, in which case I gently suggest they watch the news less.
If they're skeptical of this solution, I try to figure out whether it's because they feel ethically bound to know about bad things, or because they *want* to watch the news less but don't have the willpower to comply.
If they feel ethically bound to know about bad things, I very weakly try to push back ("Is this really helping the victims of these bad things? Wouldn't you know just as much if you only checked the news once per week? Maybe you could be slightly less informed but help in some other way like volunteering?") but eventually give up because I think this is slightly out of the proper realm of a doctor and now I'm just arguing with their lifestyle and ethics.
If they want to stop but don't know how, I talk about some willpower solutions, like giving their social media password to someone else, or getting rid of the TV, or making a specific schedule ("check the news once a week, Tuesday at 8 PM, and otherwise avoid it, and make sure all your friends know and are holding you to it"). If their problem is getting into doom spiral conversations with friends ("Did you hear about outrage X?" "Oh, you think X is bad? You are like a little baby, outrage Y is so much worse!") then I ask them if they endorse these and whether they could talk to their friends about doing this less.
If none of that works and the patient is a total wreck, I'm not above occasionally suggesting SSRIs to prevent them from being miserable all the time even if it doesn't quite meet depression/anxiety criteria. I would consider myself overprescribing if I did this for more than a small fraction of patients who come in with Politics NOS.
What writer most unexpectedly influenced you? It doesn't have to be one with an absolutely large magnitude of impact, just large relative to what you expected when you heard about / read them initially.
Two questions; feel free to answer only one and ignore the other.
One: have you seen the poem “Scooping the Loop Snooper”? It’s a proof of the impossibility of solving the halting problem in the style of Dr. Seuss. I’ve been meaning to mention it every since reading your review of the other Scott A’s quantum computing book. (If you haven’t seen it, it’s free on the author’s website; short and sweet.)
Two: have you considered making some of your old live journal posts public again? It’s really annoying to use archive.org and you’re no longer “in hiding”, right?
I like the MLK day poem and the Swerve book review and I think there were many other goodies but it’s annoying to browse now
What probability would you assign to the US getting dysfunctional and/or dictatorial enough in the next four years that you and your family will attempt to emigrate, and why?
5%. I (white, cis, etc) don't expect to be a target of Trump's, and if I'm not personally a target, then my house/family/career all being here makes me unlikely to move just because he's being horrible to someone else. I'd rather stay here, make enough money to donate to good causes, and vote against him (assuming voting still exists). I'm not sure I'd leave even if he suspended democracy and elections - obviously this would be bad, but me being somewhere else doesn't help other Americans and is probably worse for me.
I think there would either have to be such widespread economic collapse that it was worth me switching to a career other than blogging (since blogging makes me the same money regardless of where I am), some kind of civil war that makes me feel threatened by violence, or some kind of unprecedented attack on extremely basic freedoms that I don't think Trump's base would go for.
FWIW, as a parent of an older child, and someone demographically similar and thus similarly not-target-shaped, here are the scenarios I worry about that add up to somewhat more than 5% for me:
1. civil war, as you mention
2. economic collapse caused by war with China
3. pandemic where other countries get a vaccine and we don't because RFK
Regarding medicine and AI: do you think that as AI continues improving its diagnostic and treatment-recommendation skills, that psychiatrists will end up focusing more on psychotherapy again because AI will be able to do much of the rest of our jobs about as good as we can? It seems to me that the physical presence of another human is something both centrally important to psychotherapy and is something AI can't replicate. And while AI might be able to do some forms of psychotherapy kind of OK, ultimately if you're doing serious therapy with an AI bot, it's just not the real thing (and even a human-looking robot therapist would either have to disclose it isn't a human or else lie to the patient).
I think insurance reasonably wants to avoid paying psychiatrist prices for therapy, and that if mental health refocuses around human therapy then therapists with PsyDs or LMFTs or whatever will provide it.
But also, I don't see regulation as favoring AI, both for normal regulatory capture reasons and *especially* in psychiatry where we're so often prescribing controlled substances. Even if AI is able to do our job perfectly well (as it's pretty close to already) I think there will still be a psychiatrist with an MD typing the things into the AI and making the final decision.
It seems like recently the right-wing ratsphere has had a bigger impact and influence with their political sphere than the left-wing ratsphere has with theirs. Would you agree with that assessment and if so any insights into why you think that's happened?
I don't think there really is a right-wing ratsphere. If you mean like Richard Hanania, he is an interesting person with lots to say but I don't think of him as part of the rationalist movement.
I think the reason right-wing intellectuals are generally more influential than left-wing intellectuals is that there's many fewer of them (90%+ of college professors are left wing, etc) so there's much less competition.
"I think the reason right-wing intellectuals are generally more influential than left-wing intellectuals is that there's many fewer of them (90%+ of college professors are left wing, etc) so there's much less competition."
Even the sum total influence of right-wing intellectuals seems bigger to me than the sum total of left-wing intellectuals. You might even expect it should be the other way around if there are more left-wing intellectuals.
I think while the left wing 'intellectuals' are the majority in theory, the majority of them who engage with politics are not acting as intellectuals but as plain ideologues.
I'd guess that one reason more right wing intellectuals actually engage with politics is that they are going against the consensus. Left wing intellectuals don't have the same incentives.
How do you have conversations with everyday people in your life about likely rapid advances in AI, associated risks, and how they may affect their lives? I have yet to stumble on key concepts and intuition pumps that keep people open minded and guide the conversation. My peer group is mostly curious, recent liberal arts grads who mostly get AI content in the form of "LLMs are stealing", "AI art bad", "AI is just hype" and see most discussion through the lens of those arguments. When I break through to substantive discussion, we often get hung up on discussions I can win, but can't persuade. For all the "we've seen the scaling laws hold over 6 orders of magnitude, I think *you* have the burden to show they'll fail" or "No, I can't tell you how ASI would manifest, maybe through persuasion, maybe conducting business over the internet, but just as one can think Stockfish will win at chess without knowing how, we can assume AI will outmaneuver us" I might muster, it doesn't do much to convince a skeptic.
I have met a lot of people with those reactions and mostly stop trying to discuss with them. I figure if I'm on record saying things they don't believe, but don't press too hard, then eventually they'll see it for themselves and might remember that I tried to tell them. But more and more, I am starting to be able to just point to real things current AI can do. The gap between what I expect to actually happen, and what I can say before being dismissed out of hand, is shrinking. This is in some ways welcome, in some ways disturbing.
I recently got put in charge of a project at work figuring out what AI tools we should be using, because my boss knew I had an interest in the field and was trying to track advances anyway. "Here's a 50 page Deep Research report I got after an hour chatting with o1-pro and o3-mini-high to gather context, without doing any other optimization or putting in any of my own knowledge. This needs to be the minimum baseline for what we assume our consulting clients already know before they bother talking to us," is a pretty compelling argument.
A few coworkers have even started to take my views seriously on what AI could mean in 5-15 years, to the point of actually comment on how disturbing it was that those things seemed so plausible. One compared what I said to Black Mirror, and seemed to get it when I replied that Black Mirror went as far as it could without seeming too crazy for people to want to watch it, and the reality will be several times stranger, whether in good ways or bad.
Over the years I have also developed a bit of a reputation in my own circles of saying weird things that turn out to be (sometimes usefully) true, which helps. But yeah, this past six months or so has been the first time I feel comfortable just coming out and saying things like, "I think the last generation to die of old age has already been born... one way or the other," or "The last time a new and smarter kind of mind appeared on the planet was the replacement of Homo erectus with Homo sapiens, and that didn't go so well for Homo erectus, or Homo habilis before that."
I mostly don't. My parents and in-laws know it's an interest of mine and know I have some crazy beliefs about it, but we rarely go into detail. Most of my friends I met through the rationalist community and they already know.
How do you rate Irish vs US medical education? My understanding is you studied in Ireland, which I assume is similar to the UK (where I studied), and my impression has always been that US medical university is much more rigorous than in the UK (note - I’m not under the impression that it produces better doctors - telling apart the various types of pulmonary hypertension is often note useful imho)
I never got an apples-to-apples comparison, but my impression is that Ireland was lower-tech, which means they rely on clinical exams and actually knowing what you're doing, whereas US was more likely to be "here's how to call the radiologist and ask what the CT said".
I love your old story A Modern Myth! I feel like it should be made into a movie. That's not a question. My real question is how are Kai and Lyra doing?
Maybe I should’ve phrased the question in reference to ai smarter than effectively everyone but not really super intelligent. It seems really short run imaginable for this to be the case for 90% of Americans.
I tried suggesting it to my patients and they almost never got any benefit. It seems like you need some THC to "activate" the CBD, or something. CBD treatments with mild amounts of THC seem to work a little, but then they're not the safe healthy no-addiction-risk miracle drug we wanted.
My favorite SSC blog post of all time is "The Parable of The Talents". I find myself re-reading or thinking about it every few months. Question for Scott: what's an SSC or ACX blog post that you find yourself re-reading or thinking about the most?
- Yes, babies sleep less peacefully and happily on their backs, and there are various surrogate endpoints that seem worse. Nobody has ever been able to demonstrate any long-term damage, but you could imagine something very subtle.
- Yes, the amount of SIDS you prevent by back-sleeping, while nonzero, is very low. This is one of those epidemiological paradoxes. Suppose back-sleeping lowers your chances of SIDS by 1/3,000 (I think this is OOM right). Then if you're a parent, that seems so low as to hardly matter. But if you're the head of the National Institute of Health, you might think something like "3 million US babies per year, so recommending back-sleeping would save 1,000 lives per year! Think of how much tragedy is prevented!" I think these are both reasonable ways of thinking about things.
In Unsong, would the final interlude have been different if Clinton had won the 2016 election? (And if so, do you have a draft of Interlude: Clinton somewhere, or did you write the whole thing in real time?)
Invest in your relationship with the mother, and if your parents live close, with them as well. The first 6 months after birth will be a black hole, but after that normality gradually returns. The stuff about 'having no hobbies anymore' and 'never getting any time for your personal projects again' are all untrue, at least for parents who are together and have some grandparent backup.
A big part of the early exhaustion stems from stress that you don't know what you're doing, but really, really care about getting it right. Depending on your temperament reading a book on the subject might help, or might just make you second guess your instincts harder (you will eventually discover that your instincts were pretty good, and that if you could have just turned your stressometer down a few notches everything would have been better). I found grandparent advice much more useful. Asking advice from parents with kids who are older than yours and have grown up in the way you hope your kids will grow up is a good move as you had past the 6 months mark and move from survival to thrive. There is a huge amount of fad parenting nonesense and I would basically disregard advice from anyone whose kids you can't assess as a yardstick for advice quality.
Are there any areas, fields, scenes, institutions, movements, etc. in contemporary American life (being a little vague on purpose) about which you’re feeling mostly excited and/or optimistic?
For context: I’ve been feeling a bit bored and/or pessimistic about a lot of different areas of modern American life that usually excite, interest, and motivate me (the arts, writing, the behavioral and social sciences, academia, and politics being the most important). I’d love to find something outside of my personal life to feel excited or optimistic about, so I’d be very interested to hear about anything that’s making you feel excited or hopeful these days.
Do you think reading Rationalist materials/blogs, being versed in the Rationalist lingo etc, materially improves one's life? Do Rationalists make better decisions, live happier lives, make more compelling arguments, invest their money more successfully, find better mates than other folks with similar IQ/backgrounds?
Recently, Dwarkesh Patel made a fascinating point about current AI progress that I've never seen before, and I think holds water: "[T]hese things have basically the entire corpus of human knowledge memorized and they haven't been able to make a single new connection that has led to a discovery[.] Whereas if even a moderately intelligent person had this much stuff memorized, they would notice - Oh, this thing causes this symptom. This other thing also causes this symptom. There's a medical cure right here. Shouldn't we be expecting that kind of stuff?"
Do you think this is a meaningful criticism? Should we expect more discoveries? If not, why not?
Claude and I co-discovered a novel vision algorithm for a non-euclidean 2D shooter just a few weeks ago. The highest-level abstract idea was mine, but all the details of expressing it in geometrical terms were all the AI's. We needed some help from OpenAI's reasoning models to finish debugging it.
Here's a stream VOD of me working on it - unfortunately I didn't manage to capture the "aha moments" but it should still be illustrative - if I was gonna lie I'd lie about inventing it singlehanded, not lie about using AI. :) https://www.youtube.com/live/utnN6WPf5nQ
Does this serve as proof LLMs can discover things? I think so - if you ask the question correctly, and the solution requires innovation, LLMs are 100% capable of bringing their entire knowledge base to a problem and make new connections to solve it - as always, it's a problem of asking the question correctly so the right knowledge becomes salient.
You are a very culturally literate (for lack of a better term) writer who frequently makes informed references to all kinds of stuff, from ancient philosophy and classical literature to manga and edgy internet memes. (This is very cool!)
But one thing you seem to very rarely allude to is movies.* This seems very strange to me, as virtually everyone I meet IRL watches movies at least once in a while (even if they're very different movies), to the extent where I would consider the feature film as one of the basic cultural/artistic units of modern society (for lack of a better term).
Are you one of those rare people who says "I pretty much never watch movies"? Do you find that strange given your aforementioned broad cultural literacy? Or is that a false impression? Do you have a list of favorite films or directors?
*This is not a serious criticism by any means, but I remember a post from years ago in which you made a joke whose punchline was basically "Black Panther is going to get a hundred sequels" which fell flat for me because it immediately gave me the vibe of "this guy doesn't know what he's talking about" - a rare experience when reading an SSC post!
You're on your way to prevent multiple statistical deaths by donating 10% of your income in the form of a cashier's check, when you see a drowning child. If you attempt to save the drowning child, you will not be able to prevent multiple statistical deaths. What should you do?
What do you think of fame? Given the amount that you have, is it what you expected it to be? Do you ever wish you could be less famous and still have the same impact? I apologize if this is too personal a line of questioning, but I can't quite get my head around fame, not being famous myself, and I'm curious about any insights you think you might have. Thanks.
In the recent discussions about artistic taste, you didn't talk about an aspect that seems fairly central to me: boredom. I'm married to a choreographer and am friends/family with a bunch of musicians, and have noticed that the thing that seems to drive our taste is mostly "things I haven't seen or heard before". This is quite different from common taste, which seems to mostly be "things I have seen before and like, or things very similar to that". Note also the structure of pop music, which is designed to be essentially one hook that repeats over and over, vs modern classical compositions, which are often impenetrable. This makes me wonder: can we draw a distinction between financial taste (the tastes of funders), expert taste (the refined tastes of practitioners), and public taste (basic things the public likes)? The relative balance of power between these three camps might explain different trends in different artistic areas. I think architecture in particular has fairly powerful practitioners and financiers, and very weak public; whereas music has very powerful public and financiers, and fairly weak practitioners.
Question on who would be your ideal nominee for a certain job in a hypothetical Trade Offer.
The federal budget in 1969 was $186B, and the debt $353B.
In today's dollars it is $1.6T and $3T.
Today those figures are $6.8T and $35.5T.
I.e. 4.2x and 11.7x larger, even after being adjusted for an inflation that was itself almost completely produced by that budget overspending.
I picked 1969 because that's when USA landed a man on the moon while also conducting an all-out engagement in the Vietnam War. top-tier physics research, and many other things. It was even more "The Envy of the World" back then than it is now.
I understand that Elon is widely disliked. So the Hypothetical Trade Offer is that a wave of the wand and pouf - everyone agrees that Elon is a Nazi, or Unacceptably Autistic, or an Illegal Imigrant, and he gets deported to South Africa where someone can "dubul' ibhunu" him.
Granted. BUT you must nominate someone else - any person of your choice - who will slash the US government spending to less than one quarter over the next 1.5 years.
They may do so in any way they see fit. As much budget as they wish can to go to Saving Drowning Children 10.000 Miles Away. But it will be $1.6T and not a single cent more.
The Trade Offer is hypothetical. But the fact that you can run the envy of the world on less than a quarter of its current budget - without enslaving your children's children to pay off a beyond-ludicrous debt with interest - is not hypothetical in the slighest.
Why should the values be fixed total instead of per capita (then you'd need to get spending to 40% of current levels instead of 24%)? And why fixed in real dollars instead of %GDP? If we capped government spending to 1969 levels on a %GDP basis you'd only need to reduce the budget by about 20%, although the debt is still 3x more on that scale because of past decisions.
I would also ask: precisely what level and type of government spending maximizes economic growth while effectively compensating those who get hit by the negative externalities of said growth, and how do we get closer to that world? This is not a figure I expect to stay constant over time.
The US only had 60% of its current population in 1969, but you want the same amount of tax revenue to serve a far greater number of people. That doesn't seem fair.
Healthcare spending alone is more than your $1.6T target, and so would have to be massively cut. Life expectancy increased by about 9 years since 1969 despite rising obesity rates. The improvement is partially from better medical technology for treating cancer and heart disease, but those treatments are expensive and we would lose much of the benefit of these new technologies if we didn't actually pay for them.
Life expectancy had been steadily rising in dozens upon dozens of countries that didn't experience anything *near* the stellar explosion of US healthcare costs.
It was caused worldwide by general technological advances which go way beyond cancer and heart disease treatment. In 1969 diagnostic ultrasound was in its infancy; today you get amazing precision and insight out of a ~$50k machine. It's beyond compare. The list of technologies goes on, and they've all grown significantly *cheaper* and more powerful, not more expensive.
The US healthcare cost explosion has nothing to do with the cost of technology nor the cost of actually producing the medications. (cancer and heart disease included)
It has to do with the government keeping on throwing absolutely any amount of taxpayer money on the industry that the industry wants to ask for, while at the same time imposing insane byzantine levels of regulation which ensure that no true competition can exist among the few who have been selected for successfully passing that regulation, not selected for the healthcare they provide. Rinse and repeat for decades, and what do you expect?
"Serve a far greater number of people"? You probably mean welfare, right? Because right now it has to serve ~4.2 million miles of roads, which was ~3.5 million miles in 1969. Not quite a 4x difference. Meanwhile, the people who are on welfare now and wouldn't have been in 1969, and the money they draw on average compared to then, those are quite overdue for a visit by The Chainsaw Tzar.
I largely agree on the healthcare costs, but in addition the the factors you mention, we also just get more healthcare on average than we did in 1969, and this does improve health outcomes. We can be less efficient in healthcare but still have an absolute improvement.
> "Serve a far greater number of people"? You probably mean welfare, right?
I mean almost everything, including healthcare and welfare. Even the roads, in addition to the extra miles, they also get more traffic, which increases maintenance costs. You wouldn't expect a country with 1 million people to need the same expenditures as a country with 10 million people. Some costs scale sublinearly, others superlinearly, but overall, you'd expect costs to be roughly proportional to the population.
> But the fact that you can run the envy of the world on less than a quarter of its current budget - without enslaving your children's children to pay off a beyond-ludicrous debt with interest - is not hypothetical in the slighest.
Given that the children of the children of 1969 America generally consider themselves to be enslaved in the name of paying off beyond-ludicrous debt with interest, and that a not-insignificant portion of the $35T you cite *is* interest, I think you have accidentally provided a counterexample to your own claim.
Interest only grows exponentially. Whatever it is exactly for the children of the children of the 1969 America, it will be exponentially much more for the children of their children.
Financiers don't care a whiff WHAT the debt is for. "Trillions for reparations for slavery"? Yea, whatever dude, good one! Doesn't matter, just keep it rolling! As long as we can even further astronomically inflate the interest payments we're raking in, by even further enslaving the populace today, it's still less than how we'll have enslaved their children's children.
> Given that the children of the children of 1969 America generally consider themselves to be enslaved in the name of paying off beyond-ludicrous debt with interest
I’m not American but, really? The people born in 1969 are doing pretty good. Post 1990 not so much.
That's what I mean. 1969 America was the Baby Boomers, who are doing great. Their children are Gen.X, whose children are Gen.Z, who are not doing so well.
There’s some issues with how GDP is calculated but it’s the income of the country (not really wealth as is sometimes said). Therefore it’s useful to compare it to the debt of the country.
Can you try to explain/analyze, step by step, what happens in your head when you're writing jokes and coming up with creative ideas? Your posts have a lot of jokes, and some posts (like Bay Area House Parry) are made out of a series of brilliant/creative/absurd ideas, so clearly you can come up with stuff like that consistently and intentionally (as opposed to ideas "just coming to you", outside of your control). How do you do this? What happens in your head when you invent one of Bay Area House Parry startup ideas? I can't think of ideas like that, is there anything I can do to learn?
Non-authoritative answer in case Scott doesn't get around to it: I believe it is known that there are several .txt files on SA's desktop that slowly fill up over months or years with subject-related quips and curiosities (monthly links posts are a microcosm of this), so in a sense the ideas *do* just come to him out of his control.
It is similar to the common advice that if you want to be a writer, you should carry a pocket notepad and pen everywhere you go, which similarly helps random shower thoughts appear intentional when expanded upon and published.
Also, lots of practice in rhetorical role-play as a youth.
Is the methodology that social scientists have used to observe that the relationship between IQ and success in other areas of one’s life is causal rather than a correlation with other lucky factors like wealth and race robust? If so why?
I haven't looked into this super deeply, but I think so, for a few reasons:
1. It's definitely not just race, because people have looked into this within races and found the same effect.
2. Wealth is harder to disentangle. You can control for parental wealth, and when you do this I think the effect continues, but controlling for things is hard and you'll always have some residual error. I think the size of the effect is large enough that it's implausible that it's just residual error, but this is an opinion rather than a proof.
3. You can use twin studies and adoption studies to prove that the relationship between *genes* and income is at least partly causal, and probably some of those genes are for IQ, but this is just a guess - maybe it's genes for attractiveness or something.
4. You can probably do fancy statistics to show that the genetic effect is a pathway through IQ, but I don't remember any specific paper that has done this and fancy statistics are always suspect.
5. Actually, wait, the obvious right way to do this would be sibling pairs, where they both come from the same family but one of them lucked into a higher IQ. The only paper I can find on this is by Charles Murray ( https://sci-hub.st/https://www.jstor.org/stable/3083428 ) and he unsurprisingly finds it's causal.
I'm interested in the habits of successful writers. Do you have a time of day when you prefer to write? Do you write better when your schedule is empty or relatively full? Do you wait for inspiration to strike or do you block out time for writing, on a schedule?
No need to answer all of these — just curious to hear about your experience.
I'm running out of people who fit my criteria: modern-day figures who transitioned a somewhat democratic regime to a more dictatorial one and have good books about them. If you think of more, let me know!
(no, I'm not doing Trump - at least not until the dust settles and we know what his record looks like)
I actually think you could write an interesting one on Franco if you wanted to, in that he took power via an attempted coup that turned into a civil war, but it was an attempted coup in a fairly democratic regime in really, really bad shape, and pretty clearly if the country had not been a giant mess he wouldn't have pulled the coup. But that would be stretching it a lot.
There'll presumably be good Bukele books in there near future, if there aren't already. It's not focused on a specific person, but "South Africa’s Brave New World: The Beloved Country Since The End Of Apartheid," by R.W. Johnson is similar.
Scott, some time ago, I remember being able to access your blog roll (as in: people you follow and recommend) on astralcodexten.com main page. It seems to be gone (there are "recommendations" but I vaguely remember the blog roll to have been much longer and not just substacks). Where can it be found now?
Nowhere. I took it down because most of the people involved had moved to Substack and I could replace them with normal Substack recommendations (which you can still find somewhere on the blog front page). The others seemed so well-known (eg Marginal Revolution, Less Wrong) that it wasn't worth it. Sorry!
One of the people who wanted to traditional-publish wanted massive edits from a formal editor, and I didn't have the time/energy to do that. Another dropped off the face of the earth. Eventually I realized I was delaying and would never do it. At the same time, ACX user Pycea offered to help me self-publish and I realized I either took his offer or put it off forever and never published anything.
There was a convoluted correspondence between Unsong and William Blake's mythology. Erica was supposed to correspond to one of the Blake characters - I originally thought it was Orc, but later realized it should have been Vala.
That particular change didn't affect much (though headcanon will probably forever be Erica, "Val" reference pointer was already reserved by Valencia the Red from WtC), but I did spend hundreds of pages in anticipation of how the "Miss America" joke would get semantically repurposed. Ah well. Good to know the background reason. This is why it's good for the original web version to stay up!
Do mental hospitals tend to have a rule against employing people formerly confined in one? (Maybe only for those formerly involuntarily confined, maybe only for patient contact positions?)
This came up when I was a patient in one long ago - people were chatting and someone mentioned maybe wanting to come back and work here, and we got to talking about whether that was a good idea and how "lunatics running the asylum" was a well-known adage, considered a situation to be avoided.
How often do you use AI in your practice? When do you think “psychiatrist” will be automated in practice, even if you still act as a gatekeeper for drugs?
I sometimes ask it what it would do in a case, to see if there's something I'm missing. Occasionally I use it as a glorified search engine, like "remind me which diseases have this collection of symptoms".
Not really! The last time I played was to join in some friends' heavily edited homebrew version of Dragon Heist just before COVID. We wanted to do something interesting, so all of our characters were a gang of nine year old orphans. Whenever we had to interact with other people, one person with a shapechanging spell would pretend to be an adult. It was pretty funny.
This coat acts as a mundane trenchcoat while equipped by a single person, except that it resizes to fit any wearer. While equipped, the wearer may designate an ally in line of sight. When they do, their own body and anything they have equipped other than the coat become ethereal to that ally (and their equipment), such that that ally alone may pass through their body for the purpose of entering the coat and equipping it themselves. Upon fully equipping the coat, an extradimensional space materializes around the head and upper body of the prior wearer and they are blinded for the duration of the effect (sound and other senses carry through in both directions), and a separate extradimensional space materializes around the body of the new equipper below the waist. The waists of the new and previous equippers are joined.
The effect can be repeated an unlimited number of times, with each new equipper causing their lower body and the previous equipper's upper body to be transported into their respective extradimensional spaces upon equipping the coat (their waists remain joined with all other wearers). Only the most recent equipper may designate the next ally (as they are the only one able to see the material plane and affect it with their upper body).
When each wearer beyond the first equips the coat, they may will it to resize to any length from the sum total of the first equipper's height to the waist and the length of the most recent equipper's upper body, to the sum total of the first equipper's waist height and all equippers' upper bodies (excluding the head). The head, chest, and arms of the latest equipper are elevated to the corresponding appropriate height. The intermediate space between chest and legs is hidden by the closed coat, but acts as an elongation of the last equipper's torso. This and other effects of the coat are physical, not illusory. See Evard's Guide to Cosmic Horror for the effects of peeking beneath the coat or inspecting its wearers with extrasensory perception.
If the coat is opened for any reason other than a designated ally equipping or unequipping it (a full round action per wearer, in reverse order), all wearers are simultaneously expelled from their extradimensional spaces, with exits projecting vertically in order of equipping (it remains equipped to the last wearer and returns to behaving as a mundane coat). Each former wearer may attempt to grapple either or both adjacent former wearers as a swift action before any external forces take effect. Any wearer of the coat may also activate this effect at will as a free action if the coat is not held closed.
The coat may be held closed manually or with a rope or other fastener, but any fasteners attached to the coat in any way detach and fall off the coat whenever it becomes equipped or unequipped by any wearer (they may be reattached as normal immediately after falling off). As a standard action, any wearer of the coat may attempt an Escape Artist check against the method(s) of closure (regardless of their ability to physically interact) to bypass them and open the coat. Any wearer(s) of the coat may also oppose any attempt to open it (by wearers or otherwise) as a swift action, using an appropriate combat maneuver. (Any of these or other actions taken by any but the last to equip the coat suffer appropriate penalties for their blindness.)
Other effects may cause the wearers to be similarly expelled, including:
- Entering, leaving, or taking any physical action in rough terrain.
- Entering, leaving, or taking any action within the threat zone of any ally, enemy, or other creature as small as one size category below the smaller of the first or last wearer, and as large as one size category above the larger wearer.
- Beginning a round within a swarm or crowd.
- Being hit by a physical attack or combat maneuver.
- Taking fall damage.
- Performing any physical action that requires a skill check with DC 15 or greater.
- Jumping any height or distance.
All such conditions may be negated if all wearers individually succeed at a DC 10 Acrobatics or Athletics check (each wearer chooses for themselves) once per round on the turn of the first equipper. All clasps apply their bypass DC minus 10 as a bonus to all checks, but wearers may not manually hold the coat closed. Any other ally in melee range may grant advantage to all wearers as a swift action, and any effect which could break concentration grants disadvantage. These checks automatically fail if any wearer rolls a 1, and critically fail if more than one 1 is rolled.
The coat has no armor benefit, but has hardness equivalent to +1 hardened leather. If the coat takes 5 or more points of slashing damage (after DR) in one hit, its wearers are expelled regardless of clasps.
The coat gives +20 to Deception checks, made by the last wearer, to be perceived as any person of appropriate age for their apparent height. In addition to the initial check, the latest wearer must make a check once per action taken or line of dialogue delivered in the presence of a target if table talk between players exceeds 5 seconds, with an added penalty of –5 per equipped character whose player engages in table talk (excepting pantomime or shushing other players) and –10 per 10 seconds of table talk (including pantomime), to a minimum of –40 in penalties (–20 total).
Any player shouting during table talk (including to shush) grants disadvantage for any and all subsequent Deception checks (for any purpose) against the same target by the last equipper, whether or not the coat is equipped. If the coat is opened, all characters formerly wearing it suffer disadvantage to all skill checks involving social actions against anyone observing the interaction.
My pet topic is personality research, like Big 5 or Myers-Briggs. I think the zeitgeist has gone too far towards "it's not scientific/useless" when the data suggests otherwise. What do you think?
The apples to apples regression comparison ("4 scores") of MBTI vs Big 5 in that post shows a small (.15 vs .18) difference that is not indicated to be statistically significant. As in, the advantage of Big 5 over mbti is that it is Big 5 not Big 4 and that it uses a continuous score -- coincidentally the exact same reasons no one outside of academia uses it for anything.
Would you ever let someone turn ACX into an essay collection-type book, as has been done with SSC?
I know you did a retrospective on intellectual progress you made in the 2010s, and you've also searched for previous work you've done toward things you only later explicated. Do you see a connection between your essays "extremism in thought experiment is no vice," "fear and trembling at EA Global," and "the tails coming apart as a metaphor for life?"
Thank you. The omnibus ebook of SSC was a massive help to me when I was in a yeshiva without regular Internet access, and I still slightly prefer a format that I can read on my Kindle.
Noticed that ACX has been relatively quiet about artificial intelligence lately - is this a deliberate step back, or has it just seemed to you that there hasn't been much to say about it? In the past few months I've had many moments of "I wonder what Scott Alexander thinks of all this", but the only time you've commented on the subject recently was the alignment faking thing (IIRC).
Or another way of asking the same question: What happened to Machine Alignment Monday?!?!
I don't think it's been quiet. I'm helping with a different project on artificial intelligence and will post about it when it's done.
I found that people were less likely to read things with the Machine Intelligence Monday "branding" compared to regular posts, so I turned the deeper essays into regular posts. I think I was also sort of hoping to have a shallow series on the latest month's developments in AI whatever they were, but now there are dozens of blogs that do exactly that much better than I could, and Zvi even does it for alignment in particular.
Forehead-smack. I'd forgotten about the Machine Intelligence Monday "branding" when I was offering you the AGI Friday "branding". Thanks again, btw, for the encouragement. I've got the Substack set up and planning the first real post for this Friday!
Surprisingly well with the ones who contacted me, but that's obviously a biased sample. I think local commenter Isha Yiras Hashem was disappointed in it.
Is it possible to submit a question for the survey? I’d like to ask a question seeing whether people experience visual snow. I think the few studies that have surveyed it have either undercounted it to a large degree or missed its prevalence in populations the ACX community is enriched for. This is at least anecdotally true in a survey of ~100 people I happen to have brought it up in conversation with at various times.
People with mild visual snow are likely to think it’s just the standard human experience (like I did). I think it’s one of those things like post-toilet wiping stance that acts kind of like a scissor statement but for human experience.
Woah, I also have mild visual snow and assumed it was universal! (Didn't even know there was a name for it until I saw your comment and googled.) My partner has it too. We recently had a conversation trying to compare how 'big' the typical grain size of our snow was.
I also have tinnitus, which the wiki page says might be related, but I've had visual snow all my life and tinnitus only since getting COVID in 2020 so idk
This is great to see (and as always a sign that the SSC archives contain many unread riches that I haven't gotten to yet)! However, the post and survey don't really answer my question. Would you consider further exploring the topic? The survey only prompts people for persistent visual changes as a result of hallucinogen use, without prompting specifically for visual snow or similar phenomena. This is important because people who have it are unlikely to recognize it as a nonstandard experience and because it's asking for changes rather than base experience.
I am curious about the additional visual hallucinations you bring up, though - I tend to suspect that there are a lot of common but not-commonly-known minor perceptual disorders (much as there are many other common minor disorders in things that are less interesting).
I'm most of the way through the physical version of Unsong, and I've re-read the original enough to notice edits as I go. What changes most stick in your head, and what went into them?
A few seem obvious (CHANGELOG appears to be gone, which is understandable even if it makes me very sad); a few seem reasonable but unnecessary (e.g. the joke about Christianity being saved by failure to pronounce a four-letter word got replaced by one about Ohio becoming un-speakable on account of its phonetic resemblance to the Mortal Name -- which is also funny, but I"m not sure why it was a *replacement*); and a few others just mystify me (name/gender changes, mostly; kabbalistic references I'm not catching?). And some fairly significant changes to "The Image of Eternal Death" and Uriel's peace summit disaster, which I didn't much care for but I think I see the rationale in each case.
(I keep my own fiction's edit history with accompanying reasoning in Git. You can tell.)
Also, bug report(s): I spotted at least one continuity error introduced by the edits. Aaron no longer makes telepathic contact with Ana before escaping Unsong HQ, but she still later reacts as if he did. Pretty sure that was a case of intentionally editing one scene and not editing the other to match.
- Replaced Malia Ngo with Asher Bentham character. Malia was a relic of a very different earlier version of the book and never really fit in; I think the chapter focusing on Asher's childhood makes him more of a part of the story.
- Changed Erica's name to Valerie; this *was* kabbalistic, she was originally supposed to correspond to William Blake's character "Orc" but I was misreading Blake and the relevant character is actually Vala.
- Made a few of the diplomatic negotiation chapters better (foundation of Untied States, Israel/Palestine) because I kind of phoned those in originally and thought there was room for improvement.
- Took out some things that were ephemeral or related to the news when I was writing the story (eg Changelog).
Can you help me understand, in an IQ 110 or lower way, why the idea of us being in a simulation has gotten so much traction?
I’ve read the arguments and they all seem to be based on an underlying assumption that there’s no way we could know because we’d be programmed that way, ergo relatively high probability. This seems like the logic my stoner friends used in college to “prove” any number of ill-conceived theories but it’s also possible I just don’t have the intellectual horsepower to understand, or am missing a critical argument element.
To expand, it's supposed to be a trilemma: simulations are cheap and so most observer moments are simulated, we die and never do simulations, so all observer moments are real, and simulations are really really expensive or impossible, in which case most observer moments are real.
Do you listen to pop (non classical) music? And if so, could you give a spicy take that you truly believe, but think most of your readers would consider blasphemous?
Not a question, just a thank you. You've helped educate me bit by bit over the years through your work, and that's helped me in small but meaningful ways countless times.
Your fiction work finally got me off my ass to try and write something for self-publishing, in the tabletop gaming realm. So, thanks for that, too. Keep making those fiction pieces!
Any thoughts on AI in developing countries? People often talk about AI increasing economic productivity by a lot in the developed world, but if this is the case, can't the same AI be deployed in developing countries, where labor productivity is much lower? How likely do you feel this is to happen?
I was just accepted to medical school and so will have to think about what field of medicine to enter soon. What made you choose psychiatry? Any predictions about the effects of AI on the physician job market in 2035, or should I just not worry about that at all?
Can you talk about your experience, preferably first hand but second hand is fine, with gen z in therapy?
Zoomers have both much less stigma against seeking help but also claim to have a skyrocketing number of mental health problems.
Bonus question: some surveys (of dubious quality) point towards therapist being a desirable job for the youngins. Should this pan out, what would you expect the quality of the marginal therapists to be? Would the increased availability of therapy be on net positive?
Do you have a bottom-up or top-down approach to your writing? i.e., Do you have a running list of blog post ideas that you then try to confirm through content you find online, or are you reading papers and links on your own, with your writing ideas emerging along the way?
As a fellow rationalist aspiring to hone their craft, what resources and mental models do you find yourself getting the most leverage from when wrangling with a new sociological topic or issue?
When it some to national politics or policy, I find myself mostly unable to form helpful viewpoints and would like some inroads into improving.
Around a year ago several of my friends (who are women) told me some variation of "my therapist tells me adhd in women in severely underdiagnosed" and then implied that they should get tested. Hearing it from 4 unrelated, high functioning ~30 year old women made me wonder if some new research dropped or if this was a meme floating about in psychiatry. Thoughts?
Or the ones you can’t not mention in a cluster of three or four?
What do you think about the idea of someone dumping a bunch of your posts into an LLM to train it to talk like you? What do you feel like it would be missing most if it only had your writing to go by? Should this be illegal without your explicit consent?
Would you ever do a socialist/non-socialist FAQ, like you've done for libertarianism and NRX? Otherwise, do you plan on doing one for any other ideologies or plan on engaging with leftist thought in a less structured way again like your reviews of Singer on Marx or Red Plenty? Basically, I'd like to know your overall view on socialism since it's adjacent to a lot of the topics this blog normally covers and has a lot of strong arguments imo, but is relatively under-discussed here, but I'd still expect you to have a nuanced take on it.
I'd also be strongly interested in an update to the Non-Libertarian FAQ. The last version posted* does open with a disclaimer that "it no longer completely reflects my current views", but it seems like a bit more than a change in emphasis must have occurred for Scott to now be actively identifying with the label.
* 2017 on SSC? I feel like there was a more recent one on ACX, but not finding it at the moment.
Do you feel constant or regular stress, feeling it in your stomach ?
With your public exposure, with all the internet scrolling, with these huge articles you publish, I am very impressed at the level of energy that you deploy seemingly so naturally.
The topic you deals with, your readers with their high expectations : how do you manage to not freeze when choosing each word ?
Do you fight to escape the paralyzing thought circles and the physical vertigo ? Or is it natural ?
Do you have any advice for current students, or opinions about which skills should be prioritised? I'll soon be graduating in CompSci, but given the rate of AI progress I'm no longer sure what high-impact routes would look like. I was planning to go into (technical) AI Safety, but now I'm doubting I'll have time to sufficiently upskill before it matters
If a god is good and creates all possible net-positive universes, why:
- doesn't it constantly replace suffering with something that can have the same outward behavior but no experience of it? E.g., if at every single moment, most future branches are those where bad things just stop, people mostly continue in worlds where there is no evil.
- doesn't it skip simulating worlds altogether and just simulate all possible minds with good experiences, skipping over bad experiences (inserting memories of having all possible bad experiences, though skipping over the experience of living through the memories if it's bad)?
This isn't bizarre, but there is going to be an *amazing* article on misophonia in Asterisk in a few months (not by me). Also a mediocre article on people with dwarf-tunnel-torture hallucination/conspiracy theories (by me).
If you could choose some n>0 where you die n years from now, which n would you pick? No takebacks, once you choose n you're locked in.
Assume that you stop aging and that you stay mentally at the level you're at now — this isn't a question of how much Alzheimer's you're willing to tolerate. We'll cash out the immortality by saying that anything which would normally kill you just contingently doesn't happen — you don't get hit by any buses, your telomeres just don't degrade, you have time to build / acquire a space suit before entering a vacuum, and so on. Up to you whether an uploaded version of you counts as "you".
Thanks, that was helpful! Contains a lot of things that you frequently see in Scotts writing, but awesome to have it all written out explicitly in one place.
(not sure how much the trump examples hold up - but the overall point stands)
I don't intuitively feel like I have a process, so this makes things hard. Short posts come out with no thought. Long posts are painful but it just seems like thinking hard about them and dealing with the pain, not some kind of fancy outlining process.
A question from my non-subscriber wife, but hopefully you'll allow it: What was the highlight of your week? (Let's say last week since it's Monday right now)
After 8 years of uni, I will finally finish my CS/ML master a few days, but I have very little idea of what to do next (in the short term too, but primarily in the mid- to long-term). What advice do you have for figuring out one's path in life? And/or what advice would you give your younger self?
Hmmm...everyone told me to network, I hated this advice, but it's not bad. I think there's something where getting into "the elite" is more important than whatever object-level thing you're doing. This doesn't necessarily mean going to cocktail parties or something - for a CS master, it could mean knowing enough people in cool startups that you feel like the sort of person who could easily start a startup yourself (and would be able to easily navigate subtasks like talking to VCs). Even if you don't go the startup route, I think a lot of people in Silicon Valley are kind of generalists where they feel comfortable applying to jobs outside their skills-comfort-zone and trusting that their intelligence and work ethic will make it work, and I think this requires some sort of confidence/networking/etc. Being in a position where you feel like you really understand a field/industry and can do whatever thing vaguely related to it seems most interesting to you is a lot more fun than being in a position where you're resume-grinding and trying to get into the three companies that are an exact match for your skills. You rarely start at the former point but it's a useful goal to try to get to.
Also I think "network" is not a framing designed to appeal. The core components are "other people know who you are" and "you know who other people are." Who in your field is worth working with? What skills and qualities do you have, such that when other people are looking to hire someone, they think of you? How did that information flow, from the world to you and from you to the world?
I think a lot of people want the facts to 'speak for themselves' but being deliberate about communication is extremely useful!
Apologies if you’ve discussed this elsewhere, but I’d be interested in your current AI usage. Which model(s) do you prefer? In what areas do you find them most useful? Any custom prompts or unusual use cases?
I mostly just use Claude, partly out of corporate loyalty to Anthropic (most of the safety people ended up there), partly because I genuinely like it. I don't do anything fancy, mostly just ask it questions about topics I don't understand while I'm trying to write posts. My wife has a GPT subscription which I use sometimes to get images.
Oooo! You're holding a quodlibet! My medievalist side is jumping up and down like a kid in a candy store.
Here are my favorite quodlibet questions from the Middle Ages. Of course I can't remember who was asked these when, except that the last question was asked to Thomas Aquinas. Some of these (like the second and third?) may have medieval Jewish roots or counterparts.
1. If a baby is born with two heads, should it be baptized twice?
2. In the perfect original paradise in the garden of Eden, did Adam and Eve have to poop?
3. Less scatalogically: If Adam and Eve hadn't fallen, would there always have been equal numbers of men and women in Paradise? Because, obviously, if the numbers had been unequal, someone would never have found a perfect match.
4. Which is strongest: truth, the king, wine, or a woman?
>1. If a baby is born with two heads, should it be baptized twice?
What's the theory of intentionality wrt baptism? If, for instance, the formula was said in the course of immersion baptism, but the formula was "internally directed" at one head, is it valid for the other in the case of the two being two persons? What if there are identical twins the formula is "directed at" twin A when it is mistakenly twin B being immersed?
I think the most extreme case of this would be a form of monozygotic chimerism. Not even sure there's a word for that given it's likely unidentifiable, despite necessarily existing. You'd effectively have an identical twin absorbed into you in an invisible way, yet its seems to be a difference of degree rather than kind with monozygotic Siamese twins.
1. You sprinkle water on one head, then the other, so no single head gets baptized twice.
2. Obviously, God told them to eat of all trees except the forbidden one, so they were definitely eating - so where would the food have gone otherwise? However, there were no bacteria, so it was all meconium.
3. There would have been more men, because (given that we're assuming Catholicism is true) one of them would have to be Pope.
4. It's a rock paper scissors loop. Wine clouds truth, woman drinks wine, king oppresses woman, truth brings down king. This is just obvious good game design.
Approximately what level of financial grant would be necessary to incentivize you to have one more child than you would otherwise want? (I very much assume your wife would have a say, so the amount would need to be convincing to both of you!) Presume either a one-time grant at birth or distributed over the first 18 years of the child's life whichever makes the question cleaner, but pure money without any additional external assistance.
Bonus round: instead of having an additional child you would raise yourselves, a magically-zero-inconvenience medical procedure results in (genetically) the same child but born to an unrelated mother somewhere else in the United States - the child would be given your contact info when they turned 18, but no other link. What level of grant would be necessary, or would you even pay for the opportunity?
I am not Scott, but I want to take this chance to go on record saying I'm willing to pay five figures for the existence of genetic offspring in a scenario you describe (assuming the parents are healthy and likely to take good care of the resulting children), just in case this isn't purely theoretical.
I don't know. My impression was that the last mayor was slightly pro-growth but got thwarted by the supervisors. The new mayor is an unknown quantity but seems pro-growth, and the new supervisors are light-years better. But I live in Oakland and rarely leave my house, so this is all hearsay.
In the second age of Trump it appears that at least some of what were claimed to be Right Wing Conspiracies (eg where/how USAID money goes, the Deep State as an extension of the Democratic Party) are legit, while claims by the Democrats (eg the Steele Dossier) look like Left Wing Conspiracies.
I find myself every day scanning Twitter and seeing a whole lot stories that seem crazy (but also legitimate in the sense of all hanging together along with such associated data as I can find) while I scan through what the MSM has to say and it looks like Pravda-level fantasy; not exactly made-up facts so much as deliberate distortions, ignoring important details, focussing on the unimportant or trivial, and so on.
Do you have any mechanisms of heuristics for how to make truth-sense of the world in this new environment? "I choose to just accept the ideas of the prestigious people with professorships in Archaeology" was a reasonable hypothesis until we found that the professorships were all being funded by the CIA and USAID! (Not actually, not yet anyway; but certainly in Social Sciences and History I feel we are at that point. I simply don't care that Harvard Professor X claimed something about racism or trans or whatever; I no longer believe that Professor X's credentials have very much to do with what s/he's claiming is likely true.)
Interesting, my impression is that most of the things the right wing has said about USAID money has been proven false. Maybe we're thinking of different claims, but the average Twitter claim has had a half-life of a few hours before someone has proven they're making some ridiculous mistake in reading a database or something.
I guess this is my advice - follow enough people on Twitter that you frequently see debates between people claiming that some of the things one side says are false, then follow the debates, see who wins, and downgrade accordingly.
You could also ask your regular doctor for a recommendation. They likely have a list of good mental health referrals, and may have gotten a feel for which places they refer to actually work out and which ones have their client back asking for a new referral. I work in an outpatient therapy clinic and we get a lot of referrals from medical providers because we've built of a reputation for having availability and doing a good job.
My partner loves Jordan Peterson, and I'd like to introduce him to additional thinkers he might also enjoy. He's right-leaning, prefers video and podcasts to blogs, and has a contrarian streak. Do you have any content creators you'd recommend?
I'm working on a crowdfunding platform (link below) to enable systematic chemical testing of consumer products (initially focused on plasticizers and phthalates, inspired by PlasticList).
Given your interest in aggregating distributed knowledge and your writing on FDA approval processes, I'm curious:
Do you see potential in using crowdfunding models to generate publicly-available safety data? Or would such systems inevitably select for sensationalist concerns over legitimate risks?
Most of the studies I've seen on inositol have been negative, but I've never heard of that particular application before!
If by "reducing unwanted limerence" you just mean sex drive, I'm surprised you would choose it over any of the basically-SSRI-things that are known to work well for that.
In my own example, I was trying to get rid of the persistent, soul-being-torn-out feeling after ending a romantic relationship with my girlfriend. I pre-register the experiment with her (this is a very romantic thing to say to ones former girlfriend?) I used placebo capsules to self blind, but I think either inositol itself or the cognitive effects last longer than I expected. The first time I took the active side, the symptoms went away for >week. Since then, the symptoms have been sufficiently rare and intermittent that I have not been able to collect much more data.
I'd be interested in hearing about other people's experience with inositol for limerence, if any.
(Does anyone have unpleasant limerence they want to fix and want to pre-register a experiment? (Does anyone want to make me fall in love with them and then break up with me so I can run the test again myself?))
There's two things that annoy me about Astral Codex Ten
1. Unlike every other Substack, the comments are not on a separate page to the main blog post. ALL the comments are under the main blog post. I usually read Substack on my tablet, a Kindle Fire, but I can't read Astral Codex Ten on my Kindle Fire, because the page is so large it just crashes!
2. Unlike every other Substack, people can't "like" comments here. This is annoying partly because with so many comments on every post I'd like to be able to see the ones that readers have decided are the best, and partly because sometimes it's useful to be able to see, if I revisit a post, which comments I have liked.
We did have likes but many people (myself included) didn't like them (too easy to turn into a popularity contest, and then weaponise such contests) so Scott listened to the voice of the people and had them taken off. For example, I do not think that "most likes" necessarily means "best comment"; I've seen elsewhere how it just means "yeah man, stick it to 'em!" or "I don't really think this is good but I'm upvoting for the lulz" and so forth.
I think they still work in some way if you really want to give likes.
1. Substack's usual comment policy is that it shows one or two comments, then you have to click "Load Comments" every time you want to see another two or three, then Load Comments again to see another two or three more, and most people never bother and comment sections are useless. I prefer being able to see all of them at once.
2. People have consistently voted down likes for comments, because they think it would turn this into a Reddit style popularity competition. Part of this is that the community started on WordPress where those don't exist, people liked the system there, and they want to preserve it.
I really wish that there were likes. The current system worked very well at earlier points where there were (seemingly) fewer total comments. Now there are enough comments (and, to some extent, my time commitments have increased enough) that there's no way I can read them all, and there's no way to sort them into anything but effectively a random sampling (with some bias toward people who are online a lot and are in a good position to drop whatever else they were doing and comment). Being able to sort by likes is definitely imperfect, but it's much better than having nothing useful at all.
My approach (which obviously does not scale) is to free ride on the rest of the commentariat: if there are more than ~200 total comments by the time I get here, don't bother with any top-level comments that have no replies (or have only self-replies).
That does help somewhat, although in practice not enough that I actually do it except for posts where I'm especially interested in seeing what people say (this post being a clear example of such).
I see. I use a browser to read Substack - I've never used the app - and my experience is that on every other Substack when you see a post you see the top two comments under the post and then something saying "n more comments...", which when you click on it takes you to a separate comments page. And on that page you see a lot of comments (I just checked on a post on Richard Hanania's Substack and it was 54 comments) followed by "Load More", which if you click on it gives you lots more comments. I can understand that if someone can only see two or three comments at a time using the app, and they have to keep clicking to load more comments, they might not bother.
I have no strong opinion about likes (my weak opinion is I like them) but I'm strongly in favor of having all the comments visible at once. Very important for being able to grep (ctrl-f) for things. Drives me crazy that this isn't the case on any other Substack but ACX.
Seconding this. I wouldn't mind the comments as a whole being a separate page (the bit about the article crashing on loading huge threads is legit), but when comments are shown, it should be all of them. Don't break standard browser tools.
(I mean, one could probably get rid of such crashes by *not using !@#$%ing javascript*, but that's just a lost cause on today's web.)
I use an iPad a lot, and for some reason I ended up using the substack app on that. I suppose it seems convenient on that platform. But if I want to search for something, I have to figure out how to open a post in a different app, typically a browser. If you haven’t been annoyed by using a browser to read Substack, I am not sure what I can use to entice you.
The answer to both of these is that it's meant to replicate the original Slate Star Codex experience. Both of these were true of the original site, and after moving, Scott didn't want to change what worked and asked Substack to let his page be that way.
Do you beleive the recent democratic backsliding is an omen of longer and larger trends around post literate society and a new social media culture or is this a blip on a larger trendline?
What are your thoughts on the health impact of sodium consumption? Given a reasonably healthy person, does sodium consumption have much of an impact on blood pressure or other metrics? Is it worth making an effort to reduce sodium intake?
In my wildest dreams I imagine you being asked by Trump to head up the FDA. Would you take the job? Do you ever get asked by anyone related to Trump to give your thoughts on anything? And any hopes the new administration might do something to sort out IRBs?
With the caveat that this would never happen in a million years, yeah, I would take it, just because I assume I could find really good advisors and that they would be better than whoever took the job if I refused.
I have never been asked by anyone in the administration to give my thoughts on anything.
If your slate was clean - no particular attachment to family, friends, workplace, etc. and you had to decide a country to live in based purely on your expectations on how that country will legislatively handle AGI / a potential singularity / AI-risk, where would you live? Between the US, the EU, Switzerland / UK, or other?
just my personal opinion but i'd go to a developing country, provided you can earn an income somewhere. most of the scary scenarios involving AI affect developed countries before they affect undeveloped ones.
I'm a bit confused because I don't expect national level policy to matter. Even if Switzerland has amazing AI risk policy, if the US screws it up and creates unfriendly AI, Switzerland is still dead.
That's fair on ai-risk killeveryone scenarios. I suppose I was asking more for a scenario where everyone is not dead and we manage to create friendly-ish AI and which national level policy would be able to best handle the social / economic consequences. Similar dynamics apply where third countries might have entirely reactive policies to those in the US, but nevertheless would be interested in your thoughts.
I think superintelligence is likely to be pretty winner-take-all - I would be surprised if countries maintained substantial independence and different policies after it while also being anywhere near the technological frontier. I think it probably spirals into utopia or dystopia pretty quickly and regularly and cross-border policy regimes aren't likely to matter much.
What are your thoughts on the recent actions of the Trump administration? I read a lot of news on Reddit, and there are a lot of things where people talk about Trump’s actions as if they’re part of a fascist coup, or they’re going to result in the entire collapse of the U.S. economy. I know that some of this is hyperbolic because Reddit is a liberal bubble, but my impression is that Trump is also doing some legitimately bad and unprecedented things.
As some examples, I don’t have a good sense of how unreasonable it is for Elon Musk to have the access he does to government systems, nor do I have a sense of whether the tent facilities being built at Gitmo are a particularly unusual development.
I’d be curious to hear your take on which issues are real problems and which issues are totally overblown.
I don't think Musk himself is breaking this law, but it seems like he's helping Trump to break the 1973 Congressional Budget and Impoundment Control Act. And I believe there are legal protections for civil servants that Trump may have violated. IANAL.
I don't know enough about law to have an independent sense of how legal Musk's actions are. I don't really trust him to not do them if they're illegal, which is bad.
My impression was that a lot of really good programs got canceled (eg PEPFAR) along with a lot of woke garbage. I wish that he had taken a little while longer to sort the baby from the bathwater, and think that laws exist to protect us from the sorts of people who don't do that.
I think there's a possible world where he gets his act together, restarts all of the ones he shouldn't have cancelled in the first place, and it's long-term positive, but I don't really trust him to be an actual person who cares about bringing that possible world to fruition, and I would rather there be laws than me have to trust him.
See also https://scottaaronson.blog/?p=8609 (Scott Aaronson lamenting the ongoing science funding disaster; see also the link at the end to Terence Tao with a nice analogy for how harmful this can be.)
do you think we should be evaluating our chances of natural disaster and preparing for it? recently have had the realization that i spend a lot of time thinking about existential threats like AGI but not a lot of time thinking about personal physical threats specific to my location
I think more than zero, less than an amount that wastes your time. I have a few tubs of nonperishable food in my basement, a drawer with all really important documents (eg passport) so I can grab them all in an emergency, and otherwise I don't worry about it.
Is Vagal Nerve Stimulation for psychiatric conditions on your radar, and if so what are your thoughts on it?
There has been some reports of it in the press eg [1]. There are purchasable devices, although they make me a bit nervous (one advertises that it is safe because the *FCC* approved it, I guess assuming that people will mistake that for the FDA).
In 'Half an Hour Before Dawn in San Francisco', you write: 'Imagine living when the first lungfish crawled out of the primordial ooze, and missing it because the tidepool down the way had cheaper housing.'
What's do you expect your price is?
Specifically: what compensation would make you seriously consider leaving the Bay Area for 10 years, with these conditions:
It's got to be more than an extra 100K per year because I could save that in housing costs alone. I'm often on the border of moving, so I don't think too much more. Maybe 300 or 400K per year? I would miss my friends and babysitter though.
Do you think legislative secret ballots would be an improvement? Do you have any ideas on how to change the short term incentive structure of government?
I was watching a couple of GPs lament the shortage of psychiatrists and the lack of support from insurance companies for treatment. Do you think AI-powered therapy, with a convincingly human avatar—perhaps even tailored to the patient—could provide meaningful help in closing the gap in the near future?
I was planning to ask this somewhere else at some point but essentially the other day I noticed I was confused regarding the FDA.
First of all, I broadly believe RFK that the FDA is letting too much junk through as there is so much money and will behind the Pharma industry it would be impossible for one little agency to block them from profiteering. Show me the incentives and I will show you the result and the incentives for pharma regulation basically point one way.
On the other hand most of what I know about the FDA I got from you and I summarise your position as the FDA is too rulebound and slow and keeps blocking good research for no reason. I believe you have argued for this extremely convincingly in the past.
So I guess the question comes down to do you still think the FDA is still too slow and hesitant ? Do you have any sympathy at all for RFK's issues with the FDA ? Finally are my two contradictory opinions as ridiculous as they sound or is there a way to hold both intuitions ?
I think these can both be true. In a world where there's a strong expectation that an FDA-approved drug definitely works, is definitely safe, and is definitely novel and exciting, the FDA can be failing to achieve that, and so misleading consumers. But that's a higher bar than what would actually make the country best-off, where it might be better to have 10x more drugs but only 90% of them work, or something like that.
That having been said, I haven't been following RFK's complaints, and I don't know of any really egregious examples except Aduhelm.
Education and intellectual environments (eg so much stuff is written that everyone is literate) which build skills somewhere halfway between problem-solving and test-taking.
dwarkesh tweeted recently a question he asked dario in his interview:
> One question I had for you while we were talking about the intelligence stuff was, as a scientist yourself, what do you make of the fact that these things have basically the entire corpus of human knowledge memorized and they haven’t been able to make a single new connection that has led to a discovery? Whereas if even a moderately intelligent person had this much stuff memorized, they would notice — Oh, this thing causes this symptom. This other thing also causes this symptom. There’s a medical cure right here.
My favorite example of this is etymology. Did you know that "vacation" comes from literally vacating the cities? Or that a celebrity is a person who is celebrated? Or that "dream" and "trauma" come from the same root? These are all kind of obvious when you think about them, but I never noticed before reading etymology sites.
I think you don't make these connections until you have both concepts in attention at the same time, and the combinatorial explosion there means you've got to go at the same slow rate as all previous progress.
Another example: I've only just realized after all these years that "prince" literally means first, as in first in line to inherit the kingdom, and I only realized it because I was translating something into Spanish, where the word is closer to its root in Latin.
"dream" and "trauma aren't obviously related: "dream" is from proto-Germanic *draumaz meaning dream, from proto-Indo-European *dʰrewgʰ meaning "to deceive"; "trauma" is from Greek, ultimately from proto-Indo-European *terh₃ "to hurt." The "vacation" etymology also doesn't seem quite right, it originally meant something more like "freedom from obligations" and the sense of travel came later (though of course it is still related to "vacate" etymologically).
I don't know if this says anything about the difficulty of finding non-spurious connections among all the combinatorially possible ones...
Anyone know how specific the etymology example is to English? Even within English, American English spelling reform sacrificed a lot of etymological indicators.
I have a bet going with a friend, and I try to get an opinion on this from everyone I can.
Do you believe that there will be commercially available humanoid robots capable of autonomously matching human domestic labor (eg cooking, cleaning) by July 1 2034?
It doesn't have to be cheap, just available to the public for some amount of money, and it needs to be fully autonomous, with no expectation of a human being on the other end taking control. As for competence, it needs to be both capable at individual tasks and capable of acting inside a domestic space in a way that wouldn't swiftly get a human being fired. A lengthier discussion of these criteria and my thoughts about them, at the risk of self-promotion, is located here https://nicoroman.substack.com/p/against-humanoid-robots
Seems like a very underspecified bet. Do you mean ANY form of domestic human labor? If so, imdies the bet fail if the robot is bad at any one random thing? If not, how much better than a rumba are we talking here?
More in depth specifications in the linked post. In short, I’m looking at basic competence in cooking and cleaning. Nothing fancy or requiring specialized skills, just tasks which any able-bodied person off the street should be able to do: mop, sweep, tidy up, cup of morning joe and an omelet, wash the dishes by hand and stack them. This whole bet originated from my friend’s claim that, by the specified date, menial human domestic workers would only be able to compete with robots by cost.
If things started to look close, my friends and I would need to sit down and make the bet more detailed, valuing each kind of task and detailing criteria for success on each. At present, neither of us think it’s necessary: both of us expect to win the bet without resorting to squabbles over whether a given robot passed the broom benchmark.
I’d consider doing that if I saw reliable humanoid robot waitstaff: walk around a party on flat floors carrying a tray of drinks, take people’s empty glasses and offer them new ones, etc. That’s the benchmark I’m waiting to see before I get worried about losing this one.
Seems investors anticipate a massive market by then. Money is a limiting factor for domestic use, but businesses and institutions will probably be the key purchasers. You can technically get humanoid robots now absent AI and they cost a fortune.
Yes, because there will be an intelligence explosion in software-AGI and some of the resulting above-human-level AGIs will get redirected to inventing good robots.
Could you elaborate? I have the same expectation as Scott (at least in the branch where no runaway ASI turns every atom from the orbit of Mercury to the Oort cloud into a computronium Dyson swarm...). Why the reaction to household robots?
I see, Many Thanks! ( I _would_ put a probability on AI advances, currently at 80% odds of AGI within two years. Research and development is always chancy and can always run into a roadblock, though I give this work very good odds. )
What are some things you have high confidence (80%+) an arbitrarily intelligent AI will remain unable to do (e.g. faster than light travel, polynomial solutions to currently NP-hard problems). I'm trying to understand what people who lean toward intelligence as arbitrarily scalable view as limits of natural law that no amount of intelligence can get around.
I'm not sure what you're expecting other than "impossible things", of which the two you mention are good examples (assuming P =/ NP). I feel like to know what's impossible I would have to know things about math/physics/etc, which seems like a different skill from knowing things about AI.
I would not expect teleportation, time travel, hypercomputation, etc. Someone (I think Bostrom) had an interesting analysis of the maximum amount of computation that could be performed per gram of matter, but I can't find it right now.
Reflecting on the emergence of AGI on his blog yesterday, Sam Altman wrote that: "it does seem like the balance of power between capital and labor could easily get messed up, and this may require early intervention." Do you see this possibility of power imbalance as one of the most important issues around AGI, or as more of a second-order effect, less important than things like conventional alignment research?
Is there any primer or guidance for how to get educated on the sum of the community knowledge generated before SSC really got rolling?
I began reading SSC around 2017, but I’m aware that there was a formative and then maybe like an expansion stage where a lot of debate was had. I was old enough to be on lesswrong (28 when in 2009), but just wasn’t aware of it until almost a decade later. I’m not even sure what other rationalist and EA primary sites had legit community and intellectual traction before 2015.
whats ur fiction writing process? we love the bay area series and as a guy who's only seen twitter it gives me major twitter vibes. like a collection of little jokes. do you collect them and poast whenever you get to a dozen? how does a joke appear in ur mind?
I'm curious about your opinion on surveys about things like happiness or suicidality, ala "Suicidality has nothing to do with suicide to the point where it's almost a reverse indicator".
I've sorta-kinda stopped trusting these, but I'm very uncertain how justified that is. My usual intuition these days is if group X reports it's exceptionally well-functioning or exceptionally sad compared to other groups to prioritize the view that this is more a *communication style* thing than an actual, trustable indicator. So I'd expect Vegans to report greater happiness due to their diet, but I'd expect them to say that even if it made them miserable.
I put that book of text there to explain the question, not to bias you - I'm looking to have holes poked in this. I think the short version of the question is "Do we only use mood/feelings types of surveys because there's no better option, where an even slightly good one would end that practice forever? Or are they much better than I want them to be?
Happiness research is very tough and you often get different answers depending on how you ask the question, ie "are you satisfied with your life?" vs. "what percent of days do you feel really happy?" etc. When all of these different things point in the same direction, I usually trust there's something there; otherwise, there might be something there but it's probably complicated and not too worth worrying about.
I wonder what would happen if you simply asked vegans and nonvegans "how happy are you?" without mentioning that it involved veganism. I think this works for things like religion and child-having which have major effects, but maybe it would be too weak for diet in particular, or pick up the wrong thing (eg vegans are more liberal, so maybe it would instead be picking up liberalism). I don't know, seems like there might be something in this category.
I think somewhat. It depends a lot on how much capacity my wife and nanny have - when they're both doing well and feel like taking 90% of the work between them, almost no hit. When they're sick or tired and ask me to take on more of the work, significant hit.
I would be curious to see your thoughts on fatherhood, especially this part, elaborated on. I noticed I am the same, but the popular narrative of "Spend as much time with your child as you can" creates a sense of guilt when I am working or relaxing on my own. I've come to terms with it, but it is hard to talk to other parents about it, and I am interested to see you say things eloquently as a way for me to ruminate on it more.
I think one of my thoughts here is probably the same as your - i.e. that to some extent there's some training for "the right answers" where the right answers are the ones that most defend the answerer's value. So if you ask a Vegan how happy their diet makes them, they know the "right answer" is "it puts me next to god in the clouds", so long as the survey is designed in such a way that they can intuit there's a "is veganism good" question trying to be answered.
I admit freely that I'm pretty bad at actually figuring out if there's that kind of accidental/intentional guiding going on, which makes me assume it is, which makes me discount too many things as invalid. What should I be looking for in a survey to see if they controlled for it?
I kind of want to fight your hypothetical here. At least with parenting, people ask happiness related questions, and it seems like the effect you suggest would show that people say they are happy even if they aren't. But happiness research is clear that parents are much less happy than nonparents while their kids are in the house. So there's at least one case where you can get info despite signaling risks.
I think is super hard to control for though. If there is some signaling effect like that, the best you can probably do is ask other questions you expect to be correlated with the answer you want and hope signaling doesn't affect those.
Yeah. Without seeing the specific research you are talking about, I can take it as true. I think I'm uncertain enough on this entire thought that I wouldn't be surprised if I'm looking at it from entirely the wrong angle, or if there's just so many different exceptions to it it's hard to quantify.
1. How convinced are you that your original thesis was wrong?
2. Is there a specific time range that you were thinking of, where the causal effect is denser housing -> higher prices?
3. Is there a specific policy or set of policies you cared about the effect of? Density doesn't increase for no reason, so I think a question like "does density increase prices" is under-determined. See also rule 5 of https://www.reddit.com/r/badeconomics/
I'm pretty convinced that there are two separate effects - cheaper because of more supply, more expensive because of density-related amenities - and that they fight each other on various timescales. When I wrote the post, I thought these immediately canceled each other out and the stronger (the second one) was the only one that mattered. I genuinely changed my mind in the sense of disentangling these more and seeing that you could get one without the other for path-dependent reasons, at least for a long time. The survey convinced me that most real-world cases are probably ones where the first one is stronger at least for a while.
I cared about the policy situation where everything else is held constant, but YIMBYs win a policy victory and cause more housing to get built anyway.
I think induced demand has much stronger implications than people realize. The idea is: building a new home allows someone to move into the city, which creates amenities (start a business, do research, start a club, etc) that cause more people to move in. Since demand increases by more than the increase in supply, prices rise.
So if you want to avoid induced demand, you need to ban all new housing, even in the suburbs: a new resident commuting to their job downtown will also contribute to amenities.
This is what Scott thought at first, and it's contradicted by pretty much all high-quality empirical research of which I am aware. The supply effects outweigh the demand effects.
Also, you probably should be grappling with the fact that "ban all new housing" is effectively what many places in the US have tried, and they've generally seen massive increases in housing prices (exactly as intended--"home values" is a standard NIMBY rallying cry).
Also: amenities are not just "created", they are often relocated from somewhere else. Talented people move to the big city - company in a province town closes shop and reopens in the big city because they can't find talented people in the province anymore - talented people move from the province to the big city because there aren't any good jobs in the province anymore.
Increasing returns *for whom*? There are always good reasons for someone to take a small step that gets them better returns, but makes everyone else slightly worse off. I vaguely remember Scott writing an article about it...
Up to a point, agglomerization is fine - we need cities of some size to support modern civilization. The question is, are ever-increasing, ever-more dense cities an inevitable global optimum, or do they turn into a moloch trap at some point, and efforts to delay that point are fully justified?
I think it's correct that, for example, someone living in Manhattan when it was a tiny Native American village, if their only goal was to prevent Manhattan land prices from going up, should have banned people from building there (especially if they could guarantee that other jurisdictions wouldn't follow their example, and building would remain legal everywhere else).
But I think most people aren't in this position, and it's defecting (since you want growth to happen, just not near you), so governments shouldn't be making policy based on it.
Thanks. I think I understand your position better now than when I first read your posts. And I definitely think that in that policy regime, there's very strong empirical evidence that the supply effects are stronger, at least on timescales of years to decades.
You don't, by chance, have any idea if and where Lou Keep is writing these days, do you? I cannot find anything. Samzdat has been silent for years and Tauromachy nearly as long.
Thanks Scott. I've managed to avoid Twitter entirely and so can't really search there. Glad he's alive, at least, much as I desperately miss his style, perspective, and erudition.
Have you ever looked into the feasibility of the potential worldwide switch to renewable energy sources? A lot of western countries are coming out with rather aggressive net zero targets, and there seems to be some debate about whether government targets for renewable energy are even feasible. It can be hard to determine to what extent this is all just signaling or a realistic depiction of the future, and the possible consequences of that.
Regarding psychiatry (and medicine in general): in your best estimation, to what extent are psychiatrists (and doctors in general) following state-of-the-art research in performing diagnoses and prescribing their treatments; vs. using own experience as their primary guide ?
Depends on the doctor, but I think they're doing a pretty good job at using state of the art, in the sense that there are regularly-updated guidelines, most medical students/residents learn them, and most doctors continue working off them. There's also something called CME where you have to prove you've read some recent research to get your license renewed every few years.
In psychiatry, I think things change very slowly, so that your question is barely relevant. In other fields, they change more often. My father, an internist, says he disagrees with many of the recent guideline changes - but he sure does know about them.
What would an argument look like that could convince you modern poetry is Good Actually? And when, for poetry, would you date post-tartarianism? The fall?
I don't know how you do aesthetic arguments. You'd have to show me modern poems that I like. If I'm trapped in some hating-modern-poems equilibrium, maybe lead me out with a series of poems each of which is slightly more modern than the last, so that appreciating N helps me appreciate N+1.
The last poems I really like are from the 1930s. I would like to tell a story where the trauma of WWI destroyed people's hope and belief in beauty and made them create bad art, but for some reason it's WWII, which makes less sense to me!
C. S. Lewis wrote a satirical bit about the poetry scene int he 1920s and 30s in his allegorical book "The Pilgrim's Regress". Since he was a young academic at the time he would have had direct experience. In this scene our protagonist has been attending a poetry reading:
"Glugly instantly rose. She was very tall and as lean as a post: and her mouth was not quite straight in her face. When she was in the middle of the room, and silence had been obtained, she began to make gestures. First of all she set her arms a-kimbo and cleverly turned her hands the wrong way so that it looked as if her wrists were sprained. Then she waddled to and fro with her toes pointing in. After that she twisted herself to make it look as if her hip bone was out of joint. Finally she made some grunts, and said:
‘Globol obol oogle ogle globol gloogle gloo,’ and ended by pursing up her lips and making a vulgar noise such as children make in their nurseries. Then she went back to her place and sat down.
‘Thank you very much,’ said John politely.
But Glugly made no reply, for Glugly could not talk, owing to an accident in infancy.
‘I hoped you liked it,’ said young Halfways.
‘I didn’t understand her.’
‘Ah,’ said a woman in spectacles who seemed to be Glugly’s nurse or keeper, ‘that is because you are looking for beauty. You are still thinking of your Island. You have got to realize that satire is the moving force in modern music.’
‘It is the expression of a savage disillusionment,’ said someone else.
‘Reality has broken down,’ said a fat boy who had drunk a great deal of the medicine and was lying flat on his back, smiling happily.
‘Our art must be brutal,’ said Glugly’s nurse.
‘We lost our ideals when there was a war in this country,’ said a very young Clever, ‘they were ground out of us in the mud and the flood and the blood. That is why we have to be so stark and brutal.’
‘But, look here,’ cried John, ‘that war was years ago. It was your fathers who were in it: and they are all settled down and living ordinary lives.’
‘Puritanian! Bourgeois!’ cried the Clevers. Everyone seemed to have risen.
‘Hold your tongue,’ whispered Gus in John’s ear. But already someone had struck John on the head, and as he bowed under the blow someone else hit him from behind.
‘It was the mud and the blood,’ hissed the girls all round him.
‘Well,’ said John, ducking to avoid a retort that had been flung at him, ‘if you are really old enough to remember that war, why do you pretend to be so young?’
‘We are young,’ they howled; ‘we are the new movement; we are the revolt. ’
‘We have got over humanitarianism,’ bellowed one of the bearded men, kicking John on the kneecap.
‘And prudery,’ said a thin little old maid trying to wrench his clothes off from the neck. And at the same moment six girls leaped at his face with their nails, and he was kicked in the back and the belly, and tripped up so that he fell on his face, and hit again as he rose, and all the glass in the world seemed breaking round his head as he fled for his life from the laboratory. And all the dogs of Eschropolis joined in the chase as he ran along the street, and all the people followed pelting him with ordure, and crying:
Because we wanted to have kids naturally rather than go through the time/suffering/money of IVF. Turns out we're not very fertile and it still took a lot of time/suffering/money (plus enough fertility drugs that we had twins instead of a singleton), so we'll give up and do IVF next time.
I know that you have read up on Buddhism? Have you read up on Hinduism and specifically Advaita Vedanta? Does that change your thoughts on consciousness?
Do you look at Buddhism/Hinduism as philosophy, exploration of consciousness or just a religion no different than Christianity/Islam/Judaism?
What is your opinion on Moksha/Nirvana concepts in Indic philosophy?
I have read a tiny amount on the Upanishads etc. I think the theory that we are all God deluding ourselves into believing we are individuals has some philosophical/aesthetic attraction, but I'm not really sure what to do with it - like the Simulation Argument, it doesn't necessarily lend itself to life changes.
I find Buddhism/Hinduism very interesting as exploration of consciousness, but I would have to meditate a lot before I felt comfortable having an informed opinion, and I don't seem good at sticking to this.
I'm optimistic that this decade will bring more options for converting dollars into meditation results (jhana/awakening/etc). How much would you pay to press a button that, as long as you held it down, you remained in jhana (or "awakened" consciousness, or similar)?
thoughts on a potential inflammation component of attraction? something like people being able to subconsciously detect diet induced inflammation (or induced through other ways but thinking diet for most people) -- antihistamines as a minor boost to attraction in this case? do you think it's plausible/if so do you think it's significant?
As in interpersonal attraction? You're attracted to partners who have less inflammation? I think this is probably covered under the many markers of general health that we know people use, like skin quality.
How much have you been recommending Silexan/Lavender and have you found out anything new about it from your experience since the last time you posted about it?
I have found every Civ since 4 disappointing (too focused on gimmicks that take me out of immersion - great conquerers of the past just didn't spend that much time planning how their absolutely giant city districts synergized with each other) and probably won't even bother trying 7. Graphics look pretty though.
I have mixed thoughts on 7, but I think it is probably different for someone who tries all the new 4X games to see what they think about system design vs. someone who's still on Civ 4. (Like, one of my takes is "I already played Humankind, thank you very much" but--you haven't!)
There are two main changes, from my perspective. First, the ages are meaningfully distinct (but, coming from Paradox Interactive games, only weakly distinct), and your nation becomes a combination of unique age-appropriate traits instead of, like, having to balance the Greeks who peak early vs. the Americans who peak late. Second, when you place workers, they stay there, and most of your settlements are 'towns' that don't have production (but still contribute to the national economy). So there's much less management of things, and much less clicking--in a way that jives with the "the great conquerors didn't bother with that" experience but still does involve placing districts in ways that synergize. (Most of the synergies are very weak, compared to Civ VI, so this isn't that big a deal.)
I think I like it more than VI but it's hard to tell. (There's a handful of UI elements that desperately need improvements--in particular, when you overbuild a building from a previous age, the tooltip just tells you what the new building will get you, and not what you're losing from the previous building, which seems potentially tragic.) I think the city vs. town distinction is good but, uh, my favorite version of it actually dates back to Imperialism (released 1997). You can change your capital on age transition, in a way that seems historically appropriate but so far doesn't seem to be worth it gameplay-wise, which sort of sucks.
I'm assuming that you're agnostic or an atheist, but if you had to convert to an extant religion based on which one seems to have the best evidence for it or best corresponds to your existing worldview, what religion would it be?
Judaism, it's my culture, I'm pretty attached to it, and my family put in too much work keeping it to let the goyim win.
If I had to convert to a *different* religion, Christianity, because I like the altruism, Lewis and Chesterton are great, and I could finally needle online Christians who go against everything their religion believes in without being vulnerable to the meme about how I don't believe it and it's just a way to make them do what I want.
No. It really would depend on why I converted; right now I'd be tempted to find the most atheist-lite one that isn't hopelessly woke, but I assume if I was actually converting for some reason then that reason would give me more guidance.
Do you feel any pressure/responsibility to sound the alarm harder on AI risk as things have accelerated and you’ve acknowledged it’s a serious and important risk? You don’t seem very urgent about it but not sure
Yeah. I don't think I would describe myself as "not urgent" - I try not to let my blog become strident and moralizing because then I turn away everyone who isn't already convinced. But also, there's only so much I can say without being boring and repetitive. I'm helping with a pretty interesting AI risk project now that I hope will give me more to write about once it's ready.
One post I've long wished you would write is "Sugar: Much More Than You Wanted To Know", particularly now that you have children and you will have to decide on their sugar access and consumption.
* What is your confidence that sugar is unhealthy beyond excess calories?
* What limits (daily? monthly?) would you place on children's sugar intake, ignoring practical concerns and complications?
* How much (dietary % ?) do you predict your children will realistically consume during their first decade?
* Anything else you think about sugar, or present American children's diets?
You've shown interest in Jhanas. What about glimpse practices/non-dual awareness. These were made popular mostly by Sam Harris in his app in recent times I believe.
Getting there feels less structured than Jhanas, but the feeling of doing it is almost like activating a passive skill in an RPG. But you can do it whenever you like, and it just makes you feel good about everything and allows you to relax immediately about whatever important crisis you think is super important right now, but inevitably you'll wonder why you cared so much. Quite useful and incredibly interesting.
You've talked before about how your blogging topics have shifted with your interests and with age. What is the most important / most fun-to-write topic that you no longer blog about?
What have you changed your mind about most over the last 10 years/5 years/year?
How do you balance work with having small kids? Do you have any tips or tricks that help you stay productive?
You said recently [1] that you'd "like to be more observant, but the prospect of following Jewish law in its entirety feels overwhelming, and picking and choosing feels too unprincipled."
How come you're interested in observance if you're not a believer? Did you find any practice here that works for you?
[1] https://www.neonarrative.us/p/an-interview-with-scott-alexander
Can I cancel you? If not, a different question.
In an AMA like this, do you think it's better to ask just any question that comes to mind, or take the time to ask a _really_ insightful question?
A more interesting question would be: What questions have people not asked you but would be really interesting? (Besides this one.)
Have you ever closely looked at the evidence for big recent Catholic miracles? e.g. human heart tissue in Eucharistic host, the Fatima apparition, healings at Lourdes
(for disclosure, I'm a Christian but not a Catholic)
Even Catholics are not obliged to believe in those miracles, so it's something that could be extremely contentious. Faked relics and miracles have been knocking around forever, more or less; see Chaucer poking fun with the likes of the Pardoner who admits he just produces animal bones, claim they are saints' relics, and gets the sheep to be fleeced to venerate them and pay for the privilege.
Do you have a take on all the recent alien stuff / have you looked into it (the hearings / people in the goverment saying that there have been all these UFO incidents historically)?
There's a reference to Homestuck in UNSONG. Have you read Homestuck before?
I read up to about panel three hundred billion and nothing had happened yet so I gave up. I can't remember what the UNSONG reference was!
when Aaron says "I was doing this. I was making this happen." in Malia Ngo's office. Although technically that's a reference to Sweet Bro and Hella Jeff, which is like a Homestuck spinoff.
Does progress actually make people durably happier? (I feel like it can't. GDP/PPP per capita has increased 20x or 30x over the past 1000 years. If progress increases happiness, then... man, people living 1000 years ago must have been SO depressed!)
Yeah, I also find this confusing. Rich people are on average happier than poor people, and people in rich countries are on average happier than people in poor countries, but these are all smaller effects than I'd expect. I have more about this at https://slatestarcodex.com/2016/07/20/things-probably-matter/ .
But here's a useful thought experiment - would your kids dying make you sad? It seems like it has to, right, this is basically the definition of tragedy. But 50% of kids died before adulthood in the past, and very few do today. So if there hasn't been any happiness change since then, it condemns our belief that kids-not-dying creates happiness as much as our belief that economic growth does.
I think there's got to be something wrong with this - either with the research itself, or with some kind of happiness treadmill that scales back how we feel about most things (not everything) as things improve. See also https://www.astralcodexten.com/p/unpredictable-reward-predictable .
Do we actually know that there's been no change in happiness?
My intuition is that richness leads to happiness through financial security and relative status seeking, and much less so through physical comforts.
Financial security is great and leads people to feel comfortable making babies which enhances community, and on the low end maybe leads to less crime and despair.
Relative status seeking is zero sum and a culture that promotes it will always have a class of people feeling horrible. Social media amplifies. Maybe promotes crime.
Physical comforts: I don't really see how this could make a whole lot of difference in happiness. Having an extra bedroom to store crap in you never use, or having power windows, a bigger TV?
Outside of this, the pleasantness of your lived environment has to factor in somewhere. People pay A LOT, both in money and in physical comfort to live in cute walkable old towns built when you could still build those. Or beach towns or lakefront towns.
It's like the story Sam Spade tells in "The Maltese Falcon" about a missing person case he worked; the guy just disappeared one day, left his wife and family and old life behind, and couldn't be found, until five years later a chance sighting led to his discovery:
"Here's what had happened to him. Going to lunch he passed an office-building that was being put up--just the skeleton. A beam or something fell eight or ten stories down and smacked the sidewalk alongside him. It brushed pretty close to him, but didn't touch him, though a piece of the sidewalk was chipped off and flew up and hit his cheek. It only took a piece of skin off, but he still had the scar when I saw him. He rubbed it with his finger--well, affectionately--when he told me about it. He was scared stiff of course, he said, but he was more shocked than really frightened. He felt like somebody had taken the lid off life and let him look at the works."
Flitcraft had been a good citizen and a good husband and father, not by any outer compulsion, but simply because he was a man who was most comfortable in step with his surroundings. He had been raised that way. The people he knew were like that. The life he knew was a clean orderly sane responsible affair. Now a falling beam had shown him that life was fundamentally none of these things. He, the good citizen-husband-father, could be wiped out between office and restaurant by the accident of a falling beam. He knew then that men died at haphazard like that, and lived only while blind chance spared them.
It was not, primarily, the injustice of it that disturbed him: he accepted that after the first shock. What disturbed him was the discovery that in sensibly ordering his affairs he had got out of step, and not into step, with life. He said he knew before he had gone twenty feet from the fallen beam that he would never know peace again until he had adjusted himself to this new glimpse of life. By the time he had eaten his luncheon he had found his means of adjustment. Life could be ended for him at random by a falling beam: he would change his life at random by simply going away. He loved his family, he said, as much as he supposed was usual, but he knew he was leaving them adequately provided for, and his love for them was not of the sort that would make absence painful.
"He went to Seattle that afternoon," Spade said, "and from there by boat to San Francisco. For a couple of years he wandered around and then drifted back to the Northwest, and settled in Spokane and got married. His second wife didn't look like the first, but they were more alike than they were different. You know, the kind of women that play fair games of golf and bridge and like new salad-recipes. He wasn't sorry for what he had done. It seemed reasonable enough to him. I don't think he even knew he had settled back naturally into the same groove he had jumped out of in Tacoma. But that's the part of it I always liked. He adjusted himself to beams falling, and then no more of them fell, and he adjusted himself to them not falling."
In the past, child mortality was high. People were sad when their children died, but it was a fact of life: they adjusted to the beams falling. Then we got medicine and progress and all the rest of it, and child mortality decreased by a great deal. Now people might be happier about the chance their child would survive and not die. But then we got used to "your kids won't die" and that as a source of happiness faded away: we adjusted to the beams not falling.
Same with everything else in modern life. When it's normal that you have running water and indoor plumbing and heat and food and all the rest of it, that ceases to be a source of especial happiness ("we got a telephone!") but rather becomes part of what is expected and normal. Now we become unhappy about new things ("my phone isn't the latest iPhone!")
I believe that almost all of happiness is relative, not absolute. Which would explain all of these findings.
It's nearly impossible to put ourselves in the shoes of people with different expectations for quality of life than us. In response to your thought experiment: I can *almost* imagine someone 1000 years ago saying, "I was utterly devastated when my baby died. But... only 3 of my 8 children have died. My neighbor had it far worse than me. And we rarely go hungry for more than a day or two. So overall we're doing OK." It's really hard to imagine feeling this way... but the only other alternative is to believe that people were depressed beyond imagination back then, which I'm pretty sure is not true.
It is literally incomprehensible to us that improved health (and other life improvements) don't durably increase happiness in a meaningful way. And yet I'm pretty sure it's true. Which kind of implies that progress is not a very useful goal to try to achieve.
Of course, from an evolutionary POV, this is very easy to believe. Imagine that our quality of life gets 20% better over some period of time. Person A experiences happiness in absolute terms. They become durably happier due to this progress. They become more content, and so they strive less to improve their life. They take it easy.
Person B experiences happiness in relative terms. They see that they are still no better off than their neighbors. They keep striving.
B people will outcompete A people. Evolution ensures that happiness is mostly relative, but it *feels* completely absolute.
Progress causing durable happiness is evolutionarily maladaptive, unfortunately. (But *thinking* that progress will cause durable happiness is extremely adaptive!)
I totally agree. My great-grandparents had something like 8-10 children. One of them was named Maria, and she died in the first year or so. Well, then they named the one after that Maria again. I haven't met them personally, but from what I was told, they considered the second Maria as a replacement for the first one, and were satisfied with that. I don't think that "a child died" was such a terrible tragedy for them. Children were to some extent fungible back then.
And yet it's worth being careful in how we talk about this, because this kind of thinking is so alien to most people. (Myself almost included.)
Telling someone "I believe that children dying wasn't as big of a deal 100 years ago, they were somewhat fungible" will disgust a lot of people, regardless of whether it's true or not.
"I have more about this at https://slatestarcodex.com/2016/07/20/things-probably-matter/ ."
How do you remember/search for 8.5 year old blogposts like this?
It is pretty clear that there IS a kind of happiness treadmill. See, for example, the book "Never Enough" by neuroscientist and former addict Judith Grisel. The book explains that we are happy when things seem to be going better now as compared to some kind of average of the recent past. Hence the futility of seeking happiness via drugs. Yes, they can create happiness now, but they also "raise the bar" for future happiness.
What are the three topics most likely to cause you existential angst?
What's your favorite book by C. S. Lewis? Mine's "Perelandra" though that doesn't seem to be a common opinion.
"Till we have faces" is amazing.
Its probably his best work of fiction. A real page turner! And the bit with the trial of the gods is wonderful.
Section 7 of your Nonfiction Writing Advice -- https://slatestarcodex.com/2016/02/20/writing-advice/ -- from early 2016, before Trump's election, is fascinating to reread in 2025. I was just chatting with someone about how the Trump example there holds up. The 2016 claim was that Trump was skillfully, via tribal signaling, persuading conservatives about liberal ideas. Do you think, in retrospect, that's what was going on?
Assuming Prospera survived at least 20 years, expanded to 1 million people, and worked at least half as well as advertised, how much would you update toward Anarcho-capitalism?
Not much. Prospera isn't anarcho-. It's just normal, having-a-government capitalism. We already have ample evidence of how well this works and I don't think one more example in an admittedly-weird jurisdiction would update me much.
I'm always too late on these AMAs...
You left a comment on an old LessWrong post observing "premodern people seem weirdly productive compared to moderns in a lot of ways.”
I've always wondered what meant.
I can't remember what premodern people I was talking about, but a lot of them seem to wake up at the break of dawn, do twelve hours of hard labor a day, and never (noticeably) have chronic fatigue or get burnout.
I agree that a plausible solution is "if they don't do this, they'll die, and that's a great motivator!", but I do think it's helpful to notice the puzzle.
I just visited Poverty Point in Louisiana and had a similar reaction. "This mound was built by a few hundred to about 1000 people in 3 months, carrying a total of about 15 million 50 pound baskets of dirt." Today they have tram rides to take people who can't or don't feel like walking on tours around the ~2.5 mile loop to see the site.
I think you had said you did something along the lines of what Experimental Fat Loss had done to lose weight prior to your wedding. Can you share any details about that?
I lost about 30 lbs in a year on a relatively simple diet of eating the same few healthy foods every day. I was going to write about discovering the perfect diet, but I figured I should wait to see if I gained it all back first.
I went about a year without gaining it back. Then I had kids, got more stressed, overate more, and - yup - gained it all back.
Then the kids started sleeping through the night, my life stabilized again, and I tried the exact same healthy diet as before - and it had zero effect, barely even lost 1 lb in six months of trying.
Three days ago I gave up and started Zepbound (an Ozempic relative). We'll see how it goes!
The first time I tried the keto diet, it was super effective, I lost weight rapidly, my health was fine. However, I really hate the keto diet, stopped, and gained the weight back.
When I tried the keto diet a second time, I experienced almost no response.
I wonder how common that is. I haven't tried one of the Ozempic drugs, but it's on my to-do list.
Do you (still) have time to play Civilization or other map games?
Yes, regularly!
Which ones?
Scott answers this further down, search for Civ 4.
Do you think the Rationalist community could be doing more for humanity than just EA? For example, maybe if we did some copy editing on HPMOR, lobbied, and advertised, we could grow more appreciation for epistemics in public schooling.
I've found that people either find rationalist stuff instantly exciting (1%), or bounce off of it bu either finding it boring or having dumb objections that they won't drop (99%). My preferred strategy is to make sure we have enough reach to catch the 1%, then gradually bring the 99% over by exposing them to things that sort of assume it and obviously work (like prediction markets).
I know what you mean. But maybe we just need to target the right people rather than the public. Like imagine getting Niel Degrass Tyson, Bill Nye, or the right bullion gushing publicly about it and trying to get public schooling onboard.
Aren't all those people actively uncool with most tribes right now? I think I would pay money to have them vociferously oppose us.
To put it another way, a portion of rationalism won through popularity of science. Whatever the mechanism science succeeded by, with effort we should be able to get the rest of rationalism popular.
If it's through practical application and "winning", we need to find a way to amplify that into something visible.
No doubt from a politics/tribes perspective we are at a disadvantage. I guess I'm imagining spinning it whichever way is convenient (if only THEY would think RATIONALLY like us, we should make rational thinking required!) I also was hopeful people in charge actually could be convinced. Someone out there decided to reach evolution and not creationism...
Could you give examples of said dumb objections?
Some of the ones I respond to at https://www.astralcodexten.com/p/in-continued-defense-of-non-frequentist
Thank you so much for keeping this wonderful blog! Really got me through some rough times in my recent struggles with pregnancy and being a new mom. I've been really enjoying re-reading Unsong and some of the posts one the old blog, too.
To what extent did you and your wife follow the obscure pregnancy interventions you mentioned in your post on obscure pregnancy interventions? (I'm a bit curious about one in particular: the post says it's probably bad to use all painkillers including aspirin. I ended up taking baby aspirin every day for the second half of my pregnancy on recommendation by my obgyn (because of advanced maternal age, I think?). My understanding is that it was a trade-off between risk of pre-eclampsia and risk of causing cardiovascular problems for the baby. I was wondering if you also had to make any difficult decisions like that?)
Also you mentioned in one of your posts that your babies had trouble with direct breastfeeding at first. I was wondering to what extent it got better? (Been on more than two months of triple feeding (breastfeeding, then bottle feeding, then pumping every cycle), since that's what the lactation consultants and pediatrician suggested; it's been tough having no end in sight.)
Yeah, we got told the same thing - 325 mg is bad, but 81 mg is net positive because of preventing pre-eclampsia (which we were at extra risk of because of twins). I think my wife still didn't end up taking it because she got too many digestive side effects.
The breastfeeding situation didn't get better. My wife slightly blames me and my parents, because she said to wake her up at night whenever the baby cried, but we felt so bad for her being so exhausted that we bottle-fed the baby instead the first few days without waking her up, and then her milk production kind of died down. But we realistically weren't going to be able to solely breast-feed both twins anyway, so we gave up pretty easily and switched to bottles. The babies liked the bottles and bonded with us in other ways and there was no problem.
I'm working on a crowdfunding platform (link below) to enable systematic chemical testing of consumer products (initially focused on plasticizers and phthalates, inspired by PlasticList).
Given your interest in aggregating distributed knowledge and your writing on FDA approval processes, I'm curious:
Do you see potential in using crowdfunding models to generate publicly-available safety data? Or would such systems inevitably select for sensationalist concerns over legitimate risks?
https://laboratory.love/
What are your thoughts Journavx / suzetrigine the new non-opiod pain medicine that just came out. Likely to reduce the risk of opiod abuse?
I don't really know anything about it yet and haven't tried it on anyone. My very weak prior is that everyone is always coming out with new medications branded as "Exactly as good as this addictive medication you're reluctant to use! Switch all your addictive-medication patients to it!" and then it's never really as good.
Or they’re good for pain, but just as addictive!
I’d suspect this one won’t be as effective in reducing pain. But probably not addictive.
I remember you stating at one point that you have misophonia. What do you do, if anything, to manage it? (I imagine the standard headphones + music but what else?)
We’re going to the bay area soon for a visit and I was wondering if there are any rationality-related sightseeing activities available. Like, is there a plaque on the tree that dropped an apple on Eliezer Yudkowsky’s head?
Lighthaven is best, but it's not usually open. However, there's a public Sequences reading group Tuesdays at 6:30 PM, see https://www.lesswrong.com/groups/LK6GNnKp8PDCkqcxx for details. I am told that the rule is that Aella slaps anyone who hasn't done the reading beforehand (this is literally true, not a joke, although the exact penalty varies from week to week and that was just a particularly memorable one).
You can go to CFAR or see one of the many rationalist group houses in Berkeley.
If you were having kids 5 years from now, and wanted to use a new reproductive technology that could offer significant benefits, how many healthy babies from the technology would you need to see before you were comfortable using it yourself?
(And if the answer is zero, how many healthy monkey babies would you want to see?)
Obviously depends a lot on specifics. We're shaping up to be in the first 100 or so to do polygenic selection on traits, but I don't care because there aren't many ways it could go wrong (even if they mess up everything, we just get a random embryo, which is what would have happened without them).
If it's some kind of highly complex gene editing, then I would want - not necessarily more babies, but waiting longer to see if there are things that only show up later. I vaguely remember hearing that the thing where Dolly aged prematurely didn't pan out, or doesn't happen anymore, but that's the kind of thing I'd be worried about. Or maybe not, you could try to convince me that this was an unlikely failure mode.
...at least, that's the responsible answer. If it was something really cool like superbabies, I might be tempted to be early, just to support mad science and be able to say we were one of the first.
Have there been any events that have made you much more pessimistic for humanity's future over the last year?
Similarly, have there been any events that have made you significantly more optimistic?
The Trump election wasn't great. I don't necessarily think he'll be some kind of extreme perma-dictator or something, but I don't think his administration is likely to be very good at handling the Singularity and I think there's a decent chance it happens before 2029.
Can you share your thoughts/research on how you can help your little one develop language, reasoning etc?
What kind of change in policy and/or in the underlying culture would lead you to update your currently negative outlook on Italy? Also, somewhat related, what is your opinion on the value of studying classical languages and literature, both in general and more specifically for someone who identifies as an intellectual?
Do you think it's possible to train yourself to be more agentic? If so, how?
is masturbation healthy? Which psychiatrists are right/wrong?
How many ceteris would have to paribus to have made it worthwhile to have kids earlier, for you? Every Sad Singles' Day that goes by has me thinking of the rising startup costs of reaching the same milestone myself...what's the Pareto optimal point of age-enhanced wistom and resources, vs. the waning ardour of youth?
I got really lucky, in that I married someone younger than myself and got the advantages of a stable career etc without having to compromise on fertility.
I don't know if I could recommend this strategy to other men, because at some point your chances of getting a woman sufficiently younger than yourself for it to work go down, and it's not worth the risk. And I'm not in the right place to recommend/disrecommend it to women at all. But if it works, it works really well.
Would you consider trying to get a collection of your short stories/humorous published in a collection to try to make a splash on the literary scene and perhaps submit for award consideration? Maybe try to write a novel in the tradition of the "novel of ideas" in the future? You've written some lovely short stories in the past that show a good command of emotional modulation, humor, and character development in addition to wrestling with philosophical questions.
Not really. I think blogging is now big enough that the sort of splash I could expect from the literary scene wouldn't be a huge benefit. Exceptions if I got a Pulitzer or something but I think that's out of reach, both because it's not my style (I'm not that serious and sophisticated and in touch) and for the usual reasons that it's crazy hubris to think you could win a Pulitzer.
What do you think about the value of MDMA for everyday non-clinical therapy? I am especially thinking about couple therapy, where in my experience MDMA enables people to talk to the partner in a very connected way about difficult issues between them.
Do you save for retirement, given your views on AI timelines?
Yes.
Partly this is out of normal outside view "don't bet your whole future on your idiosyncratic crazy beliefs being true" considerations, and partly it's out of the possibility that a Singularity which doesn't kill all humans will massively multiply capital and obsolete labor, so you'll want to have money saved up if it happens.
For a while you have read biographies of various dictators. I think you have looked for patterns in their strategies for dictatorship, so how they consolidated their power after they came into office.
Would you say that there is a big pattern (like, which checks and balances to dismantle, how much to build up their popularity, how much to control the media etc), or are the persons and countries too different for comparisons?
I think the main patterns I've noticed:
- Usually this doesn't happen to healthy countries. On a cosmic level, nobody "deserves" dictatorship, but on a mundane level, the countries it happens to have really been screwing their populations over for a long time.
- Usually comes on the back of right-wing anti-elite sentiment, and the elites usually have been extremely out of touch and terrible leaders.
- Often the future dictator builds credibility by being a genuinely good leader beforehand, or at least lucking into a good track record. Often they are more honest than their opponents, or at least differently corrupt, or at least good at seeming this way.
- Usually less about "literally kill every opponent" and more about someone who could naturally get 55% support in an election tweaking things enough to get 65% or 75% support; but they still sometimes lose elections if they really screw up!
- For some reason dictators really like sports.
Thanks for the great answer!
In some of the media I follow, a lot of hoopla is being made of late about near-death and other mystical-type experiences challenging materialist scientific presumptions about how things work (eg. Sebastian Junger, Ross Douthat). It seems like such experiences can only be one of 3 things: lies, hallucinations, or veridical in some sense. What probability do you place on the third explanation (veridicality) being correct, ever, and can you please explain why?
Pretty low. I remember reading about a cool study where some operation tended to produce near death experiences. The patients would say they were floating above their body or something. One of the surgeons put some kind of really bizarre thing on top of a bookshelf, such that a normal person couldn't see it, but if you were floating on top of the room you definitely would. They asked the patients who said they saw the whole thing from above what was on top of the bookshelf. Nobody could answer.
This study sounded really interesting so I decided to try and hunt it down. I think this comment is conflating two (similar in concept and similarly-named) studies by the same researcher, Sam Parnia.
The first study, AWARE (https://doi.org/10.1016/j.resuscitation.2014.09.004), used the experimental design described here. The operation in question is CPR following cardiac arrest. Basically what the researchers did is they installed the aforementioned shelves in roughly 1000 hospital rooms in critical care wards where they expected cardiac arrests to take place. The items placed on the shelves were images of "a combination of nationalistic and religious symbols, people, animals, and major newspaper headlines". Unfortunately, most cardiac arrests in the participating hospitals didn't actually happen in these rooms, and of the 2 subjects who survived cardiac arrest and reported out-of-body experiences, neither were in a room with an installed shelf.
The second study, AWARE II (https://doi.org/10.1016/j.resuscitation.2023.109903), used a somewhat different design, wherein researchers wore pagers that alerted them to cardiac arrests, upon which they would bring a tablet into the room showing an image randomly selected from a set of 10 images, and they would also put bluetooth headphones on the patient playing audio that consisted of the words "apple", "pear", and "banana". Among 28 survivors that this was done on, 1 was able to recall the correct fruits, and none were able to identify the image on the tablet.
I was honestly just trying to find out what the object on the shelf was and didn't expect to go down a rabbit hole.
Thanks for the papers! The second study is particularly interesting. To expand a bit:
They had over 500 subjects, but only 53 of them survived and due to health concerns they were only able to interview 28 of them about their experiences. Of those 28 only 11 had experiences. Of those 11, 6 had the classic "out of body experience" thing. Though the specifics varied, all 6 reported 5 similar stages of the experience: Being separated from their body, heading towards a destination, undergoing a meaningful re-evaluation of their life, arriving at a good place (heaven like), and then returning to their bodies again. Notably all 6 reported feeling lucid during the experience, aware of their situation and identity while being able to reason and remember.
As far as the tablet images go, none of the 6 reported noticing the tablet. None of them reported seeing a different or wrong image on the tablet then what was displayed, and none of them said there wasn't a tablet there. So it's not exactly a smoking gun refutation: presumably if the experience was veridical it wouldn't' be unreasonable for someone not to notice a tablet computer on a bookshelf while they're watching themselves on the table and then traveling to heaven or wherever. One of the subjects reported watching the doctors put electrodes on his chest and shock him, and the researchers were able to confirm that he was indeed shocked with electrodes. None of the other subjects provided any information that could be checked like that.
Some fun quotes from the 6 who had the classic experience:
"I was no longer in my body. I floated without weight or physicality. I was above my body and directly below the ceiling of the intensive therapy room. I observed the scene that was taking place below me ... I, who no longer was the body that had belonged to me just a moment prior, found myself in a position which was … more elevated. It was a place that had nothing to do with any kind of … material experience.”
“I remember entering a … tunnel. The feelings I experienced … were much more intense than [usual]. The first feeling was a feeling of intense peace. It was so calm and serene with an incredible amount of tranquility. All of my … worries, thoughts, fears, and opinions were gone. The intensity of the tranquility was so incredible and overwhelming that there was no fear in what I was experiencing. I had no fear about where I was going and what to expect when I arrived there. Then I felt warmth … Then there was the desire to be home.”
“I do remember a being of light … standing near me. It was looming over me like a great tower of strength, yet radiating only warmth and love … I caught glimpses of my life and felt pride, love, joy, and sadness, all pouring into me. Each images was of me, but from the standpoint of a being standing with me or looking on… I was shown the consequences of my life, thousands of people that I'd interacted with and felt what they felt about me, saw their life and how I had impacted them. Next I saw the consequences of my life and the influence of my actions.”
"“I went directly to a place of light. It was calm and immediate … The place where I was I perceived to be analogous in a way to the exterior of an entry way…There was one major being of love and many other beings of love … There was nothing but love, goodness, truth, and all things to do with love. There was no room for fear or evil or anything but this love. It was more wonderful than any of my best hopes or experiences [in this place]. It was beyond perfect and loving, as we in our human state know it. There are no words to describe it. I was so happy to be there.”
“I was asked if I wanted to come home (meaning there) or wanted to come back here. I told them that my two sons needed me and I had to go back. I was suddenly in my body again feeling my achy joints flaring in pain. I really don't remember what was going on around me at that point, just that I hurt.”
Related: The out-of-body experience can be artificially induced by neural stimulation. You don't need to almost die for it.
https://pubmed.ncbi.nlm.nih.gov/12239558/
Ah, but true near-death experience has never been tried!
As a fellow healthcare provider, I’ve been seeing a lot of folks lately who are very distressed about the state of the world — more than usual. What approach do you use to address these fears? It’s a tricky subject and it has a lot of potential to derail the appointment, but some people seem almost paralyzed with worry.
I think first step is to figure out whether this is a sign of a deeper psych problem or just nonpathologic (albeit often unreasonable) worry. Normal psych screening can help with this.
If it's nonpathologic, then - while validating their concerns enough to let them know I'm not objecting merely out of love for the other political party - I try to figure out the pattern of what's bothering them and what points of leverage we have. Often it's something like the news triggering them, in which case I gently suggest they watch the news less.
If they're skeptical of this solution, I try to figure out whether it's because they feel ethically bound to know about bad things, or because they *want* to watch the news less but don't have the willpower to comply.
If they feel ethically bound to know about bad things, I very weakly try to push back ("Is this really helping the victims of these bad things? Wouldn't you know just as much if you only checked the news once per week? Maybe you could be slightly less informed but help in some other way like volunteering?") but eventually give up because I think this is slightly out of the proper realm of a doctor and now I'm just arguing with their lifestyle and ethics.
If they want to stop but don't know how, I talk about some willpower solutions, like giving their social media password to someone else, or getting rid of the TV, or making a specific schedule ("check the news once a week, Tuesday at 8 PM, and otherwise avoid it, and make sure all your friends know and are holding you to it"). If their problem is getting into doom spiral conversations with friends ("Did you hear about outrage X?" "Oh, you think X is bad? You are like a little baby, outrage Y is so much worse!") then I ask them if they endorse these and whether they could talk to their friends about doing this less.
If none of that works and the patient is a total wreck, I'm not above occasionally suggesting SSRIs to prevent them from being miserable all the time even if it doesn't quite meet depression/anxiety criteria. I would consider myself overprescribing if I did this for more than a small fraction of patients who come in with Politics NOS.
Thanks! I like Politics NOS!
What writer most unexpectedly influenced you? It doesn't have to be one with an absolutely large magnitude of impact, just large relative to what you expected when you heard about / read them initially.
Two questions; feel free to answer only one and ignore the other.
One: have you seen the poem “Scooping the Loop Snooper”? It’s a proof of the impossibility of solving the halting problem in the style of Dr. Seuss. I’ve been meaning to mention it every since reading your review of the other Scott A’s quantum computing book. (If you haven’t seen it, it’s free on the author’s website; short and sweet.)
Two: have you considered making some of your old live journal posts public again? It’s really annoying to use archive.org and you’re no longer “in hiding”, right?
I like the MLK day poem and the Swerve book review and I think there were many other goodies but it’s annoying to browse now
What probability would you assign to the US getting dysfunctional and/or dictatorial enough in the next four years that you and your family will attempt to emigrate, and why?
5%. I (white, cis, etc) don't expect to be a target of Trump's, and if I'm not personally a target, then my house/family/career all being here makes me unlikely to move just because he's being horrible to someone else. I'd rather stay here, make enough money to donate to good causes, and vote against him (assuming voting still exists). I'm not sure I'd leave even if he suspended democracy and elections - obviously this would be bad, but me being somewhere else doesn't help other Americans and is probably worse for me.
I think there would either have to be such widespread economic collapse that it was worth me switching to a career other than blogging (since blogging makes me the same money regardless of where I am), some kind of civil war that makes me feel threatened by violence, or some kind of unprecedented attack on extremely basic freedoms that I don't think Trump's base would go for.
Thanks, this is good perspective.
FWIW, as a parent of an older child, and someone demographically similar and thus similarly not-target-shaped, here are the scenarios I worry about that add up to somewhat more than 5% for me:
1. civil war, as you mention
2. economic collapse caused by war with China
3. pandemic where other countries get a vaccine and we don't because RFK
Regarding medicine and AI: do you think that as AI continues improving its diagnostic and treatment-recommendation skills, that psychiatrists will end up focusing more on psychotherapy again because AI will be able to do much of the rest of our jobs about as good as we can? It seems to me that the physical presence of another human is something both centrally important to psychotherapy and is something AI can't replicate. And while AI might be able to do some forms of psychotherapy kind of OK, ultimately if you're doing serious therapy with an AI bot, it's just not the real thing (and even a human-looking robot therapist would either have to disclose it isn't a human or else lie to the patient).
I think insurance reasonably wants to avoid paying psychiatrist prices for therapy, and that if mental health refocuses around human therapy then therapists with PsyDs or LMFTs or whatever will provide it.
But also, I don't see regulation as favoring AI, both for normal regulatory capture reasons and *especially* in psychiatry where we're so often prescribing controlled substances. Even if AI is able to do our job perfectly well (as it's pretty close to already) I think there will still be a psychiatrist with an MD typing the things into the AI and making the final decision.
Makes sense, appreciate the reply!
It seems like recently the right-wing ratsphere has had a bigger impact and influence with their political sphere than the left-wing ratsphere has with theirs. Would you agree with that assessment and if so any insights into why you think that's happened?
I don't think there really is a right-wing ratsphere. If you mean like Richard Hanania, he is an interesting person with lots to say but I don't think of him as part of the rationalist movement.
I think the reason right-wing intellectuals are generally more influential than left-wing intellectuals is that there's many fewer of them (90%+ of college professors are left wing, etc) so there's much less competition.
"I think the reason right-wing intellectuals are generally more influential than left-wing intellectuals is that there's many fewer of them (90%+ of college professors are left wing, etc) so there's much less competition."
Even the sum total influence of right-wing intellectuals seems bigger to me than the sum total of left-wing intellectuals. You might even expect it should be the other way around if there are more left-wing intellectuals.
Good point. See section 2 of this post for my best guess: https://slatestarcodex.com/2018/05/23/can-things-be-both-popular-and-silenced/
I think while the left wing 'intellectuals' are the majority in theory, the majority of them who engage with politics are not acting as intellectuals but as plain ideologues.
I'd guess that one reason more right wing intellectuals actually engage with politics is that they are going against the consensus. Left wing intellectuals don't have the same incentives.
How do you have conversations with everyday people in your life about likely rapid advances in AI, associated risks, and how they may affect their lives? I have yet to stumble on key concepts and intuition pumps that keep people open minded and guide the conversation. My peer group is mostly curious, recent liberal arts grads who mostly get AI content in the form of "LLMs are stealing", "AI art bad", "AI is just hype" and see most discussion through the lens of those arguments. When I break through to substantive discussion, we often get hung up on discussions I can win, but can't persuade. For all the "we've seen the scaling laws hold over 6 orders of magnitude, I think *you* have the burden to show they'll fail" or "No, I can't tell you how ASI would manifest, maybe through persuasion, maybe conducting business over the internet, but just as one can think Stockfish will win at chess without knowing how, we can assume AI will outmaneuver us" I might muster, it doesn't do much to convince a skeptic.
I have met a lot of people with those reactions and mostly stop trying to discuss with them. I figure if I'm on record saying things they don't believe, but don't press too hard, then eventually they'll see it for themselves and might remember that I tried to tell them. But more and more, I am starting to be able to just point to real things current AI can do. The gap between what I expect to actually happen, and what I can say before being dismissed out of hand, is shrinking. This is in some ways welcome, in some ways disturbing.
I recently got put in charge of a project at work figuring out what AI tools we should be using, because my boss knew I had an interest in the field and was trying to track advances anyway. "Here's a 50 page Deep Research report I got after an hour chatting with o1-pro and o3-mini-high to gather context, without doing any other optimization or putting in any of my own knowledge. This needs to be the minimum baseline for what we assume our consulting clients already know before they bother talking to us," is a pretty compelling argument.
A few coworkers have even started to take my views seriously on what AI could mean in 5-15 years, to the point of actually comment on how disturbing it was that those things seemed so plausible. One compared what I said to Black Mirror, and seemed to get it when I replied that Black Mirror went as far as it could without seeming too crazy for people to want to watch it, and the reality will be several times stranger, whether in good ways or bad.
Over the years I have also developed a bit of a reputation in my own circles of saying weird things that turn out to be (sometimes usefully) true, which helps. But yeah, this past six months or so has been the first time I feel comfortable just coming out and saying things like, "I think the last generation to die of old age has already been born... one way or the other," or "The last time a new and smarter kind of mind appeared on the planet was the replacement of Homo erectus with Homo sapiens, and that didn't go so well for Homo erectus, or Homo habilis before that."
I mostly don't. My parents and in-laws know it's an interest of mine and know I have some crazy beliefs about it, but we rarely go into detail. Most of my friends I met through the rationalist community and they already know.
How do you think about induced demand now?
https://www.astralcodexten.com/p/change-my-mind-density-increases
See downthread, https://www.astralcodexten.com/p/ask-me-anything-22025/comment/92555827
I may have a small amount of policy making power over AI in the next twelve months. What would you like me to know?
I think you should get in contact with the AI policy establishment if you aren't already. https://www.safe.ai/ , https://encodeai.org/ , and https://www.centeraipolicy.org/ are three good US orgs; if you're in a different country, there are probably local versions.
Otherwise, it depends a lot what kind of position you're in. Feel free to email me with more details, scott@slatestarcodex.com.
How do you rate Irish vs US medical education? My understanding is you studied in Ireland, which I assume is similar to the UK (where I studied), and my impression has always been that US medical university is much more rigorous than in the UK (note - I’m not under the impression that it produces better doctors - telling apart the various types of pulmonary hypertension is often note useful imho)
I never got an apples-to-apples comparison, but my impression is that Ireland was lower-tech, which means they rely on clinical exams and actually knowing what you're doing, whereas US was more likely to be "here's how to call the radiologist and ask what the CT said".
I love your old story A Modern Myth! I feel like it should be made into a movie. That's not a question. My real question is how are Kai and Lyra doing?
How do you think you’ll spend your time in a post ASI world where iq is no longer a valuable commodity?
Sorry if this is too personal, but how do you think about this when raising kids?
Post ASI world is pretty unimaginable, such that I'm not sure it makes sense to talk about IQ being obsolete and holding everything else constant. See https://www.astralcodexten.com/p/book-review-deep-utopia for more.
Maybe I should’ve phrased the question in reference to ai smarter than effectively everyone but not really super intelligent. It seems really short run imaginable for this to be the case for 90% of Americans.
What are your thoughts about CBD as a mild antianxiety medication?
I tried suggesting it to my patients and they almost never got any benefit. It seems like you need some THC to "activate" the CBD, or something. CBD treatments with mild amounts of THC seem to work a little, but then they're not the safe healthy no-addiction-risk miracle drug we wanted.
My favorite SSC blog post of all time is "The Parable of The Talents". I find myself re-reading or thinking about it every few months. Question for Scott: what's an SSC or ACX blog post that you find yourself re-reading or thinking about the most?
(link: https://slatestarcodex.com/2015/01/31/the-parable-of-the-talents/)
Can I get your opinion on this critique of back sleeping recommendation for babies:
https://www.sensible-med.com/p/a-new-series-on-the-back-to-sleep
Also happy for other people with relevant knowledge to comment.
My impression is:
- Yes, babies sleep less peacefully and happily on their backs, and there are various surrogate endpoints that seem worse. Nobody has ever been able to demonstrate any long-term damage, but you could imagine something very subtle.
- Yes, the amount of SIDS you prevent by back-sleeping, while nonzero, is very low. This is one of those epidemiological paradoxes. Suppose back-sleeping lowers your chances of SIDS by 1/3,000 (I think this is OOM right). Then if you're a parent, that seems so low as to hardly matter. But if you're the head of the National Institute of Health, you might think something like "3 million US babies per year, so recommending back-sleeping would save 1,000 lives per year! Think of how much tragedy is prevented!" I think these are both reasonable ways of thinking about things.
Thanks for the reply.
Did you put your twins to sleep on their backs and if/when you have more babies will you put them to sleep on their backs?
In Unsong, would the final interlude have been different if Clinton had won the 2016 election? (And if so, do you have a draft of Interlude: Clinton somewhere, or did you write the whole thing in real time?)
I didn't even think of writing that until after the election was over.
What business or other thing do you most wish existed in your neighborhood?
Jet's Pizza. I got it all the time in Michigan, and it's still one of my favorite things about going to the East Coast, but there are none in CA.
If all goes according to plan I will be a father before 2026. What are the high leverage things I should do before the due date?
Invest in your relationship with the mother, and if your parents live close, with them as well. The first 6 months after birth will be a black hole, but after that normality gradually returns. The stuff about 'having no hobbies anymore' and 'never getting any time for your personal projects again' are all untrue, at least for parents who are together and have some grandparent backup.
A big part of the early exhaustion stems from stress that you don't know what you're doing, but really, really care about getting it right. Depending on your temperament reading a book on the subject might help, or might just make you second guess your instincts harder (you will eventually discover that your instincts were pretty good, and that if you could have just turned your stressometer down a few notches everything would have been better). I found grandparent advice much more useful. Asking advice from parents with kids who are older than yours and have grown up in the way you hope your kids will grow up is a good move as you had past the 6 months mark and move from survival to thrive. There is a huge amount of fad parenting nonesense and I would basically disregard advice from anyone whose kids you can't assess as a yardstick for advice quality.
Get as much stuff done as possible while your time is still yours, haha. Seriously.
Are there any areas, fields, scenes, institutions, movements, etc. in contemporary American life (being a little vague on purpose) about which you’re feeling mostly excited and/or optimistic?
For context: I’ve been feeling a bit bored and/or pessimistic about a lot of different areas of modern American life that usually excite, interest, and motivate me (the arts, writing, the behavioral and social sciences, academia, and politics being the most important). I’d love to find something outside of my personal life to feel excited or optimistic about, so I’d be very interested to hear about anything that’s making you feel excited or hopeful these days.
Do you think reading Rationalist materials/blogs, being versed in the Rationalist lingo etc, materially improves one's life? Do Rationalists make better decisions, live happier lives, make more compelling arguments, invest their money more successfully, find better mates than other folks with similar IQ/backgrounds?
Recently, Dwarkesh Patel made a fascinating point about current AI progress that I've never seen before, and I think holds water: "[T]hese things have basically the entire corpus of human knowledge memorized and they haven't been able to make a single new connection that has led to a discovery[.] Whereas if even a moderately intelligent person had this much stuff memorized, they would notice - Oh, this thing causes this symptom. This other thing also causes this symptom. There's a medical cure right here. Shouldn't we be expecting that kind of stuff?"
Do you think this is a meaningful criticism? Should we expect more discoveries? If not, why not?
See my response downthread: https://www.astralcodexten.com/p/ask-me-anything-22025/comment/92565275
Claude and I co-discovered a novel vision algorithm for a non-euclidean 2D shooter just a few weeks ago. The highest-level abstract idea was mine, but all the details of expressing it in geometrical terms were all the AI's. We needed some help from OpenAI's reasoning models to finish debugging it.
Here's a stream VOD of me working on it - unfortunately I didn't manage to capture the "aha moments" but it should still be illustrative - if I was gonna lie I'd lie about inventing it singlehanded, not lie about using AI. :) https://www.youtube.com/live/utnN6WPf5nQ
Does this serve as proof LLMs can discover things? I think so - if you ask the question correctly, and the solution requires innovation, LLMs are 100% capable of bringing their entire knowledge base to a problem and make new connections to solve it - as always, it's a problem of asking the question correctly so the right knowledge becomes salient.
Has your p(doom) estimate changed much since you posted about it a couple years ago?
On a related note to Mo's question:
Can you tell us any more about the AGI 2 years, ASI 3 years projection, and/or where to look for the article where it will appear?
Any chance of a hardcover edition of Unsong?
How about original-vs-edited editions?
Have you tried OpenAI's deep research? If so, do you find it useful in researching things like the blog posts you write?
P.S: If you haven't, I think you maybe would, and I think you should try it!
Also more generally, how (if at all) do you use LLMs on a daily basis?
You are a very culturally literate (for lack of a better term) writer who frequently makes informed references to all kinds of stuff, from ancient philosophy and classical literature to manga and edgy internet memes. (This is very cool!)
But one thing you seem to very rarely allude to is movies.* This seems very strange to me, as virtually everyone I meet IRL watches movies at least once in a while (even if they're very different movies), to the extent where I would consider the feature film as one of the basic cultural/artistic units of modern society (for lack of a better term).
Are you one of those rare people who says "I pretty much never watch movies"? Do you find that strange given your aforementioned broad cultural literacy? Or is that a false impression? Do you have a list of favorite films or directors?
*This is not a serious criticism by any means, but I remember a post from years ago in which you made a joke whose punchline was basically "Black Panther is going to get a hundred sequels" which fell flat for me because it immediately gave me the vibe of "this guy doesn't know what he's talking about" - a rare experience when reading an SSC post!
Just guessing: Scott likes to read, and doesn't listen to podcasts. Maybe he doesn't like the video format much either.
I'm also like that, like to read many things at my pace, or skim thru them, but will only listen or watch for really amazing stuff.
This variation on the "drowning child:"
You're on your way to prevent multiple statistical deaths by donating 10% of your income in the form of a cashier's check, when you see a drowning child. If you attempt to save the drowning child, you will not be able to prevent multiple statistical deaths. What should you do?
What do you think of fame? Given the amount that you have, is it what you expected it to be? Do you ever wish you could be less famous and still have the same impact? I apologize if this is too personal a line of questioning, but I can't quite get my head around fame, not being famous myself, and I'm curious about any insights you think you might have. Thanks.
In the recent discussions about artistic taste, you didn't talk about an aspect that seems fairly central to me: boredom. I'm married to a choreographer and am friends/family with a bunch of musicians, and have noticed that the thing that seems to drive our taste is mostly "things I haven't seen or heard before". This is quite different from common taste, which seems to mostly be "things I have seen before and like, or things very similar to that". Note also the structure of pop music, which is designed to be essentially one hook that repeats over and over, vs modern classical compositions, which are often impenetrable. This makes me wonder: can we draw a distinction between financial taste (the tastes of funders), expert taste (the refined tastes of practitioners), and public taste (basic things the public likes)? The relative balance of power between these three camps might explain different trends in different artistic areas. I think architecture in particular has fairly powerful practitioners and financiers, and very weak public; whereas music has very powerful public and financiers, and fairly weak practitioners.
Hang on. Am I paying for this?
Hah
I knew what I was feeling, but the words would not come to me. Thank you.
Uh oh, the jig is up.
At these prices, Scott, we expect some truly appalling disclosures.
Such as "he likes/dislikes pineapple on pizza" (delete as applicable to your taste re: pineapple on pizza)?
For someone on the fence who never has before, would you recommend trying psychedelics?
Who do you think is the smartest person of all time?
John von Neuman? But this is only hearsay for me.
This is my answer and why I went through a lot of hassle to record an interview with his daughter!
What do you think of the face starvation autism theory mentioned here? https://slimemoldtimemold.com/2023/09/07/mysteries-contest-winners/
Question on who would be your ideal nominee for a certain job in a hypothetical Trade Offer.
The federal budget in 1969 was $186B, and the debt $353B.
In today's dollars it is $1.6T and $3T.
Today those figures are $6.8T and $35.5T.
I.e. 4.2x and 11.7x larger, even after being adjusted for an inflation that was itself almost completely produced by that budget overspending.
I picked 1969 because that's when USA landed a man on the moon while also conducting an all-out engagement in the Vietnam War. top-tier physics research, and many other things. It was even more "The Envy of the World" back then than it is now.
I understand that Elon is widely disliked. So the Hypothetical Trade Offer is that a wave of the wand and pouf - everyone agrees that Elon is a Nazi, or Unacceptably Autistic, or an Illegal Imigrant, and he gets deported to South Africa where someone can "dubul' ibhunu" him.
Granted. BUT you must nominate someone else - any person of your choice - who will slash the US government spending to less than one quarter over the next 1.5 years.
They may do so in any way they see fit. As much budget as they wish can to go to Saving Drowning Children 10.000 Miles Away. But it will be $1.6T and not a single cent more.
The Trade Offer is hypothetical. But the fact that you can run the envy of the world on less than a quarter of its current budget - without enslaving your children's children to pay off a beyond-ludicrous debt with interest - is not hypothetical in the slighest.
Whom would you nominate to be The Chainsaw Tzar?
Why should the values be fixed total instead of per capita (then you'd need to get spending to 40% of current levels instead of 24%)? And why fixed in real dollars instead of %GDP? If we capped government spending to 1969 levels on a %GDP basis you'd only need to reduce the budget by about 20%, although the debt is still 3x more on that scale because of past decisions.
I would also ask: precisely what level and type of government spending maximizes economic growth while effectively compensating those who get hit by the negative externalities of said growth, and how do we get closer to that world? This is not a figure I expect to stay constant over time.
The US only had 60% of its current population in 1969, but you want the same amount of tax revenue to serve a far greater number of people. That doesn't seem fair.
Healthcare spending alone is more than your $1.6T target, and so would have to be massively cut. Life expectancy increased by about 9 years since 1969 despite rising obesity rates. The improvement is partially from better medical technology for treating cancer and heart disease, but those treatments are expensive and we would lose much of the benefit of these new technologies if we didn't actually pay for them.
USA is ~50th in life expectancy.
Life expectancy had been steadily rising in dozens upon dozens of countries that didn't experience anything *near* the stellar explosion of US healthcare costs.
It was caused worldwide by general technological advances which go way beyond cancer and heart disease treatment. In 1969 diagnostic ultrasound was in its infancy; today you get amazing precision and insight out of a ~$50k machine. It's beyond compare. The list of technologies goes on, and they've all grown significantly *cheaper* and more powerful, not more expensive.
The US healthcare cost explosion has nothing to do with the cost of technology nor the cost of actually producing the medications. (cancer and heart disease included)
It has to do with the government keeping on throwing absolutely any amount of taxpayer money on the industry that the industry wants to ask for, while at the same time imposing insane byzantine levels of regulation which ensure that no true competition can exist among the few who have been selected for successfully passing that regulation, not selected for the healthcare they provide. Rinse and repeat for decades, and what do you expect?
"Serve a far greater number of people"? You probably mean welfare, right? Because right now it has to serve ~4.2 million miles of roads, which was ~3.5 million miles in 1969. Not quite a 4x difference. Meanwhile, the people who are on welfare now and wouldn't have been in 1969, and the money they draw on average compared to then, those are quite overdue for a visit by The Chainsaw Tzar.
I largely agree on the healthcare costs, but in addition the the factors you mention, we also just get more healthcare on average than we did in 1969, and this does improve health outcomes. We can be less efficient in healthcare but still have an absolute improvement.
> "Serve a far greater number of people"? You probably mean welfare, right?
I mean almost everything, including healthcare and welfare. Even the roads, in addition to the extra miles, they also get more traffic, which increases maintenance costs. You wouldn't expect a country with 1 million people to need the same expenditures as a country with 10 million people. Some costs scale sublinearly, others superlinearly, but overall, you'd expect costs to be roughly proportional to the population.
> But the fact that you can run the envy of the world on less than a quarter of its current budget - without enslaving your children's children to pay off a beyond-ludicrous debt with interest - is not hypothetical in the slighest.
Given that the children of the children of 1969 America generally consider themselves to be enslaved in the name of paying off beyond-ludicrous debt with interest, and that a not-insignificant portion of the $35T you cite *is* interest, I think you have accidentally provided a counterexample to your own claim.
It's an example, not a counterexample.
Interest only grows exponentially. Whatever it is exactly for the children of the children of the 1969 America, it will be exponentially much more for the children of their children.
Financiers don't care a whiff WHAT the debt is for. "Trillions for reparations for slavery"? Yea, whatever dude, good one! Doesn't matter, just keep it rolling! As long as we can even further astronomically inflate the interest payments we're raking in, by even further enslaving the populace today, it's still less than how we'll have enslaved their children's children.
> Given that the children of the children of 1969 America generally consider themselves to be enslaved in the name of paying off beyond-ludicrous debt with interest
I’m not American but, really? The people born in 1969 are doing pretty good. Post 1990 not so much.
That's what I mean. 1969 America was the Baby Boomers, who are doing great. Their children are Gen.X, whose children are Gen.Z, who are not doing so well.
Note: Real GDP has also grown by 4.5x, so the budget is actually *less* as a % of GDP.
GDPs are accounting voodoo and fiction, as far as actual people are concerned.
Budget dollars are very real.
Debt is very real.
There’s some issues with how GDP is calculated but it’s the income of the country (not really wealth as is sometimes said). Therefore it’s useful to compare it to the debt of the country.
Can you try to explain/analyze, step by step, what happens in your head when you're writing jokes and coming up with creative ideas? Your posts have a lot of jokes, and some posts (like Bay Area House Parry) are made out of a series of brilliant/creative/absurd ideas, so clearly you can come up with stuff like that consistently and intentionally (as opposed to ideas "just coming to you", outside of your control). How do you do this? What happens in your head when you invent one of Bay Area House Parry startup ideas? I can't think of ideas like that, is there anything I can do to learn?
Non-authoritative answer in case Scott doesn't get around to it: I believe it is known that there are several .txt files on SA's desktop that slowly fill up over months or years with subject-related quips and curiosities (monthly links posts are a microcosm of this), so in a sense the ideas *do* just come to him out of his control.
It is similar to the common advice that if you want to be a writer, you should carry a pocket notepad and pen everywhere you go, which similarly helps random shower thoughts appear intentional when expanded upon and published.
Also, lots of practice in rhetorical role-play as a youth.
Is the methodology that social scientists have used to observe that the relationship between IQ and success in other areas of one’s life is causal rather than a correlation with other lucky factors like wealth and race robust? If so why?
I haven't looked into this super deeply, but I think so, for a few reasons:
1. It's definitely not just race, because people have looked into this within races and found the same effect.
2. Wealth is harder to disentangle. You can control for parental wealth, and when you do this I think the effect continues, but controlling for things is hard and you'll always have some residual error. I think the size of the effect is large enough that it's implausible that it's just residual error, but this is an opinion rather than a proof.
3. You can use twin studies and adoption studies to prove that the relationship between *genes* and income is at least partly causal, and probably some of those genes are for IQ, but this is just a guess - maybe it's genes for attractiveness or something.
4. You can probably do fancy statistics to show that the genetic effect is a pathway through IQ, but I don't remember any specific paper that has done this and fancy statistics are always suspect.
5. Actually, wait, the obvious right way to do this would be sibling pairs, where they both come from the same family but one of them lucked into a higher IQ. The only paper I can find on this is by Charles Murray ( https://sci-hub.st/https://www.jstor.org/stable/3083428 ) and he unsurprisingly finds it's causal.
Thanks! Can you link to the studies you like that find it isn’t race?
Are you going to continue to add more essays on Lorien Psychiatry about psychiatric conditions/medications?
I'm interested in the habits of successful writers. Do you have a time of day when you prefer to write? Do you write better when your schedule is empty or relatively full? Do you wait for inspiration to strike or do you block out time for writing, on a schedule?
No need to answer all of these — just curious to hear about your experience.
Also, Scott, any plans to continue the Dictator Book Club?
I'm running out of people who fit my criteria: modern-day figures who transitioned a somewhat democratic regime to a more dictatorial one and have good books about them. If you think of more, let me know!
(no, I'm not doing Trump - at least not until the dust settles and we know what his record looks like)
I actually think you could write an interesting one on Franco if you wanted to, in that he took power via an attempted coup that turned into a civil war, but it was an attempted coup in a fairly democratic regime in really, really bad shape, and pretty clearly if the country had not been a giant mess he wouldn't have pulled the coup. But that would be stretching it a lot.
There'll presumably be good Bukele books in there near future, if there aren't already. It's not focused on a specific person, but "South Africa’s Brave New World: The Beloved Country Since The End Of Apartheid," by R.W. Johnson is similar.
(Review by John Psmith: https://www.thepsmiths.com/p/review-south-africas-brave-new-world)
That reminds me, I think I have the full list here (let me know if I've missed any):
1. Erdogan in Turkey: https://astralcodexten.substack.com/p/book-review-the-new-sultan
2. Modi in India: https://www.astralcodexten.com/p/book-review-modi-a-political-biography
3. Orban in Hungary: https://www.astralcodexten.com/p/dictator-book-club-orban
4. Xi in China: https://astralcodexten.substack.com/p/dictator-book-club-xi-jinping
5. Putin in Russia: https://www.astralcodexten.com/publish/post/134180409
6. Chavez in Venezuela: https://www.astralcodexten.com/p/dictator-book-club-chavez (arguably the most Trump-like?)
7. Hitler in Germany: https://www.astralcodexten.com/p/your-book-review-the-rise-and-fall (not by Scott Alexander)
I enjoy that you say Chavez is the most Trump-like. I imagine he's the one of those that trump would least like to be compared to
I think you're right though
Thanks. It's nice to have these in one place.
What nootropics do you use or recommend using?
+1 for this -- there's a bunch of older posts about this but I'd be curious what your current thoughts are!
Scott, some time ago, I remember being able to access your blog roll (as in: people you follow and recommend) on astralcodexten.com main page. It seems to be gone (there are "recommendations" but I vaguely remember the blog roll to have been much longer and not just substacks). Where can it be found now?
Nowhere. I took it down because most of the people involved had moved to Substack and I could replace them with normal Substack recommendations (which you can still find somewhere on the blog front page). The others seemed so well-known (eg Marginal Revolution, Less Wrong) that it wasn't worth it. Sorry!
Why’d you self-publish Unsong instead of traditionally publishing (as you planned originally iirc)?
One of the people who wanted to traditional-publish wanted massive edits from a formal editor, and I didn't have the time/energy to do that. Another dropped off the face of the earth. Eventually I realized I was delaying and would never do it. At the same time, ACX user Pycea offered to help me self-publish and I realized I either took his offer or put it off forever and never published anything.
Do you read sci-fi or fantasy for pleasure? If so, any favorite authors?
What article from 2024 is your favorite (or are you the proudest)? How about least favorite (or least proud)?
What was the kabbalistic significance of renaming Erica in the print version of Unsong?
There was a convoluted correspondence between Unsong and William Blake's mythology. Erica was supposed to correspond to one of the Blake characters - I originally thought it was Orc, but later realized it should have been Vala.
That particular change didn't affect much (though headcanon will probably forever be Erica, "Val" reference pointer was already reserved by Valencia the Red from WtC), but I did spend hundreds of pages in anticipation of how the "Miss America" joke would get semantically repurposed. Ah well. Good to know the background reason. This is why it's good for the original web version to stay up!
Do mental hospitals tend to have a rule against employing people formerly confined in one? (Maybe only for those formerly involuntarily confined, maybe only for patient contact positions?)
This came up when I was a patient in one long ago - people were chatting and someone mentioned maybe wanting to come back and work here, and we got to talking about whether that was a good idea and how "lunatics running the asylum" was a well-known adage, considered a situation to be avoided.
I don't think the HR office even has access to confinement records.
How often do you use AI in your practice? When do you think “psychiatrist” will be automated in practice, even if you still act as a gatekeeper for drugs?
I sometimes ask it what it would do in a case, to see if there's something I'm missing. Occasionally I use it as a glorified search engine, like "remind me which diseases have this collection of symptoms".
How interesting or compelling do you find the works of Christopher Alexander?
What 19th century writers and thinkers do you like?
You seem like the kind of person who plays dungeons and dragons. do you? what kind of game?
See dungeons and discourse on the old blog
Not really! The last time I played was to join in some friends' heavily edited homebrew version of Dragon Heist just before COVID. We wanted to do something interesting, so all of our characters were a gang of nine year old orphans. Whenever we had to interact with other people, one person with a shapechanging spell would pretend to be an adult. It was pretty funny.
Unique item—Totem Coat of the Kobold:
This coat acts as a mundane trenchcoat while equipped by a single person, except that it resizes to fit any wearer. While equipped, the wearer may designate an ally in line of sight. When they do, their own body and anything they have equipped other than the coat become ethereal to that ally (and their equipment), such that that ally alone may pass through their body for the purpose of entering the coat and equipping it themselves. Upon fully equipping the coat, an extradimensional space materializes around the head and upper body of the prior wearer and they are blinded for the duration of the effect (sound and other senses carry through in both directions), and a separate extradimensional space materializes around the body of the new equipper below the waist. The waists of the new and previous equippers are joined.
The effect can be repeated an unlimited number of times, with each new equipper causing their lower body and the previous equipper's upper body to be transported into their respective extradimensional spaces upon equipping the coat (their waists remain joined with all other wearers). Only the most recent equipper may designate the next ally (as they are the only one able to see the material plane and affect it with their upper body).
When each wearer beyond the first equips the coat, they may will it to resize to any length from the sum total of the first equipper's height to the waist and the length of the most recent equipper's upper body, to the sum total of the first equipper's waist height and all equippers' upper bodies (excluding the head). The head, chest, and arms of the latest equipper are elevated to the corresponding appropriate height. The intermediate space between chest and legs is hidden by the closed coat, but acts as an elongation of the last equipper's torso. This and other effects of the coat are physical, not illusory. See Evard's Guide to Cosmic Horror for the effects of peeking beneath the coat or inspecting its wearers with extrasensory perception.
If the coat is opened for any reason other than a designated ally equipping or unequipping it (a full round action per wearer, in reverse order), all wearers are simultaneously expelled from their extradimensional spaces, with exits projecting vertically in order of equipping (it remains equipped to the last wearer and returns to behaving as a mundane coat). Each former wearer may attempt to grapple either or both adjacent former wearers as a swift action before any external forces take effect. Any wearer of the coat may also activate this effect at will as a free action if the coat is not held closed.
The coat may be held closed manually or with a rope or other fastener, but any fasteners attached to the coat in any way detach and fall off the coat whenever it becomes equipped or unequipped by any wearer (they may be reattached as normal immediately after falling off). As a standard action, any wearer of the coat may attempt an Escape Artist check against the method(s) of closure (regardless of their ability to physically interact) to bypass them and open the coat. Any wearer(s) of the coat may also oppose any attempt to open it (by wearers or otherwise) as a swift action, using an appropriate combat maneuver. (Any of these or other actions taken by any but the last to equip the coat suffer appropriate penalties for their blindness.)
Other effects may cause the wearers to be similarly expelled, including:
- Entering, leaving, or taking any physical action in rough terrain.
- Entering, leaving, or taking any action within the threat zone of any ally, enemy, or other creature as small as one size category below the smaller of the first or last wearer, and as large as one size category above the larger wearer.
- Beginning a round within a swarm or crowd.
- Being hit by a physical attack or combat maneuver.
- Taking fall damage.
- Performing any physical action that requires a skill check with DC 15 or greater.
- Jumping any height or distance.
All such conditions may be negated if all wearers individually succeed at a DC 10 Acrobatics or Athletics check (each wearer chooses for themselves) once per round on the turn of the first equipper. All clasps apply their bypass DC minus 10 as a bonus to all checks, but wearers may not manually hold the coat closed. Any other ally in melee range may grant advantage to all wearers as a swift action, and any effect which could break concentration grants disadvantage. These checks automatically fail if any wearer rolls a 1, and critically fail if more than one 1 is rolled.
The coat has no armor benefit, but has hardness equivalent to +1 hardened leather. If the coat takes 5 or more points of slashing damage (after DR) in one hit, its wearers are expelled regardless of clasps.
The coat gives +20 to Deception checks, made by the last wearer, to be perceived as any person of appropriate age for their apparent height. In addition to the initial check, the latest wearer must make a check once per action taken or line of dialogue delivered in the presence of a target if table talk between players exceeds 5 seconds, with an added penalty of –5 per equipped character whose player engages in table talk (excepting pantomime or shushing other players) and –10 per 10 seconds of table talk (including pantomime), to a minimum of –40 in penalties (–20 total).
Any player shouting during table talk (including to shush) grants disadvantage for any and all subsequent Deception checks (for any purpose) against the same target by the last equipper, whether or not the coat is equipped. If the coat is opened, all characters formerly wearing it suffer disadvantage to all skill checks involving social actions against anyone observing the interaction.
Which book reviews did you vote for in each year?
My pet topic is personality research, like Big 5 or Myers-Briggs. I think the zeitgeist has gone too far towards "it's not scientific/useless" when the data suggests otherwise. What do you think?
I originally agreed with you - see https://slatestarcodex.com/2014/05/27/on-types-of-typologies/ - but I think Spencer Greenberg has done a good job quantifying exactly how good each type of test is - see https://www.clearerthinking.org/post/how-accurate-are-popular-personality-test-frameworks-at-predicting-life-outcomes-a-detailed-investi
The apples to apples regression comparison ("4 scores") of MBTI vs Big 5 in that post shows a small (.15 vs .18) difference that is not indicated to be statistically significant. As in, the advantage of Big 5 over mbti is that it is Big 5 not Big 4 and that it uses a continuous score -- coincidentally the exact same reasons no one outside of academia uses it for anything.
This is exactly right. The binary and grouped nature of mbti is what makes it useful in day to day life.
https://www.vectorsofmind.com/p/the-big-five-are-word-vectors if you haven't seen it you'll find it interesting!
Would you ever let someone turn ACX into an essay collection-type book, as has been done with SSC?
I know you did a retrospective on intellectual progress you made in the 2010s, and you've also searched for previous work you've done toward things you only later explicated. Do you see a connection between your essays "extremism in thought experiment is no vice," "fear and trembling at EA Global," and "the tails coming apart as a metaphor for life?"
I did this for all the creative writing done on ACX (plus some other random old SSC essays that I loved and hadn't been published in other books I have, and a couple Eliezer fiction pieces that I wanted). You can find it here: https://www.lesswrong.com/posts/foKpTq3WQ7w5s9rZE/printable-book-of-some-rationalist-creative-writing-from
Thanks. I'll check it out.
Yeah, anyone who wants to can do this.
Thank you. The omnibus ebook of SSC was a massive help to me when I was in a yeshiva without regular Internet access, and I still slightly prefer a format that I can read on my Kindle.
Noticed that ACX has been relatively quiet about artificial intelligence lately - is this a deliberate step back, or has it just seemed to you that there hasn't been much to say about it? In the past few months I've had many moments of "I wonder what Scott Alexander thinks of all this", but the only time you've commented on the subject recently was the alignment faking thing (IIRC).
Or another way of asking the same question: What happened to Machine Alignment Monday?!?!
I don't think it's been quiet. I'm helping with a different project on artificial intelligence and will post about it when it's done.
I found that people were less likely to read things with the Machine Intelligence Monday "branding" compared to regular posts, so I turned the deeper essays into regular posts. I think I was also sort of hoping to have a shallow series on the latest month's developments in AI whatever they were, but now there are dozens of blogs that do exactly that much better than I could, and Zvi even does it for alignment in particular.
Forehead-smack. I'd forgotten about the Machine Intelligence Monday "branding" when I was offering you the AGI Friday "branding". Thanks again, btw, for the encouragement. I've got the Substack set up and planning the first real post for this Friday!
How was Unsong received by religious Jews?
Surprisingly well with the ones who contacted me, but that's obviously a biased sample. I think local commenter Isha Yiras Hashem was disappointed in it.
Is it possible to submit a question for the survey? I’d like to ask a question seeing whether people experience visual snow. I think the few studies that have surveyed it have either undercounted it to a large degree or missed its prevalence in populations the ACX community is enriched for. This is at least anecdotally true in a survey of ~100 people I happen to have brought it up in conversation with at various times.
People with mild visual snow are likely to think it’s just the standard human experience (like I did). I think it’s one of those things like post-toilet wiping stance that acts kind of like a scissor statement but for human experience.
Woah, I also have mild visual snow and assumed it was universal! (Didn't even know there was a name for it until I saw your comment and googled.) My partner has it too. We recently had a conversation trying to compare how 'big' the typical grain size of our snow was.
I also have tinnitus, which the wiki page says might be related, but I've had visual snow all my life and tinnitus only since getting COVID in 2020 so idk
Always happy to let someone in on it :)
See https://slatestarcodex.com/2019/09/11/lots-of-people-going-around-with-mild-hallucinations-all-the-time/ . If the post itself doesn't answer your questions, I think there's more data on the 2019 survey ( https://slatestarcodex.com/2019/01/13/ssc-survey-results-2019/ )
This is great to see (and as always a sign that the SSC archives contain many unread riches that I haven't gotten to yet)! However, the post and survey don't really answer my question. Would you consider further exploring the topic? The survey only prompts people for persistent visual changes as a result of hallucinogen use, without prompting specifically for visual snow or similar phenomena. This is important because people who have it are unlikely to recognize it as a nonstandard experience and because it's asking for changes rather than base experience.
I am curious about the additional visual hallucinations you bring up, though - I tend to suspect that there are a lot of common but not-commonly-known minor perceptual disorders (much as there are many other common minor disorders in things that are less interesting).
I'm most of the way through the physical version of Unsong, and I've re-read the original enough to notice edits as I go. What changes most stick in your head, and what went into them?
A few seem obvious (CHANGELOG appears to be gone, which is understandable even if it makes me very sad); a few seem reasonable but unnecessary (e.g. the joke about Christianity being saved by failure to pronounce a four-letter word got replaced by one about Ohio becoming un-speakable on account of its phonetic resemblance to the Mortal Name -- which is also funny, but I"m not sure why it was a *replacement*); and a few others just mystify me (name/gender changes, mostly; kabbalistic references I'm not catching?). And some fairly significant changes to "The Image of Eternal Death" and Uriel's peace summit disaster, which I didn't much care for but I think I see the rationale in each case.
(I keep my own fiction's edit history with accompanying reasoning in Git. You can tell.)
Also, bug report(s): I spotted at least one continuity error introduced by the edits. Aaron no longer makes telepathic contact with Ana before escaping Unsong HQ, but she still later reacts as if he did. Pretty sure that was a case of intentionally editing one scene and not editing the other to match.
Biggest changes:
- Replaced Malia Ngo with Asher Bentham character. Malia was a relic of a very different earlier version of the book and never really fit in; I think the chapter focusing on Asher's childhood makes him more of a part of the story.
- Changed Erica's name to Valerie; this *was* kabbalistic, she was originally supposed to correspond to William Blake's character "Orc" but I was misreading Blake and the relevant character is actually Vala.
- Made a few of the diplomatic negotiation chapters better (foundation of Untied States, Israel/Palestine) because I kind of phoned those in originally and thought there was room for improvement.
- Took out some things that were ephemeral or related to the news when I was writing the story (eg Changelog).
I was disappointed that the Spiders Georg joke was removed. :P
Can you help me understand, in an IQ 110 or lower way, why the idea of us being in a simulation has gotten so much traction?
I’ve read the arguments and they all seem to be based on an underlying assumption that there’s no way we could know because we’d be programmed that way, ergo relatively high probability. This seems like the logic my stoner friends used in college to “prove” any number of ill-conceived theories but it’s also possible I just don’t have the intellectual horsepower to understand, or am missing a critical argument element.
I thought the argument was that most people will be in simulations, because each real world will have many simulations in it.
I agree there's not much you can do with this once you've thought of it.
To expand, it's supposed to be a trilemma: simulations are cheap and so most observer moments are simulated, we die and never do simulations, so all observer moments are real, and simulations are really really expensive or impossible, in which case most observer moments are real.
Do you listen to pop (non classical) music? And if so, could you give a spicy take that you truly believe, but think most of your readers would consider blasphemous?
eg Foo Fighters are better than Nirvana
Foo Fighters are better than Nirvana tho
Not really, sorry.
Do you believe that humans have a soul?
What writers/thinkers/schools of thought do you think the ACX/rat community put too little value on?
Do you expect that an AI future is utopian or dystopian. And what probabilities would you assign here.
Not a question, just a thank you. You've helped educate me bit by bit over the years through your work, and that's helped me in small but meaningful ways countless times.
Your fiction work finally got me off my ass to try and write something for self-publishing, in the tabletop gaming realm. So, thanks for that, too. Keep making those fiction pieces!
You're welcome!
Any thoughts on AI in developing countries? People often talk about AI increasing economic productivity by a lot in the developed world, but if this is the case, can't the same AI be deployed in developing countries, where labor productivity is much lower? How likely do you feel this is to happen?
I was just accepted to medical school and so will have to think about what field of medicine to enter soon. What made you choose psychiatry? Any predictions about the effects of AI on the physician job market in 2035, or should I just not worry about that at all?
Can you talk about your experience, preferably first hand but second hand is fine, with gen z in therapy?
Zoomers have both much less stigma against seeking help but also claim to have a skyrocketing number of mental health problems.
Bonus question: some surveys (of dubious quality) point towards therapist being a desirable job for the youngins. Should this pan out, what would you expect the quality of the marginal therapists to be? Would the increased availability of therapy be on net positive?
Do you have a bottom-up or top-down approach to your writing? i.e., Do you have a running list of blog post ideas that you then try to confirm through content you find online, or are you reading papers and links on your own, with your writing ideas emerging along the way?
As a fellow rationalist aspiring to hone their craft, what resources and mental models do you find yourself getting the most leverage from when wrangling with a new sociological topic or issue?
When it some to national politics or policy, I find myself mostly unable to form helpful viewpoints and would like some inroads into improving.
Around a year ago several of my friends (who are women) told me some variation of "my therapist tells me adhd in women in severely underdiagnosed" and then implied that they should get tested. Hearing it from 4 unrelated, high functioning ~30 year old women made me wonder if some new research dropped or if this was a meme floating about in psychiatry. Thoughts?
Best scifi novel and series?
Best fantasy novel and series?
Or the ones you can’t not mention in a cluster of three or four?
What do you think about the idea of someone dumping a bunch of your posts into an LLM to train it to talk like you? What do you feel like it would be missing most if it only had your writing to go by? Should this be illegal without your explicit consent?
Would you ever do a socialist/non-socialist FAQ, like you've done for libertarianism and NRX? Otherwise, do you plan on doing one for any other ideologies or plan on engaging with leftist thought in a less structured way again like your reviews of Singer on Marx or Red Plenty? Basically, I'd like to know your overall view on socialism since it's adjacent to a lot of the topics this blog normally covers and has a lot of strong arguments imo, but is relatively under-discussed here, but I'd still expect you to have a nuanced take on it.
I'd also be strongly interested in an update to the Non-Libertarian FAQ. The last version posted* does open with a disclaimer that "it no longer completely reflects my current views", but it seems like a bit more than a change in emphasis must have occurred for Scott to now be actively identifying with the label.
* 2017 on SSC? I feel like there was a more recent one on ACX, but not finding it at the moment.
Maybe too personal but do you still publicly identify as asexual?
What would you do if you were 10x more agentic?
Two things I haven’t seen you write about that I’d be interested in your take on:
1. How you maintain and improve your relationship with your wife over time.
2. What tools and practices do you use regarding personal finance (also how you and your wife deal with shared finances).
Do you feel constant or regular stress, feeling it in your stomach ?
With your public exposure, with all the internet scrolling, with these huge articles you publish, I am very impressed at the level of energy that you deploy seemingly so naturally.
The topic you deals with, your readers with their high expectations : how do you manage to not freeze when choosing each word ?
Do you fight to escape the paralyzing thought circles and the physical vertigo ? Or is it natural ?
And is it easier with time or harder ?
Do you have any advice for current students, or opinions about which skills should be prioritised? I'll soon be graduating in CompSci, but given the rate of AI progress I'm no longer sure what high-impact routes would look like. I was planning to go into (technical) AI Safety, but now I'm doubting I'll have time to sufficiently upskill before it matters
If a god is good and creates all possible net-positive universes, why:
- doesn't it constantly replace suffering with something that can have the same outward behavior but no experience of it? E.g., if at every single moment, most future branches are those where bad things just stop, people mostly continue in worlds where there is no evil.
- doesn't it skip simulating worlds altogether and just simulate all possible minds with good experiences, skipping over bad experiences (inserting memories of having all possible bad experiences, though skipping over the experience of living through the memories if it's bad)?
Do you have any evidence He doesn't?
What bizarre psychiatry/neurology case studies have you heard about lately?
This isn't bizarre, but there is going to be an *amazing* article on misophonia in Asterisk in a few months (not by me). Also a mediocre article on people with dwarf-tunnel-torture hallucination/conspiracy theories (by me).
Appears to be out now, unless there's another planned: https://asteriskmag.com/issues/09/the-unbearable-loudness-of-chewing
Oh! You're right! Thanks! In that case hopefully I can get around to writing an ACX post about it soon.
If you could choose some n>0 where you die n years from now, which n would you pick? No takebacks, once you choose n you're locked in.
Assume that you stop aging and that you stay mentally at the level you're at now — this isn't a question of how much Alzheimer's you're willing to tolerate. We'll cash out the immortality by saying that anything which would normally kill you just contingently doesn't happen — you don't get hit by any buses, your telomeres just don't degrade, you have time to build / acquire a space suit before entering a vacuum, and so on. Up to you whether an uploaded version of you counts as "you".
How is Lorien Psychiatry going? Any chance we’ll get a post on your learnings there?
Answer to the second question at https://www.astralcodexten.com/p/from-the-mailbag is still pretty much what I've got.
Recommendations for books about therapy?
Now that you've successfully had children, what would you change about your preparation with the benefit of hindsight?
A related question: how have your "theories of child rearing" changed?
Have you written anywhere about your process? Still find it hard to believe that Scott Alexander is only one person. (this is a compliment)
Do you have people helping you with research etc?
There's this classic: https://slatestarcodex.com/2016/02/20/writing-advice/
It's the one that coined "concept handle".
Thanks, that was helpful! Contains a lot of things that you frequently see in Scotts writing, but awesome to have it all written out explicitly in one place.
(not sure how much the trump examples hold up - but the overall point stands)
I don't intuitively feel like I have a process, so this makes things hard. Short posts come out with no thought. Long posts are painful but it just seems like thinking hard about them and dealing with the pain, not some kind of fancy outlining process.
That makes sense
A question from my non-subscriber wife, but hopefully you'll allow it: What was the highlight of your week? (Let's say last week since it's Monday right now)
After 8 years of uni, I will finally finish my CS/ML master a few days, but I have very little idea of what to do next (in the short term too, but primarily in the mid- to long-term). What advice do you have for figuring out one's path in life? And/or what advice would you give your younger self?
Strongly recommend the "80.000 hours" site and the resources there. Wish I had them when I was younger.
Hmmm...everyone told me to network, I hated this advice, but it's not bad. I think there's something where getting into "the elite" is more important than whatever object-level thing you're doing. This doesn't necessarily mean going to cocktail parties or something - for a CS master, it could mean knowing enough people in cool startups that you feel like the sort of person who could easily start a startup yourself (and would be able to easily navigate subtasks like talking to VCs). Even if you don't go the startup route, I think a lot of people in Silicon Valley are kind of generalists where they feel comfortable applying to jobs outside their skills-comfort-zone and trusting that their intelligence and work ethic will make it work, and I think this requires some sort of confidence/networking/etc. Being in a position where you feel like you really understand a field/industry and can do whatever thing vaguely related to it seems most interesting to you is a lot more fun than being in a position where you're resume-grinding and trying to get into the three companies that are an exact match for your skills. You rarely start at the former point but it's a useful goal to try to get to.
Also I think "network" is not a framing designed to appeal. The core components are "other people know who you are" and "you know who other people are." Who in your field is worth working with? What skills and qualities do you have, such that when other people are looking to hire someone, they think of you? How did that information flow, from the world to you and from you to the world?
I think a lot of people want the facts to 'speak for themselves' but being deliberate about communication is extremely useful!
Apologies if you’ve discussed this elsewhere, but I’d be interested in your current AI usage. Which model(s) do you prefer? In what areas do you find them most useful? Any custom prompts or unusual use cases?
I mostly just use Claude, partly out of corporate loyalty to Anthropic (most of the safety people ended up there), partly because I genuinely like it. I don't do anything fancy, mostly just ask it questions about topics I don't understand while I'm trying to write posts. My wife has a GPT subscription which I use sometimes to get images.
related: would love to know what your complains are about current models/what you are looking forward to
Oooo! You're holding a quodlibet! My medievalist side is jumping up and down like a kid in a candy store.
Here are my favorite quodlibet questions from the Middle Ages. Of course I can't remember who was asked these when, except that the last question was asked to Thomas Aquinas. Some of these (like the second and third?) may have medieval Jewish roots or counterparts.
1. If a baby is born with two heads, should it be baptized twice?
2. In the perfect original paradise in the garden of Eden, did Adam and Eve have to poop?
3. Less scatalogically: If Adam and Eve hadn't fallen, would there always have been equal numbers of men and women in Paradise? Because, obviously, if the numbers had been unequal, someone would never have found a perfect match.
4. Which is strongest: truth, the king, wine, or a woman?
>1. If a baby is born with two heads, should it be baptized twice?
What's the theory of intentionality wrt baptism? If, for instance, the formula was said in the course of immersion baptism, but the formula was "internally directed" at one head, is it valid for the other in the case of the two being two persons? What if there are identical twins the formula is "directed at" twin A when it is mistakenly twin B being immersed?
I think the most extreme case of this would be a form of monozygotic chimerism. Not even sure there's a word for that given it's likely unidentifiable, despite necessarily existing. You'd effectively have an identical twin absorbed into you in an invisible way, yet its seems to be a difference of degree rather than kind with monozygotic Siamese twins.
1. You sprinkle water on one head, then the other, so no single head gets baptized twice.
2. Obviously, God told them to eat of all trees except the forbidden one, so they were definitely eating - so where would the food have gone otherwise? However, there were no bacteria, so it was all meconium.
3. There would have been more men, because (given that we're assuming Catholicism is true) one of them would have to be Pope.
4. It's a rock paper scissors loop. Wine clouds truth, woman drinks wine, king oppresses woman, truth brings down king. This is just obvious good game design.
Whoa there! If Adam and Eve never fell, why did Christ incarnate as a human, die, rise again, and found the Church, of which the Pope is the head?
This ignores the deep roots of the Papacy - a close reading of the Bible will find that Melchizedek was Pope two thousand years before Christ.
Personally I like the theory that Adam and Eve don't poop they just keep getting bigger and bigger over time
Approximately what level of financial grant would be necessary to incentivize you to have one more child than you would otherwise want? (I very much assume your wife would have a say, so the amount would need to be convincing to both of you!) Presume either a one-time grant at birth or distributed over the first 18 years of the child's life whichever makes the question cleaner, but pure money without any additional external assistance.
Bonus round: instead of having an additional child you would raise yourselves, a magically-zero-inconvenience medical procedure results in (genetically) the same child but born to an unrelated mother somewhere else in the United States - the child would be given your contact info when they turned 18, but no other link. What level of grant would be necessary, or would you even pay for the opportunity?
I am not Scott, but I want to take this chance to go on record saying I'm willing to pay five figures for the existence of genetic offspring in a scenario you describe (assuming the parents are healthy and likely to take good care of the resulting children), just in case this isn't purely theoretical.
Probably $300K, and depends how sure I am that the unrelated mother would be a good parent.
Are the pro-growth vibes getting better in SF?
I don't know. My impression was that the last mayor was slightly pro-growth but got thwarted by the supervisors. The new mayor is an unknown quantity but seems pro-growth, and the new supervisors are light-years better. But I live in Oakland and rarely leave my house, so this is all hearsay.
If you had to take a bet one one IQ improvement intervention for adults what would it be?
What are your 3 favorite podcasts, substacks, and YouTube channels?
In the second age of Trump it appears that at least some of what were claimed to be Right Wing Conspiracies (eg where/how USAID money goes, the Deep State as an extension of the Democratic Party) are legit, while claims by the Democrats (eg the Steele Dossier) look like Left Wing Conspiracies.
I find myself every day scanning Twitter and seeing a whole lot stories that seem crazy (but also legitimate in the sense of all hanging together along with such associated data as I can find) while I scan through what the MSM has to say and it looks like Pravda-level fantasy; not exactly made-up facts so much as deliberate distortions, ignoring important details, focussing on the unimportant or trivial, and so on.
It feels somewhat like a rerun of https://slatestarcodex.com/2019/06/03/repost-epistemic-learned-helplessness/
Do you have any mechanisms of heuristics for how to make truth-sense of the world in this new environment? "I choose to just accept the ideas of the prestigious people with professorships in Archaeology" was a reasonable hypothesis until we found that the professorships were all being funded by the CIA and USAID! (Not actually, not yet anyway; but certainly in Social Sciences and History I feel we are at that point. I simply don't care that Harvard Professor X claimed something about racism or trans or whatever; I no longer believe that Professor X's credentials have very much to do with what s/he's claiming is likely true.)
Interesting, my impression is that most of the things the right wing has said about USAID money has been proven false. Maybe we're thinking of different claims, but the average Twitter claim has had a half-life of a few hours before someone has proven they're making some ridiculous mistake in reading a database or something.
I guess this is my advice - follow enough people on Twitter that you frequently see debates between people claiming that some of the things one side says are false, then follow the debates, see who wins, and downgrade accordingly.
If social referrals were not an option, what is the best way to find a therapist?
You could also ask your regular doctor for a recommendation. They likely have a list of good mental health referrals, and may have gotten a feel for which places they refer to actually work out and which ones have their client back asking for a new referral. I work in an outpatient therapy clinic and we get a lot of referrals from medical providers because we've built of a reputation for having availability and doing a good job.
There is no good option, but psychologytoday.com at least exists and (eventually) works.
What new or existing governance structures are likely to work best for long term spacefaring colonization efforts?
My partner loves Jordan Peterson, and I'd like to introduce him to additional thinkers he might also enjoy. He's right-leaning, prefers video and podcasts to blogs, and has a contrarian streak. Do you have any content creators you'd recommend?
Not Scott, but this blog, Hanania, Bentham's Bulldog, Tylwer Cowen all come to mind
I'm working on a crowdfunding platform (link below) to enable systematic chemical testing of consumer products (initially focused on plasticizers and phthalates, inspired by PlasticList).
Given your interest in aggregating distributed knowledge and your writing on FDA approval processes, I'm curious:
Do you see potential in using crowdfunding models to generate publicly-available safety data? Or would such systems inevitably select for sensationalist concerns over legitimate risks?
https://laboratory.love/
Not Scott but this is a great idea!
Do you ascribe to Richard Hannah's idea we'll have millions of make beleive jobs or are we the 1900's donkey's?
What are your thoughts on inositol, specifically for reducing unwanted limerence?
Most of the studies I've seen on inositol have been negative, but I've never heard of that particular application before!
If by "reducing unwanted limerence" you just mean sex drive, I'm surprised you would choose it over any of the basically-SSRI-things that are known to work well for that.
The limerence I'm talking about is unrelated to sex drive. More of a specific flavor of OCD.
This is the only example (other than myself) which I know of: https://www.lesswrong.com/posts/Jk9yMXpBLMWNTFLzh/limerence-messes-up-your-rationality-real-bad-yo?commentId=bddudmFAqxjo3dGPj
In my own example, I was trying to get rid of the persistent, soul-being-torn-out feeling after ending a romantic relationship with my girlfriend. I pre-register the experiment with her (this is a very romantic thing to say to ones former girlfriend?) I used placebo capsules to self blind, but I think either inositol itself or the cognitive effects last longer than I expected. The first time I took the active side, the symptoms went away for >week. Since then, the symptoms have been sufficiently rare and intermittent that I have not been able to collect much more data.
I'd be interested in hearing about other people's experience with inositol for limerence, if any.
(Does anyone have unpleasant limerence they want to fix and want to pre-register a experiment? (Does anyone want to make me fall in love with them and then break up with me so I can run the test again myself?))
There's two things that annoy me about Astral Codex Ten
1. Unlike every other Substack, the comments are not on a separate page to the main blog post. ALL the comments are under the main blog post. I usually read Substack on my tablet, a Kindle Fire, but I can't read Astral Codex Ten on my Kindle Fire, because the page is so large it just crashes!
2. Unlike every other Substack, people can't "like" comments here. This is annoying partly because with so many comments on every post I'd like to be able to see the ones that readers have decided are the best, and partly because sometimes it's useful to be able to see, if I revisit a post, which comments I have liked.
My question is: Why?
We did have likes but many people (myself included) didn't like them (too easy to turn into a popularity contest, and then weaponise such contests) so Scott listened to the voice of the people and had them taken off. For example, I do not think that "most likes" necessarily means "best comment"; I've seen elsewhere how it just means "yeah man, stick it to 'em!" or "I don't really think this is good but I'm upvoting for the lulz" and so forth.
I think they still work in some way if you really want to give likes.
These are both deliberate.
1. Substack's usual comment policy is that it shows one or two comments, then you have to click "Load Comments" every time you want to see another two or three, then Load Comments again to see another two or three more, and most people never bother and comment sections are useless. I prefer being able to see all of them at once.
2. People have consistently voted down likes for comments, because they think it would turn this into a Reddit style popularity competition. Part of this is that the community started on WordPress where those don't exist, people liked the system there, and they want to preserve it.
I really wish that there were likes. The current system worked very well at earlier points where there were (seemingly) fewer total comments. Now there are enough comments (and, to some extent, my time commitments have increased enough) that there's no way I can read them all, and there's no way to sort them into anything but effectively a random sampling (with some bias toward people who are online a lot and are in a good position to drop whatever else they were doing and comment). Being able to sort by likes is definitely imperfect, but it's much better than having nothing useful at all.
My approach (which obviously does not scale) is to free ride on the rest of the commentariat: if there are more than ~200 total comments by the time I get here, don't bother with any top-level comments that have no replies (or have only self-replies).
That does help somewhat, although in practice not enough that I actually do it except for posts where I'm especially interested in seeing what people say (this post being a clear example of such).
I see. I use a browser to read Substack - I've never used the app - and my experience is that on every other Substack when you see a post you see the top two comments under the post and then something saying "n more comments...", which when you click on it takes you to a separate comments page. And on that page you see a lot of comments (I just checked on a post on Richard Hanania's Substack and it was 54 comments) followed by "Load More", which if you click on it gives you lots more comments. I can understand that if someone can only see two or three comments at a time using the app, and they have to keep clicking to load more comments, they might not bother.
I have no strong opinion about likes (my weak opinion is I like them) but I'm strongly in favor of having all the comments visible at once. Very important for being able to grep (ctrl-f) for things. Drives me crazy that this isn't the case on any other Substack but ACX.
Seconding this. I wouldn't mind the comments as a whole being a separate page (the bit about the article crashing on loading huge threads is legit), but when comments are shown, it should be all of them. Don't break standard browser tools.
(I mean, one could probably get rid of such crashes by *not using !@#$%ing javascript*, but that's just a lost cause on today's web.)
How horrifying is it that the Substack phone app has no search function?
I've actually never tried the Substack app. Should I?
I use an iPad a lot, and for some reason I ended up using the substack app on that. I suppose it seems convenient on that platform. But if I want to search for something, I have to figure out how to open a post in a different app, typically a browser. If you haven’t been annoyed by using a browser to read Substack, I am not sure what I can use to entice you.
The answer to both of these is that it's meant to replicate the original Slate Star Codex experience. Both of these were true of the original site, and after moving, Scott didn't want to change what worked and asked Substack to let his page be that way.
Taking these points as suggestions for improvement, I second the former (and am fine with the latter). Same issue on my phone
Do you beleive the recent democratic backsliding is an omen of longer and larger trends around post literate society and a new social media culture or is this a blip on a larger trendline?
What are your thoughts on the health impact of sodium consumption? Given a reasonably healthy person, does sodium consumption have much of an impact on blood pressure or other metrics? Is it worth making an effort to reduce sodium intake?
In my wildest dreams I imagine you being asked by Trump to head up the FDA. Would you take the job? Do you ever get asked by anyone related to Trump to give your thoughts on anything? And any hopes the new administration might do something to sort out IRBs?
With the caveat that this would never happen in a million years, yeah, I would take it, just because I assume I could find really good advisors and that they would be better than whoever took the job if I refused.
I have never been asked by anyone in the administration to give my thoughts on anything.
If your slate was clean - no particular attachment to family, friends, workplace, etc. and you had to decide a country to live in based purely on your expectations on how that country will legislatively handle AGI / a potential singularity / AI-risk, where would you live? Between the US, the EU, Switzerland / UK, or other?
just my personal opinion but i'd go to a developing country, provided you can earn an income somewhere. most of the scary scenarios involving AI affect developed countries before they affect undeveloped ones.
I'm a bit confused because I don't expect national level policy to matter. Even if Switzerland has amazing AI risk policy, if the US screws it up and creates unfriendly AI, Switzerland is still dead.
That's fair on ai-risk killeveryone scenarios. I suppose I was asking more for a scenario where everyone is not dead and we manage to create friendly-ish AI and which national level policy would be able to best handle the social / economic consequences. Similar dynamics apply where third countries might have entirely reactive policies to those in the US, but nevertheless would be interested in your thoughts.
I think superintelligence is likely to be pretty winner-take-all - I would be surprised if countries maintained substantial independence and different policies after it while also being anywhere near the technological frontier. I think it probably spirals into utopia or dystopia pretty quickly and regularly and cross-border policy regimes aren't likely to matter much.
What are your thoughts on the recent actions of the Trump administration? I read a lot of news on Reddit, and there are a lot of things where people talk about Trump’s actions as if they’re part of a fascist coup, or they’re going to result in the entire collapse of the U.S. economy. I know that some of this is hyperbolic because Reddit is a liberal bubble, but my impression is that Trump is also doing some legitimately bad and unprecedented things.
As some examples, I don’t have a good sense of how unreasonable it is for Elon Musk to have the access he does to government systems, nor do I have a sense of whether the tent facilities being built at Gitmo are a particularly unusual development.
I’d be curious to hear your take on which issues are real problems and which issues are totally overblown.
I don't think Musk himself is breaking this law, but it seems like he's helping Trump to break the 1973 Congressional Budget and Impoundment Control Act. And I believe there are legal protections for civil servants that Trump may have violated. IANAL.
I don't know enough about law to have an independent sense of how legal Musk's actions are. I don't really trust him to not do them if they're illegal, which is bad.
My impression was that a lot of really good programs got canceled (eg PEPFAR) along with a lot of woke garbage. I wish that he had taken a little while longer to sort the baby from the bathwater, and think that laws exist to protect us from the sorts of people who don't do that.
I think there's a possible world where he gets his act together, restarts all of the ones he shouldn't have cancelled in the first place, and it's long-term positive, but I don't really trust him to be an actual person who cares about bringing that possible world to fruition, and I would rather there be laws than me have to trust him.
See also https://scottaaronson.blog/?p=8609 (Scott Aaronson lamenting the ongoing science funding disaster; see also the link at the end to Terence Tao with a nice analogy for how harmful this can be.)
Any thoughts on seed oils / linoleic acid since you wrote this in 2020? https://slatestarcodex.com/2020/03/10/for-then-against-high-saturated-fat-diets/
Some interesting recent summaries of evidence and different viewpoints: https://www.lesswrong.com/posts/DHkkL2GxhxoceLzua/thoughts-on-seed-oil
https://www.exfatloss.com/p/what-causes-obesity-and-how-to-reverse
Dynomight convinced me to worry less about seed oils: https://dynomight.net/seed-oil/
(I'm definitely interested in Scott's take as well.)
do you think we should be evaluating our chances of natural disaster and preparing for it? recently have had the realization that i spend a lot of time thinking about existential threats like AGI but not a lot of time thinking about personal physical threats specific to my location
I think more than zero, less than an amount that wastes your time. I have a few tubs of nonperishable food in my basement, a drawer with all really important documents (eg passport) so I can grab them all in an emergency, and otherwise I don't worry about it.
Water is more critical to stock than food!
Most Americans could survive weeks without eating but we die of thirst much faster.
Is Vagal Nerve Stimulation for psychiatric conditions on your radar, and if so what are your thoughts on it?
There has been some reports of it in the press eg [1]. There are purchasable devices, although they make me a bit nervous (one advertises that it is safe because the *FCC* approved it, I guess assuming that people will mistake that for the FDA).
[1] https://www.theguardian.com/society/2023/aug/23/the-key-to-depression-obesity-alcoholism-and-more-why-the-vagus-nerve-is-so-exciting-to-scientists
I don't know much more than you. It sounds promising, but there are always promising things that don't pan out.
In 'Half an Hour Before Dawn in San Francisco', you write: 'Imagine living when the first lungfish crawled out of the primordial ooze, and missing it because the tidepool down the way had cheaper housing.'
What's do you expect your price is?
Specifically: what compensation would make you seriously consider leaving the Bay Area for 10 years, with these conditions:
- Must live >150 miles from SF
- Can visit for up to 2 weeks at a time
- Maximum 90 days in Bay Area per year
It's got to be more than an extra 100K per year because I could save that in housing costs alone. I'm often on the border of moving, so I don't think too much more. Maybe 300 or 400K per year? I would miss my friends and babysitter though.
Do you think legislative secret ballots would be an improvement? Do you have any ideas on how to change the short term incentive structure of government?
Post a pic of your bookshelf
that's a good one.
+1
Yeah, I love looking at other people's bookshelves
Which one? :-)
I was watching a couple of GPs lament the shortage of psychiatrists and the lack of support from insurance companies for treatment. Do you think AI-powered therapy, with a convincingly human avatar—perhaps even tailored to the patient—could provide meaningful help in closing the gap in the near future?
I was planning to ask this somewhere else at some point but essentially the other day I noticed I was confused regarding the FDA.
First of all, I broadly believe RFK that the FDA is letting too much junk through as there is so much money and will behind the Pharma industry it would be impossible for one little agency to block them from profiteering. Show me the incentives and I will show you the result and the incentives for pharma regulation basically point one way.
On the other hand most of what I know about the FDA I got from you and I summarise your position as the FDA is too rulebound and slow and keeps blocking good research for no reason. I believe you have argued for this extremely convincingly in the past.
So I guess the question comes down to do you still think the FDA is still too slow and hesitant ? Do you have any sympathy at all for RFK's issues with the FDA ? Finally are my two contradictory opinions as ridiculous as they sound or is there a way to hold both intuitions ?
To what extent does the junk fall into the F rather than the D?
Too lenient on food additives & too strict on new drug candidates does not obviously conflict with anything I've heard about the FDA.
I think these can both be true. In a world where there's a strong expectation that an FDA-approved drug definitely works, is definitely safe, and is definitely novel and exciting, the FDA can be failing to achieve that, and so misleading consumers. But that's a higher bar than what would actually make the country best-off, where it might be better to have 10x more drugs but only 90% of them work, or something like that.
That having been said, I haven't been following RFK's complaints, and I don't know of any really egregious examples except Aduhelm.
I talk a bit more about options at https://www.astralcodexten.com/p/beyond-abolish-the-fda
What would your swag at the causes of the Flynn effect be?
Education and intellectual environments (eg so much stuff is written that everyone is literate) which build skills somewhere halfway between problem-solving and test-taking.
How do you prepare for AGI? How do you not despair?
dwarkesh tweeted recently a question he asked dario in his interview:
> One question I had for you while we were talking about the intelligence stuff was, as a scientist yourself, what do you make of the fact that these things have basically the entire corpus of human knowledge memorized and they haven’t been able to make a single new connection that has led to a discovery? Whereas if even a moderately intelligent person had this much stuff memorized, they would notice — Oh, this thing causes this symptom. This other thing also causes this symptom. There’s a medical cure right here.
Shouldn’t we be expecting that kind of stuff?
do you have any thoughts/intuitions on this?
Humans also aren't logically omniscient.
My favorite example of this is etymology. Did you know that "vacation" comes from literally vacating the cities? Or that a celebrity is a person who is celebrated? Or that "dream" and "trauma" come from the same root? These are all kind of obvious when you think about them, but I never noticed before reading etymology sites.
I think you don't make these connections until you have both concepts in attention at the same time, and the combinatorial explosion there means you've got to go at the same slow rate as all previous progress.
Another example: I've only just realized after all these years that "prince" literally means first, as in first in line to inherit the kingdom, and I only realized it because I was translating something into Spanish, where the word is closer to its root in Latin.
"dream" and "trauma aren't obviously related: "dream" is from proto-Germanic *draumaz meaning dream, from proto-Indo-European *dʰrewgʰ meaning "to deceive"; "trauma" is from Greek, ultimately from proto-Indo-European *terh₃ "to hurt." The "vacation" etymology also doesn't seem quite right, it originally meant something more like "freedom from obligations" and the sense of travel came later (though of course it is still related to "vacate" etymologically).
I don't know if this says anything about the difficulty of finding non-spurious connections among all the combinatorially possible ones...
(sources:
https://en.m.wiktionary.org/wiki/trauma
https://en.m.wiktionary.org/wiki/dream
https://www.etymonline.com/word/vacation)
Anyone know how specific the etymology example is to English? Even within English, American English spelling reform sacrificed a lot of etymological indicators.
i did not; neat!
I have a bet going with a friend, and I try to get an opinion on this from everyone I can.
Do you believe that there will be commercially available humanoid robots capable of autonomously matching human domestic labor (eg cooking, cleaning) by July 1 2034?
It doesn't have to be cheap, just available to the public for some amount of money, and it needs to be fully autonomous, with no expectation of a human being on the other end taking control. As for competence, it needs to be both capable at individual tasks and capable of acting inside a domestic space in a way that wouldn't swiftly get a human being fired. A lengthier discussion of these criteria and my thoughts about them, at the risk of self-promotion, is located here https://nicoroman.substack.com/p/against-humanoid-robots
Seems like a very underspecified bet. Do you mean ANY form of domestic human labor? If so, imdies the bet fail if the robot is bad at any one random thing? If not, how much better than a rumba are we talking here?
More in depth specifications in the linked post. In short, I’m looking at basic competence in cooking and cleaning. Nothing fancy or requiring specialized skills, just tasks which any able-bodied person off the street should be able to do: mop, sweep, tidy up, cup of morning joe and an omelet, wash the dishes by hand and stack them. This whole bet originated from my friend’s claim that, by the specified date, menial human domestic workers would only be able to compete with robots by cost.
If things started to look close, my friends and I would need to sit down and make the bet more detailed, valuing each kind of task and detailing criteria for success on each. At present, neither of us think it’s necessary: both of us expect to win the bet without resorting to squabbles over whether a given robot passed the broom benchmark.
I’d consider doing that if I saw reliable humanoid robot waitstaff: walk around a party on flat floors carrying a tray of drinks, take people’s empty glasses and offer them new ones, etc. That’s the benchmark I’m waiting to see before I get worried about losing this one.
Seems investors anticipate a massive market by then. Money is a limiting factor for domestic use, but businesses and institutions will probably be the key purchasers. You can technically get humanoid robots now absent AI and they cost a fortune.
https://www.grandviewresearch.com/industry-analysis/humanoid-robot-market-report
Yes, because there will be an intelligence explosion in software-AGI and some of the resulting above-human-level AGIs will get redirected to inventing good robots.
I'd be very happy to offer you 2-to-1 odds if you're interested in a bet.
A man after my own heart!
bloody hell....
Could you elaborate? I have the same expectation as Scott (at least in the branch where no runaway ASI turns every atom from the orbit of Mercury to the Oort cloud into a computronium Dyson swarm...). Why the reaction to household robots?
It was more Scott explicitly saying "there will be an intelligence explosion" without even a probability modifier.
I see, Many Thanks! ( I _would_ put a probability on AI advances, currently at 80% odds of AGI within two years. Research and development is always chancy and can always run into a roadblock, though I give this work very good odds. )
What are some things you have high confidence (80%+) an arbitrarily intelligent AI will remain unable to do (e.g. faster than light travel, polynomial solutions to currently NP-hard problems). I'm trying to understand what people who lean toward intelligence as arbitrarily scalable view as limits of natural law that no amount of intelligence can get around.
I'm not sure what you're expecting other than "impossible things", of which the two you mention are good examples (assuming P =/ NP). I feel like to know what's impossible I would have to know things about math/physics/etc, which seems like a different skill from knowing things about AI.
I would not expect teleportation, time travel, hypercomputation, etc. Someone (I think Bostrom) had an interesting analysis of the maximum amount of computation that could be performed per gram of matter, but I can't find it right now.
Possibly you're thinking of "Ultimate physical limits to computation" https://arxiv.org/pdf/quant-ph/9908043
Reflecting on the emergence of AGI on his blog yesterday, Sam Altman wrote that: "it does seem like the balance of power between capital and labor could easily get messed up, and this may require early intervention." Do you see this possibility of power imbalance as one of the most important issues around AGI, or as more of a second-order effect, less important than things like conventional alignment research?
Your phrasing made me read this as AGI emerging on Altman's blog yesterday. If this happened to me, I might also be in a reflective mood :D
Is there any primer or guidance for how to get educated on the sum of the community knowledge generated before SSC really got rolling?
I began reading SSC around 2017, but I’m aware that there was a formative and then maybe like an expansion stage where a lot of debate was had. I was old enough to be on lesswrong (28 when in 2009), but just wasn’t aware of it until almost a decade later. I’m not even sure what other rationalist and EA primary sites had legit community and intellectual traction before 2015.
Have you tried readthesequences.com (also, early Overcoming Bias has a lot of gems!)
If you had to live somewhere that isn't san Francisco, los angeles, or new York city, where would it be
Suburbs of Seattle, sorry for the boring answer.
But you liked Switzerland so much ...
Excluding AI, what is the most dire risk facing humanity ? And what do you believe should be done about it ?
whats ur fiction writing process? we love the bay area series and as a guy who's only seen twitter it gives me major twitter vibes. like a collection of little jokes. do you collect them and poast whenever you get to a dozen? how does a joke appear in ur mind?
What do you suspect is the principal cause of social capital deterioration, and is there any actionable policy response to it?
I'm curious about your opinion on surveys about things like happiness or suicidality, ala "Suicidality has nothing to do with suicide to the point where it's almost a reverse indicator".
I've sorta-kinda stopped trusting these, but I'm very uncertain how justified that is. My usual intuition these days is if group X reports it's exceptionally well-functioning or exceptionally sad compared to other groups to prioritize the view that this is more a *communication style* thing than an actual, trustable indicator. So I'd expect Vegans to report greater happiness due to their diet, but I'd expect them to say that even if it made them miserable.
I put that book of text there to explain the question, not to bias you - I'm looking to have holes poked in this. I think the short version of the question is "Do we only use mood/feelings types of surveys because there's no better option, where an even slightly good one would end that practice forever? Or are they much better than I want them to be?
Happiness research is very tough and you often get different answers depending on how you ask the question, ie "are you satisfied with your life?" vs. "what percent of days do you feel really happy?" etc. When all of these different things point in the same direction, I usually trust there's something there; otherwise, there might be something there but it's probably complicated and not too worth worrying about.
I wonder what would happen if you simply asked vegans and nonvegans "how happy are you?" without mentioning that it involved veganism. I think this works for things like religion and child-having which have major effects, but maybe it would be too weak for diet in particular, or pick up the wrong thing (eg vegans are more liberal, so maybe it would instead be picking up liberalism). I don't know, seems like there might be something in this category.
> child-having which have major effects
Piggybacking on this - did you notice your happiness taking a hit after you got children?
I think somewhat. It depends a lot on how much capacity my wife and nanny have - when they're both doing well and feel like taking 90% of the work between them, almost no hit. When they're sick or tired and ask me to take on more of the work, significant hit.
I would be curious to see your thoughts on fatherhood, especially this part, elaborated on. I noticed I am the same, but the popular narrative of "Spend as much time with your child as you can" creates a sense of guilt when I am working or relaxing on my own. I've come to terms with it, but it is hard to talk to other parents about it, and I am interested to see you say things eloquently as a way for me to ruminate on it more.
There's a poem about it:
The Two Parents
by Hugh MacDiarmid
I love my little son, and yet when he was ill
I could not confine myself to his bedside.
I was impatient of his squalid little needs,
His laboured breathing, and the fretful way he cried,
And longed for my wide range of interests again,
Whereas his mother sank without another care
To that dread level of nothing but life itself
And stayed, day and night, till he was better, there.
Women may pretend, yet they always dismiss
Everything but mere being just like this.
I think one of my thoughts here is probably the same as your - i.e. that to some extent there's some training for "the right answers" where the right answers are the ones that most defend the answerer's value. So if you ask a Vegan how happy their diet makes them, they know the "right answer" is "it puts me next to god in the clouds", so long as the survey is designed in such a way that they can intuit there's a "is veganism good" question trying to be answered.
I admit freely that I'm pretty bad at actually figuring out if there's that kind of accidental/intentional guiding going on, which makes me assume it is, which makes me discount too many things as invalid. What should I be looking for in a survey to see if they controlled for it?
I kind of want to fight your hypothetical here. At least with parenting, people ask happiness related questions, and it seems like the effect you suggest would show that people say they are happy even if they aren't. But happiness research is clear that parents are much less happy than nonparents while their kids are in the house. So there's at least one case where you can get info despite signaling risks.
I think is super hard to control for though. If there is some signaling effect like that, the best you can probably do is ask other questions you expect to be correlated with the answer you want and hope signaling doesn't affect those.
Yeah. Without seeing the specific research you are talking about, I can take it as true. I think I'm uncertain enough on this entire thought that I wouldn't be surprised if I'm looking at it from entirely the wrong angle, or if there's just so many different exceptions to it it's hard to quantify.
In https://www.astralcodexten.com/p/highlights-from-the-comments-on-housing and the following open thread (https://www.astralcodexten.com/p/open-thread-276?utm_source=publication-search), it felt to me like you changed your mind more because of the survey results than because you were convinced by the object-level arguments. I also got the impression that there was some talking past each other between you and commenters (possibly including me) on what the actual question is.
1. How convinced are you that your original thesis was wrong?
2. Is there a specific time range that you were thinking of, where the causal effect is denser housing -> higher prices?
3. Is there a specific policy or set of policies you cared about the effect of? Density doesn't increase for no reason, so I think a question like "does density increase prices" is under-determined. See also rule 5 of https://www.reddit.com/r/badeconomics/
I'm pretty convinced that there are two separate effects - cheaper because of more supply, more expensive because of density-related amenities - and that they fight each other on various timescales. When I wrote the post, I thought these immediately canceled each other out and the stronger (the second one) was the only one that mattered. I genuinely changed my mind in the sense of disentangling these more and seeing that you could get one without the other for path-dependent reasons, at least for a long time. The survey convinced me that most real-world cases are probably ones where the first one is stronger at least for a while.
I cared about the policy situation where everything else is held constant, but YIMBYs win a policy victory and cause more housing to get built anyway.
I think induced demand has much stronger implications than people realize. The idea is: building a new home allows someone to move into the city, which creates amenities (start a business, do research, start a club, etc) that cause more people to move in. Since demand increases by more than the increase in supply, prices rise.
So if you want to avoid induced demand, you need to ban all new housing, even in the suburbs: a new resident commuting to their job downtown will also contribute to amenities.
https://x.com/michael_wiebe/status/1829019082043339196
This is what Scott thought at first, and it's contradicted by pretty much all high-quality empirical research of which I am aware. The supply effects outweigh the demand effects.
Also, you probably should be grappling with the fact that "ban all new housing" is effectively what many places in the US have tried, and they've generally seen massive increases in housing prices (exactly as intended--"home values" is a standard NIMBY rallying cry).
Also: amenities are not just "created", they are often relocated from somewhere else. Talented people move to the big city - company in a province town closes shop and reopens in the big city because they can't find talented people in the province anymore - talented people move from the province to the big city because there aren't any good jobs in the province anymore.
But the whole idea of agglomeration effects is increasing returns: you can do things in the big city that you can't do elsewhere.
Increasing returns *for whom*? There are always good reasons for someone to take a small step that gets them better returns, but makes everyone else slightly worse off. I vaguely remember Scott writing an article about it...
Up to a point, agglomerization is fine - we need cities of some size to support modern civilization. The question is, are ever-increasing, ever-more dense cities an inevitable global optimum, or do they turn into a moloch trap at some point, and efforts to delay that point are fully justified?
Yes, the small town loses, but with increasing returns, the gains in the big city are larger, so the net effect is positive.
I think it's correct that, for example, someone living in Manhattan when it was a tiny Native American village, if their only goal was to prevent Manhattan land prices from going up, should have banned people from building there (especially if they could guarantee that other jurisdictions wouldn't follow their example, and building would remain legal everywhere else).
But I think most people aren't in this position, and it's defecting (since you want growth to happen, just not near you), so governments shouldn't be making policy based on it.
Thanks. I think I understand your position better now than when I first read your posts. And I definitely think that in that policy regime, there's very strong empirical evidence that the supply effects are stronger, at least on timescales of years to decades.
You don't, by chance, have any idea if and where Lou Keep is writing these days, do you? I cannot find anything. Samzdat has been silent for years and Tauromachy nearly as long.
https://loukeep.substack.com/p/vagaries
Yeah, silent since October 2020, though. 😭
He has some tweets from this month (https://x.com/lou_keep), but they don't suggest he's writing anything longer.
Thanks Scott. I've managed to avoid Twitter entirely and so can't really search there. Glad he's alive, at least, much as I desperately miss his style, perspective, and erudition.
Have you ever looked into the feasibility of the potential worldwide switch to renewable energy sources? A lot of western countries are coming out with rather aggressive net zero targets, and there seems to be some debate about whether government targets for renewable energy are even feasible. It can be hard to determine to what extent this is all just signaling or a realistic depiction of the future, and the possible consequences of that.
Haven't really looked. Based on things I heard at https://www.astralcodexten.com/p/notes-from-the-progress-studies-conference , solar is spreading in a pretty amazing way, and there's room for nuclear too, but I don't know if the existing targets overshoot even that.
Is it possible for governments being overly optimistic about solar to coexist with https://www.reddit.com/r/solar/comments/1dknl7x/predictions_vs_reality_for_solar_energy_growth/ ?
Regarding psychiatry (and medicine in general): in your best estimation, to what extent are psychiatrists (and doctors in general) following state-of-the-art research in performing diagnoses and prescribing their treatments; vs. using own experience as their primary guide ?
Depends on the doctor, but I think they're doing a pretty good job at using state of the art, in the sense that there are regularly-updated guidelines, most medical students/residents learn them, and most doctors continue working off them. There's also something called CME where you have to prove you've read some recent research to get your license renewed every few years.
In psychiatry, I think things change very slowly, so that your question is barely relevant. In other fields, they change more often. My father, an internist, says he disagrees with many of the recent guideline changes - but he sure does know about them.
What would an argument look like that could convince you modern poetry is Good Actually? And when, for poetry, would you date post-tartarianism? The fall?
I don't know how you do aesthetic arguments. You'd have to show me modern poems that I like. If I'm trapped in some hating-modern-poems equilibrium, maybe lead me out with a series of poems each of which is slightly more modern than the last, so that appreciating N helps me appreciate N+1.
The last poems I really like are from the 1930s. I would like to tell a story where the trauma of WWI destroyed people's hope and belief in beauty and made them create bad art, but for some reason it's WWII, which makes less sense to me!
C. S. Lewis wrote a satirical bit about the poetry scene int he 1920s and 30s in his allegorical book "The Pilgrim's Regress". Since he was a young academic at the time he would have had direct experience. In this scene our protagonist has been attending a poetry reading:
"Glugly instantly rose. She was very tall and as lean as a post: and her mouth was not quite straight in her face. When she was in the middle of the room, and silence had been obtained, she began to make gestures. First of all she set her arms a-kimbo and cleverly turned her hands the wrong way so that it looked as if her wrists were sprained. Then she waddled to and fro with her toes pointing in. After that she twisted herself to make it look as if her hip bone was out of joint. Finally she made some grunts, and said:
‘Globol obol oogle ogle globol gloogle gloo,’ and ended by pursing up her lips and making a vulgar noise such as children make in their nurseries. Then she went back to her place and sat down.
‘Thank you very much,’ said John politely.
But Glugly made no reply, for Glugly could not talk, owing to an accident in infancy.
‘I hoped you liked it,’ said young Halfways.
‘I didn’t understand her.’
‘Ah,’ said a woman in spectacles who seemed to be Glugly’s nurse or keeper, ‘that is because you are looking for beauty. You are still thinking of your Island. You have got to realize that satire is the moving force in modern music.’
‘It is the expression of a savage disillusionment,’ said someone else.
‘Reality has broken down,’ said a fat boy who had drunk a great deal of the medicine and was lying flat on his back, smiling happily.
‘Our art must be brutal,’ said Glugly’s nurse.
‘We lost our ideals when there was a war in this country,’ said a very young Clever, ‘they were ground out of us in the mud and the flood and the blood. That is why we have to be so stark and brutal.’
‘But, look here,’ cried John, ‘that war was years ago. It was your fathers who were in it: and they are all settled down and living ordinary lives.’
‘Puritanian! Bourgeois!’ cried the Clevers. Everyone seemed to have risen.
‘Hold your tongue,’ whispered Gus in John’s ear. But already someone had struck John on the head, and as he bowed under the blow someone else hit him from behind.
‘It was the mud and the blood,’ hissed the girls all round him.
‘Well,’ said John, ducking to avoid a retort that had been flung at him, ‘if you are really old enough to remember that war, why do you pretend to be so young?’
‘We are young,’ they howled; ‘we are the new movement; we are the revolt. ’
‘We have got over humanitarianism,’ bellowed one of the bearded men, kicking John on the kneecap.
‘And prudery,’ said a thin little old maid trying to wrench his clothes off from the neck. And at the same moment six girls leaped at his face with their nails, and he was kicked in the back and the belly, and tripped up so that he fell on his face, and hit again as he rose, and all the glass in the world seemed breaking round his head as he fled for his life from the laboratory. And all the dogs of Eschropolis joined in the chase as he ran along the street, and all the people followed pelting him with ordure, and crying:
‘Puritanian! Bourgeois! Prurient!’"
Have you read https://gradual-disempowerment.ai/? What's your take on it? Do you plan to write about it?
It's on my list but not yet.
Did you use polygenic embryo screening to select your kids? If so what did you select for?
Not the current twins. We plan to use it for #3, but are still deciding details.
Mind if I ask why not?
Because we wanted to have kids naturally rather than go through the time/suffering/money of IVF. Turns out we're not very fertile and it still took a lot of time/suffering/money (plus enough fertility drugs that we had twins instead of a singleton), so we'll give up and do IVF next time.
I know that you have read up on Buddhism? Have you read up on Hinduism and specifically Advaita Vedanta? Does that change your thoughts on consciousness?
Do you look at Buddhism/Hinduism as philosophy, exploration of consciousness or just a religion no different than Christianity/Islam/Judaism?
What is your opinion on Moksha/Nirvana concepts in Indic philosophy?
I have read a tiny amount on the Upanishads etc. I think the theory that we are all God deluding ourselves into believing we are individuals has some philosophical/aesthetic attraction, but I'm not really sure what to do with it - like the Simulation Argument, it doesn't necessarily lend itself to life changes.
I find Buddhism/Hinduism very interesting as exploration of consciousness, but I would have to meditate a lot before I felt comfortable having an informed opinion, and I don't seem good at sticking to this.
I'm optimistic that this decade will bring more options for converting dollars into meditation results (jhana/awakening/etc). How much would you pay to press a button that, as long as you held it down, you remained in jhana (or "awakened" consciousness, or similar)?
thoughts on a potential inflammation component of attraction? something like people being able to subconsciously detect diet induced inflammation (or induced through other ways but thinking diet for most people) -- antihistamines as a minor boost to attraction in this case? do you think it's plausible/if so do you think it's significant?
As in interpersonal attraction? You're attracted to partners who have less inflammation? I think this is probably covered under the many markers of general health that we know people use, like skin quality.
How much have you been recommending Silexan/Lavender and have you found out anything new about it from your experience since the last time you posted about it?
I recommend it occasionally, most people (including me) notice an effect, but it's still just that one biased group doing studies.
What video games are you playing these days?
Still Civ4. Most common mods are Fall From Heaven, PIE's Ancient Europe, and Realism Invictus.
Any thoughts on Civ 7? I've been hearing good things from fans of the series.
I have found every Civ since 4 disappointing (too focused on gimmicks that take me out of immersion - great conquerers of the past just didn't spend that much time planning how their absolutely giant city districts synergized with each other) and probably won't even bother trying 7. Graphics look pretty though.
I have mixed thoughts on 7, but I think it is probably different for someone who tries all the new 4X games to see what they think about system design vs. someone who's still on Civ 4. (Like, one of my takes is "I already played Humankind, thank you very much" but--you haven't!)
There are two main changes, from my perspective. First, the ages are meaningfully distinct (but, coming from Paradox Interactive games, only weakly distinct), and your nation becomes a combination of unique age-appropriate traits instead of, like, having to balance the Greeks who peak early vs. the Americans who peak late. Second, when you place workers, they stay there, and most of your settlements are 'towns' that don't have production (but still contribute to the national economy). So there's much less management of things, and much less clicking--in a way that jives with the "the great conquerors didn't bother with that" experience but still does involve placing districts in ways that synergize. (Most of the synergies are very weak, compared to Civ VI, so this isn't that big a deal.)
I think I like it more than VI but it's hard to tell. (There's a handful of UI elements that desperately need improvements--in particular, when you overbuild a building from a previous age, the tooltip just tells you what the new building will get you, and not what you're losing from the previous building, which seems potentially tragic.) I think the city vs. town distinction is good but, uh, my favorite version of it actually dates back to Imperialism (released 1997). You can change your capital on age transition, in a way that seems historically appropriate but so far doesn't seem to be worth it gameplay-wise, which sort of sucks.
Talk about serendipity, I tune in to my regular Tuesday "Tasting History" and this week's episode is sponsored by "Sid Meier's Civilisation VII" 😁
https://www.youtube.com/watch?v=U9134Rs_0co
Feeding Emperor Augustus Caesar - Handmade Roman Cheese
You don't come on podcasts. Not even mine!! But do you listen to them or are they too inefficient as a way of getting information?
I don't listen to them.
I'm assuming that you're agnostic or an atheist, but if you had to convert to an extant religion based on which one seems to have the best evidence for it or best corresponds to your existing worldview, what religion would it be?
Judaism, it's my culture, I'm pretty attached to it, and my family put in too much work keeping it to let the goyim win.
If I had to convert to a *different* religion, Christianity, because I like the altruism, Lewis and Chesterton are great, and I could finally needle online Christians who go against everything their religion believes in without being vulnerable to the meme about how I don't believe it and it's just a way to make them do what I want.
Do you have a particular denomination in mind?
No. It really would depend on why I converted; right now I'd be tempted to find the most atheist-lite one that isn't hopelessly woke, but I assume if I was actually converting for some reason then that reason would give me more guidance.
>I'd be tempted to find the most atheist-lite one that isn't hopelessly woke
Cue Pseudo-Dionysius vs Richard Dawkins saying "There is no such thing as 'God'" in the Human Resources meme format
Do you feel any pressure/responsibility to sound the alarm harder on AI risk as things have accelerated and you’ve acknowledged it’s a serious and important risk? You don’t seem very urgent about it but not sure
Yeah. I don't think I would describe myself as "not urgent" - I try not to let my blog become strident and moralizing because then I turn away everyone who isn't already convinced. But also, there's only so much I can say without being boring and repetitive. I'm helping with a pretty interesting AI risk project now that I hope will give me more to write about once it's ready.
How are the twins? I’ve got a 3 month old son, and have enjoyed your previous family posts a great deal. Cheers!
Congrats on the child! They're doing fine. Wish they would learn to talk a little faster but they're making up for it with constant babble.
Are they learning baby sign language?