873 Comments

I remember once coming across a wordpress blog from the early 10s where a guy reads books and posts his reactions chapter by chapter. I remember reading his series on The Lord of the Rings, and remember that for the last chapter of Fellowship, he included a bunch of photos of himself making faces while talking to the characters. Unfortunately, I don't remember the name or much else about it. It's a bit of a longshot, but does anyone know what I'm talking about?

Expand full comment
Sep 23, 2022·edited Sep 23, 2022

Has Scott (or Freddie, or...) written about language learning?

Inspired by a New Yorker piece (https://www.newyorker.com/news/annals-of-education/the-rise-and-fall-of-vibes-based-literacy) and, since it was paywalled, ensuing Motte discussion (https://www.themotte.org/post/75/culture-war-roundup-for-the-week/9574?context=8#context). Allegedly, schools are clinging to "whole language" learning despite it being debunked and well refuted compared to phonics. This is used as a jumping-off point for attacks on "vibes" based policy and general wailing and gnashing of teeth.

Except...

* Phonics is treated as if it were synonymous with entry-level phonetics, but also might be a framework built *on* those fundamentals

* Most of the cited evidence is about remedial usage

* It's not clear, to me, if that evidence or earlier 60s evidence has survived the replication crisis

* Whole-language approaches may or may not be marketed as an alternative "when phonics doesn't work" (https://www.newchapterlearning.net/phonics-vs-whole-word.html)

* Actually, the whole debate smacks of marketing; I suspect this is par for the course in education solutions

Is (the stronger form of) phonics actually well-supported? Why has this particular bit of High Modernism reversed the usual political valence? Are we all just thinking about https://en.wikipedia.org/wiki/Hooked_on_Phonics?

Expand full comment

The Arabic 'caliph' and the Spanish 'California' are likely related, via a 16th-century novel "Las Sergas de Esplandián":

https://www.reddit.com/r/etymology/comments/xkclmk/the_arabic_caliph_and_the_spanish_california_are/

Expand full comment

Hello folks!

I am glad to announce the Sixth of a continuing series of Orange County ACX/LW meetups. Meeting this Saturday and most Saturdays. The first few meetings were great (approximately 8 to 10 people), and I hope to see many of you at this one. Snacks will be available.

Saturday, 9/24/22, 2 pm

1900 Port Carlow Place, Newport Beach, 92660

The Picnic tables outside the community clubhouse

33.6173166789459, -117.85885652037152

https://goo.gl/maps/WmzxQhBM2vdpJvz39

Plus code 8554J48R+WFJ

Contact me, Michael, at michaelmichalchik+acxlw@gmail.com with questions or requests.

Activities (all activities are optional)

A) Two conversation starter topics this week will be. (readings at the end)

1) The FermiParadox (why don’t we see a lot of aliens)

2) The collapse of civilizations

B) We will also have the card game Predictably Irrational. Feel free to bring your own favorite games or distractions. This is a pet-friendly park and meeting.

C) There will be opportunities to go for a walk and talk about an hour after the meeting starts and use some gas barbeques if anyone wants to grill something. There are two easy-access mini-malls nearby with takeout hot food available. Search for Gelson's or Pavilions in the zipcode 92660. I will provide some snacks and water.

D) Share a surprise! Tell the group about something that happened that was unexpected or changed the way you look at the universe.

E) Make a prediction and give a probability and end condition.

F) Contribute ideas to the future direction of the group. Topics, types of meetings, activities, etc.

Conversation Starter Readings:

Suggested readings for this week are these summaries. These readings are optional, but if you do them, think about what you find interesting, surprising, useful, questionable, vexing, or exciting. This week we are going to try some videos.

1) The only popularizer I have ever seen get the essential elements of the Fermi Paradox correct is Isaac Arthur. Every other popularizer I have seen misses or misunderstands some essential piece that makes a wide variety of solutions seem plausible. He has a lisp, but if you can cope with that, this is a great overview of the main ideas and what seems to be wrong with them.

The Fermi Paradox Compendium

https://youtu.be/rDPj5zI66LA

2) The collapse of civilizations. This is a summary of an MIT analysis that from decades ago predicted that we would face civilizational collapse. A recent review indicates that based on the parameters they were looking at, we are ahead of schedule. Do you think this is a good assessment? Do you think it is comprehensive enough? Do you think we are in a decline but have gotten so used to the ways our civilization is decaying that we underestimate how far it has gone? Are the factors leading to civilizational collapse invisible or ignored b the general population? Do technological progress and cultural innovations change what we should be looking at?

MIT Has Predicted that Society Will Collapse in 2040 | Economics Explained

https://youtu.be/kVOTPAxrrP4

Expand full comment

For anyone who's been to one of the meetups before, how awkward is it? Do the people there often know each other? Does it get weird at all?

Is everyone relatively similar?

Expand full comment

The US government suppresses viewpoints they don't like (COVID, 2020 election, Hunter Biden's laptop), up to and including telling social media corps what to ban :

https://www.tabletmag.com/sections/arts-letters/articles/government-privatized-censorship-regime

Expand full comment
Sep 22, 2022·edited Sep 22, 2022

Headlines:

> Biden denounces Putin’s nuclear threats as ‘reckless’ in UN address

> Cornered by war, Putin makes another nuclear threat

> Putin Raises Nuclear Threat Following Battlefield Losses

> Putin announces partial mobilisation and threatens nuclear retaliation in escalation of Ukraine war

> Putin is not bluffing, says Serbian president on 'nuclear threat'

I'm slightly baffled that most of the media seems to have misunderstood Putin's speech as a threat to use nuclear weapons. It's standard Russian propaganda that NATO supposedly threatens Russia with nukes, as they make-believe they're fighting NATO in Ukraine. Putin's just saying "don't worry, if they use nukes, we'll use nukes" as a domestic propaganda message. https://twitter.com/DPiepgrass/status/1572814367946870785

With 300,000 conscripts shipping to Ukraine soon, it turns out calling the Russians "orcs" is about to become more apt. Hordes of men indoctrinated by Russian state TV will be given guns and pointed toward the "ukranazis"*. It had just become clear that Ukraine was winning, but while they were winning for multiple reasons, the biggest reason was Russia's manpower shortage — which is about to be fixed. Putin is probably willing to dump Russia's entire conventional arsenal on Ukraine (and already conscripted its weapons factories a couple months ago, though precision-guided munitions will probably remain in short supply). Sure, they will be poorly-trained, but "quantity has a quality all its own", and it's not like typical Ukrainian soldiers are well-trained either.

Ukraine can do offensives for a little while longer, and is likely to retake Kherson, but I'm thinking that "containment" and guerilla/partisan war will become the only option for Ukraine next year. World War I for the Ukrainians once more. I hope we'll give them all the weapons they need — and humanitarian aid, and homes for refugees.

Anybody from Europe? How bad is the energy crisis? I've seen conflicting reports on whether Russia can afford to continue cutting off its own income like this...

Anyone else think that banning Russians from getting EU visitor visas is a really stupid move that bolsters Russian propaganda about EU's "russophobia" and "naziism"? Let me be the first to say that I am pro-Russian, just anti-Kremlin, and more than that, pro-Ukrainian. Saw hordes of twitterites cheering on the bans against Russians; shook my head. People seem to have no common sense about sanctions, either — not realizing that if you slam Russian middle and lower-middle classes with sanctions, you disempower them relative to the regime, which makes them less likely to fight Putin, not more. I don't, however, know what it would take to create a rebellion in Russia, especially not one that would put someone better than Putin in charge. The time for action was 25 years ago.

Latest video from Vlad Vexler is good as usual: https://www.youtube.com/watch?v=HravTYSIVu4

And Perun: https://www.youtube.com/channel/UCC3ehuUksTyQ7bbjGntmx3Q/videos

* it's bad enough that the U.S. toppled a dictatorship and temporarily occupied it on a flimsy pretext like "WMDs" — Putin invaded a democracy on a pretext of name-calling "nazis" and "russophobes" and intends to occupy it forever. Sometimes they try out WMDs and other messages too though — because for Putin, Russia is one big messaging focus group.

Expand full comment

Is there any good open source prediction market software that I could run internally at my company (I should be able to get it running)?

Expand full comment
Sep 21, 2022·edited Sep 21, 2022

I am far less impressed with AI supposedly behaving intelligently than some.

For quite some time, we have known both that dumb models (think linear regressions and PID controllers) can often act in an intelligent-seeming manner in some situations while only understanding surface features, and that humans experts basically suck at writing them.

Think of naive-Bayes translation AIs that realized that nurses should have female pronouns and doctors should have male pronouns. It's not that they know that doctors tend to be male and nurses tend to be female, just that "doctor" is often connected with "he" and "nurse" is often connected with "she" in the training dataset.

Therefore, "model X acts intelligently in situation Y after being trained with a lot of data" feels like a fairly plausible consequence of "model X managed to act as the right sort of dumb statistical model in situation Y". And when you have a lot of data, you often don't need to think because the solution might very well appear in your training set.

And you can't just say "humanity is smart because it already has a very dataset". The dataset exists because it was written by humans.

Expand full comment

I once read a blog post about how Computational Complexity is the reason why Perfect Information games are interesting. Consider a game like Chess, it's trivial given infinite memory and infinite computation, where infinite means arbitarily unbounded (Potential Infinity, you can always request more). You can simply extend a tree of every single possible move down from the start state. Given such a tree, algorithms like e.g. MiniMax can demolish any opponent.

But such a tree would have ~10^157 nodes (game state, a 64-cell board with associated pieces), the estimated total number of subatomic particles in the observable universe is 10^80. Even if your memory encoding efficiency is 1 chess board per subatomic particle, you won't make it. You would need each subatomic particle to hide an entire other universe to just make the memory (10^80 observable universe * 10^80 particles per universe = 10^160) necessary to represent the tree. Iterating over that tree is a whole other thing.

This was a fantastic observation that still rings in my mind till today. We all instinctively know that physical games are only interesting because of the human body's limitations, but somehow intellectual games being the same surprised me. This also explains why the library of babel contains no actual human-comprehensible knowledge, why a computer program that enumerates every possible binary string of every length will contain no actual human-comprehensible knowledge, why mathematics is interesting despite all its theorems being tautologies from its axioms, why evolution needs selection pressure and brutal death to progress, etc... It all comes down to the impossibility of Brute Force Search. In a sense, all intellect\intelligence\algorithms are an approximation of Brute Force Search.

I can't find this blog post no matter how hard I try. It's on computer-science oriented blog, another post in the blog that I remember explores the notion of code optimization, specifically how optimization always entails code specialization (making the code less generic), there were code examples in Lisp and Javascript, the code was drawing the mandelbrot set. I tried alot using this info to find the blog but no success. Please help if this rings any bells.

Expand full comment

Scott, would you ever consider writing something on AvPD for Lorien Psychiatry?

Expand full comment

This is a dumb question. What makes for a natural harbor? I imagine it's something like water that isn't too shallow too fast that is largely protected from waves. Is that right and are there other important considerations?

Expand full comment

Has nobody started a Ukraine thread yet?

How do the situations in Azerbeijan/Armenia and Georgia change things? It looks to me like the focus on Ukraine is happening at the expense of Russia's ability to project power across its sphere of influence in the Caucausus. Unless they can stabilise things there, this might change the 'Russia still hasn't lost anything they didn't have before the invasion therefore they won't accept defeat' line of reasoning.

Expand full comment

Hello folks. Looking for an article. It was an introduction to a theory of consciousness. It was explaining how people model each other. It had pictures of people thinking about each other as they talk to each other. And within their thoughts about each other there was thoughts about each others' thoughts about each other. Sound familiar to anyone?

Expand full comment

I was listening to an podcast episode a few days ago where one host commented that he went to a convention and met someone who refused to read anything he wasn't sure was written by a human. I don't remember if the episode was new or old, but the anecdote was being told as though it happened several years before.

Do you think it's possible to follow such a rule nowadays without refusing to read anything from the past ten years in general? If so, how would you go about it? How would you apply the rule to AI art (such as if an artist you already followed started using AI generation)? How would you apply to rule to AI voices?

Expand full comment
Sep 20, 2022·edited Sep 20, 2022

A lot of writing about genetics, especially for a general audience, goes to great lengths to insist that heritability doesn't mean what you think it means, and that it's only about the proportion of population phenotype variance is accounted for by genetic variance within that population.

This is...somewhat frustrating. Often it's addressing what seem to me like strawmen of the lay understanding of heritability, for example that environment cannot affect the trait, or that the trait is completely immutable.

It's obvious to most people that some traits of an individual (nose shape, say) are quite likely to be similar to those of one or both of their parents, whereas others (I don't know, let's say favorite song) aren't. Is there some *other* word that geneticists would rather we use to refer to the scale on which those traits vary? Very informally, something like "the degree to which a trait is likely to be similar in an individual and one or more of their parents"?

Or is "heritability" the word that comes closest to denoting that scale, and the idea is just to mention the various caveats around it? That's how it seems to me from my reading, but I would love to get input on it from someone with expertise in the field.

Expand full comment

Rationalist-esque office decorations up! Pretty proud of how they turned out:

https://imgur.com/HHVnmp9

https://imgur.com/cDK8fK4

https://imgur.com/9l5NVY7

Expand full comment

Ray Kurzweil on AI on Lex Friedman's podcast. - https://www.youtube.com/watch?v=ykY69lSpDdo

He maintains his views and you can see what he says with Lex on YT.

One thought I had while watching was about not getting lost in the weeds of this or that capability of current language models and other AI or AI related tech.

We are already in the grey area in terms of an AI which could arguably be conscious. We're not clearly across the line and there likely is no such clear line, nor one we could all agree on.

But I'd argue we might find broad agreement that we've already reached some sort of uncanny valley where the top chat bots are able to converse for hours with a regular person and pass some forms of a Turing test.

It is simply an observable fact now that for many a person off the street and even for those with some exposure to AI concepts such as say...a classic science fiction fan....would be unable to reliably tell the difference between the top natural language models and another person chatting with them. This is simply a fact that many people can be 'tricked' and this can happen in an over 1 hour conversation with an AI.

Kurzweil noted this point in his conversation with Lex on what a 'valid' Turning test would look like and he said it would take an AI expert to try every trick they knew in order to confirm it, but that already even many an AI engineer can be fooled if they don't know the specific failings of a given natural language AI.

Whatever broad grey zone of lower achievement in AI and full achievement in AI exists....we've already cross into that area. Can I get a Keanu saying 'Woah...'

Expanding more broadly....the AI is already better than an infant, or a stupid person, or even a person who doesn't speak the same language as me. A chat exchange with another person wouldn't even be possible if we didn't share a common language....unless we used some AI translator or another human to translate.

So in many ways, the bandwidth of my ability to chat with an advanced AI bot is already higher than my ability to speak with the majority of humans on the planet, with whom I do not share a common language. A pedantic point perhaps, but no less true for it being 'silly'.

Expand full comment

Greg Egan has a distinctive and extremly Computer Science-y way of writing scifi that I never saw elsewhere to this degree. I'm talking about Permutation City and Diaspora. The only thing that sorta come close is the Quantum Thief trilogy, but still not quite as masterful as Egan.

People who read permutation City and Diaspora, recomend me something like.

Expand full comment

Does anyone have any recommendations for late-19th century, early 20th century commentary/writings on divorce?

I'm asking because, reading through Lasch, he made a passing comment on late 19th century controversies over divorce. Which is pretty weird because, like, all modern divorce commentary focuses on feminism and the 1960s-70s. So I did a bit of digging and the best source I can find is the 1972 CDC report(1). Haven't read it all but the core takeaway is that yes, there was a probably a rising divorce trend from 1867-1900 and definitely one from 1901-1928. The relevant graph is figure 3 on p.10 and the scare quote is:

"During the first 60 years under consideration the divorce rate increased on the average about 75 percent every 20 years, and if this increase had continued, rates of 2.8 in 1947 and 4.9 in 1967 could have been expected. The year 1947 fell into the period of post-World War II increase, so the observed rate was much higher than expected, but the following year, 1948, it was 2.8. In spite of rapid growth in the divorce rate since 1963, the observed rate for 1967 was still much lower than that which could be expected on the basis of the 1867-1927 experience."

These figures indicate that, rather than the 1960s-70s divorce spike being an outlier, it was actually a return to pre-existing trends tracing back to, basically, industrialization. I don't want to oversell this, especially on something as touchy as feminism/sex/etc but this seemed surprising and underdiscussed. It must have drawn contemporary commentary, and Lasch references such, but I'm not finding any. Does anyone have any info to share or, preferably, contemporary or modern commentaries to recommend for me to read?

(1) https://www.cdc.gov/nchs/data/series/sr_21/sr21_024.pdf

Expand full comment

I asked this late during the last open thread, but I'll try again.

A question to any readers who are in medicine: what is your relationship to radiology results and the radiology department? Do you get to see the images themselves, or have to rely only on the narrative report? Do you - or any specific colleagues of yours - have enough expertise (or think they do) to question the radiologists' conclusions?

Also, do you communicate with a specific radiologist in your hospital, or do you have to outsource to a central source, where you can't actually talk to the person who made the diagnosis or report?

I will also say, as a patient, the fact that I can log into a portal and see my radiology and test results is really nice, even though I've heard physicians don't really like it.

Expand full comment

According to this article

https://www.science.org/doi/10.1126/science.abn7950

Climate change may be potentially far worse than previously thought. Even if we manage to reach the Paris agreement target, this may still be over the tipping point leading to positive feedback and runaway climate change, in contrast to the IPCC report.

Expand full comment

Over the weekend, I saw this article about Sweden and Covid: https://maximumtruth.substack.com/p/sweden-and-covid-unraveling-the-mystery

I've generally thought we need covid to be "over" before we try to draw lessons from it. During covid discussions where way too partisan to be fruitful. The also tended to look at only US data, which isn't helpful as the US population seemed to act in partisan ways. While covid isn't "over", it effects on day-to-day life seem mostly gone. It is probably time to start looking at what countries did and how well it worked.

I'm quite interested in Sweden's approach. Their approach seems very different to any nation I am aware of. At first it seems like trying to take the approach of doing the least restrictions that will get the biggest impact. Then I heard they also discouraged people from wearing masks which seems harmful. So I've been interested to see how well they do and if it would be the approach other nations should consider in the future.

Most of the article makes it pretty clear that it is hard to say. In some comparisons it looks good. In others it looks worse. There is also the issue of relevant comparison countries being less honest about their covid data.

The one interesting part that I think leans clearly in Sweden's favor is the cost-benefit analysis. In computing the days-of-life-saved. Which ranges between 0 and 2.9 days. It seems hard to justify 2 years of lockdowns to save 2.9 days of life.

Expand full comment

'Hippocrates Writings' translated by GER Lloyd. The classic on medicine and source of the four humors theory. 'The human body contains the sanguine, phlegmatic, melancholy and choler'. Sort of means blood, spit, liver bile, puke.

BUT Hippocrates is as innocent of blood types and chemically isolated fluids as of building a '57 Chevy. He fingers crusties and slimes on a sicko and says- Wet or Dry? Hot or Cold? Then combines it as HotWet, ColdWet, HotDry, ColdDry. We took that and invented the Four Humors. It's not his fault! He warned us!

'I am utterly at a loss to see how those who prefer these hypothetical arguments and reduce the science ever cure anyone on the basis of their assumptions. I do not think they have ever found anything purely 'hot' 'cold' 'wet' or 'dry' without it sharing some other qualities.'

He warned us, but we persisted. He was a smart guy who gave good information, useful enough to be recopied and preserved. He became a classic. By Galen's time, as far from Hippocrates as Jonas Salk from Paracelsus, Galen was responding to critics who said 'hot and cold are the terms of bathhouse attendants, not of doctors'. And at some point, I don't find it before the Elizabethan poets and Albrecht Durer (who take it for granted), we started thinking the four humors are psychological types. No bath house indignities, not icky like crusties and slimes from a sicko. The high science of the soul. The dignified classic writing in Classical Greek.

And it works. Once the four humors are etherealized into psychology we have a useful way to type people. As biochemistry improved the medical science of humors went mythical, like phlogiston. But we still type people. She's a moist little piece. Well I'm hot-blooded, check it and see. She has a dry sense of humor. And the OODA loop and DISC character typing still works. And they are still infested with kitsch.

'Want of skill is a poor treasure. It robs a man of contentment and tranquility and leads to cowardice and recklessness, the one a mark of weakness, the other of ignorance'.

Kitsch shows a want of skill.

I have no Greek. I don't know when the four humors became psychological types, but pulling a wild hair from my ass I suspect the Four Apostles medieval mystery plays. When I read my Bible I don't see Matthew, Mark, Luke and John as a comedy of humors, but lots of better Christians who knew their Bibles better did. You always have to dumb down characterization to get it into the actors, much less across the footlights. I have no Latin either and have not read the mystery plays.

Expand full comment

'Englishmen, Frenchmen, Spaniards' by Salvador Madariaga. Madariaga was a Spanish diplomat in the 1920's, using the League of Nations, the Soviet Union, and the rearming Germans against US and British power in Spanish America. (In hindsight Spanish links with Soviets and rightwing Germans probably helped start the Spanish Civil War). The book reads like a fix-up of high table conversations about national personalities- the English Man of Action, the French Man of Reason, the Spanish Man of Passion.

Three out of four classic humors or of the OODA loop. Who gets the fourth humor? I guess anyone outside Madariaga's scheme gets to be Choleric. Soviets, Germans, Americans . . .

Written in a really charming 1920's high table style.

Expand full comment
Sep 19, 2022·edited Sep 19, 2022

Please recommend me a trustworthy and all-around good cryptocurrency exchange, i.e. a website, app, or whatever, where I can exchange different cryptocurrencies (I don't need to exchange them for Fiat money), without KYC/AML bullshit.

In particular, I'm interested in BTC, ETH, USDT TRC20, BUSD BEP20, Tezos, Tron, Monero.

Expand full comment

To those here who might be interested.

https://timself.substack.com/p/the-war-on-cancer

I published a long essay about progress (or lack thereof) in cancer pharmacology in recent years.

Expand full comment

In re anti-aging research: Do we even have good ideas about whether the important thing is to attack the problem directly, or whether the best thing is better tools? Maybe what's needed is improvement in computers (maybe cheaper is needed more than faster), mathematics, physics, chemistry, etc.

Expand full comment

Essential new Substack from Sam Kriss https://samkriss.substack.com/p/the-internet-is-already-over

Expand full comment
(Banned)Sep 19, 2022·edited Sep 19, 2022

Scott Aaronson has a dream : https://scottaaronson.blog/?p=6718

I want that made into a 1000-page scifi novel.

Expand full comment

I took a few lessons in piano performance as a kid and messed around with it for a while before setting it aside for decades. Nothing even remotely impressive - I got to the point of doing a decent job with stuff like Valse 69 Nr 2 and I think it was the Pathetique that finally did me in.

I'm thinking about dabbling in it again as an exercise in triggering and sustaining flow and concentration. The goal is to try to get as far up the technical complexity scale as possible, with a time investment of something like seven hours a week. Anyone got any hacks to make this go a little more efficiently?

Expand full comment

I find myself in the odd position now of working for a store branch that earns over $1m in revenue each week, translating to $PROFIT of unknown but assumed-positive quantity...and yet things have deteriorated so far that managers are now offering random grunt employees field promotions to manager. (They're all wise enough to decline, it's a bad deal. Skin in the game doesn't let frontline workers lie to themselves like that.) Normally, promotion is a months- or years-long arduous quest with formal processes, applications to write to the regional directors, peer review...etc, etc.

Should I be shrugging this off as Just Some Unusual Times In A String Of Continued Boondoggles, or actually possibly worried about continued employment? I can't imagine *every* large chain grocer in America is undergoing this kind of absurdity. Don't we have riproaring full employment these days? It's not like the company has never hired bosses externally rather than doing internal promotions...

Expand full comment

In the past year or so, the mainstream media has started writing "Black" with the "B" capitalized when referring to a black aka African-American person while still referring to a white person as "white" with the "w" in lowercase.

I don't know the reasoning for this, but my best guess is that some influential academics have decided that black should be considered an ethnicity like Hispanic or Jewish and on those grounds should be capitalized as a Proper Noun, whereas "white" should not be considered an ethnicity because it is a false social construct or something (even though, if so, why do they still use the word at all and what does it mean?)

The whole thing seems ludicrous and even makes me angry and want to cancel subscriptions because I feel insulted.

Does anyone here want to attempt to justify this style change in the mainstream press?

Expand full comment
Sep 19, 2022·edited Sep 19, 2022

Should the USA ever bluff in foreign policy? I just read that Biden says the USA would send US troops to defend Taiwan if China were to invade. If that turns out to be a bluff under Biden's watch, doesn't that in turn weaken NATO?

For that matter, hasn't the USA already lost credibility on such things after Obama's "red line" in Syria was crossed without meaningful consequences?

Seems to me one should only bluff when one holds a weak hand. Does the USA have a weak hand?

Expand full comment

I think Scott was premature in claiming victory, but I don't see anything nefarious in Google's reluctance to give Marcus access to its models; it has only his word that he is an objective inquirer, and he has been quite explicit in stating that he thinks the field of machine learning is currently headed down a cul-de-sac. One could understandably get the impression that he is in thrall to a preconception, unwilling to give Imagen and PARTI a fair hearing. Wanting to avoid bad PR from a biased or disingenuous observer is not at all dishonest.

Expand full comment

I am pushing 60 and somewhat pre-diabetic, so I can probably get metformin prescribed. I don't really need it, but should I, anyway, for the supposed longevity benefits?

Expand full comment

Does anyone have any recommendations for a psychiatrist in the Northern Virginia area who is taking new patients and takes CareFirst insurance? I've looked on the CareFirst website, but their rating system is not working.

Expand full comment

Book noted. I will leave this for Scott, or somebody else with psychiatry chops to read and review:

"‘Strangers to Ourselves’ Review: The Mystery of Mental Illness" By Rachel Aviv

Review by Elizabeth Winkler

https://www.wsj.com/articles/strangers-to-ourselves-book-review-psychology-the-mystery-of-mental-illness-11663340407

Sept. 16, 2022

"When Rachel Aviv was 6, she stopped eating. Psychiatrists diagnosed her with anorexia nervosa, a disorder typically brought on by reading magazines that present thinness as the ideal of femininity. But young Rachel was only just learning to read; she didn’t yet have a concept of ideal femininity. Her case was the earliest recorded onset of anorexia in America. During her hospitalization, she met other girls in the anorexia unit, including Hava, a 12-year-old whose circumstances mirrored her own. Both girls came from Jewish families (Rachel got the idea of fasting from Yom Kippur); had parents engaged in a long, hostile divorce; and heard derogatory jokes about obese people. But while Rachel soon began eating again and returned to normal life, Hava became a “career” anorexic—in and out of hospitals her entire life until her premature death in her early 40s.

"Why do some people recover from mental illness and others don’t? Why doesn’t having insight into one’s condition provide a cure? By all accounts, Hava at 12 had excellent insight, precociously recorded in her journals; at 6, Rachel had none. In “Strangers to Ourselves: Unsettled Minds and the Stories That Make Us,” Ms. Aviv, now a staff writer for the New Yorker, draws on her own brush with mental illness to explore the limits of psychiatric frameworks for understanding minds in crisis."

Expand full comment
Sep 19, 2022·edited Sep 19, 2022

I see Gary Marcus, I have to post my standard reference on why compositionality is the wrong way to try and understand how human language works: https://arxiv.org/abs/1904.03991

tl;dr: Meaning is discriminative, not compositional.

Also, having just skimmed his post, I don't like his tone, specifically how he pulls rank as a scientist. I also happen to be a scientist, technically -- a linguist, to boot -- and I've been thinking a lot recently about how we as a research community managed to get sidetracked by compositionality for a hundred years, when people who actually care about clear communication and avoiding misunderstandings inevitably and intuitively reach the conclusion that meaning is discriminative (cf. Hartley and Shannon, referenced in the linked paper as "the founding fathers of information theory"), even though they might not use these exact words (cf. e.g. https://www.lesswrong.com/posts/57sq9qA3wurjres4K/ruling-out-everything-else, but also a lot of what Eliezer Yudkowsky writes on communication in the sequences).

My answer is basically skin in the game / Nassim Taleb. So I get particularly irked when Marcus chides Scott for not being scientific enough, when Scott is quite clear from the get go that he's in a somewhat different -- skin-in-the-game -- paradigm. He made a bet with very specific terms, someone accepted it, and then he won that bet. Nothing more, nothing less. Marcus is free to discuss what the result means, why he personally wouldn't have made that particular bet, ideally propose a bet he *would* be willing to accept. Instead, he's all condescending that this is not “real science”. Well maybe, but if “real science” is the reason Marcus still believes in compositionality...

Expand full comment
Sep 19, 2022·edited Sep 19, 2022

Substack in it's infinite wisdom has given me three one-month subscriptions to Astral Codex Ten that I can give away to anyone I'd like.

If anythree here are curious about what's behind the paywall, reply with your ROT13 email

Expand full comment

Much of Christianity seems delusional to me, but professed Christians seem far happier then average (though it may be that it’s their happiness that makes them Christians rather than the other way around), and there’s plenty of studies that show that (except for Orthodox Christians among the denominations https://theconversation.com/amp/are-religious-people-happier-than-non-religious-people-87394 )

If I could somehow “catch belief” instead of being as miserable as I am, why not?

Expand full comment

> I want to make it clear that whatever the merits of my bet or his arguments, Google did not “snooker” me. They had no part in this: I went around begging for someone to run my prompts through PARTI and Imagen, one of their employees asked their bosses’ permission and then agreed to do so, and ran them exactly as I asked.

But that's what getting snookered looks like. Where are you disputing the idea that they snookered you? How do you think they made the decision to give you what you wanted?

Expand full comment

Pretty sure to the extent a corporate behemoth like Google ever responds to the requests of "random strangers" it's because the suits have thought it out carefully and believe it's in their business interest to do so. They're not naive, or idiots, and they're not doing charity work.

That is, they weren't doing favors for you, they think *you* were doing favors for *them* -- and that would be Marcus's point.

I think it's a reasonable caution to keep in mind. Google isn't a bunch of plucky nerds fiddling around in someone's garage, and hasn't been for many years. It's a great big corporation run by lawyers, accountants, business executives and wolves from Wall Street, just like Exxon Mobil, Procter & Gamble, or Ford. If Ford decided to give away one of what they describe as their totally awesome paradigm-shifting electric vehicles to a random amateur columnist, instead of to the cynical pros who run The Truth About Cars -- yeah, I'd be a little bit cautious about their motives. No difference here.

Expand full comment

What should Mexico (or the US) be doing about Mexico's increasingly violent drug cartels? Has any country ever successfully 'solved' organized crime gangs that have sort of morphed into paramilitary/terrorist organizations? Mexico's specific problem is a bit sui generis in that they're literally nextdoor to the world's largest market for illegal drugs, a country with both incredible wealth and incredible demand.

My impression as someone who doesn't follow the subject super-closely is that cracking down on the largest cartels just increases the violence, as now smaller orgs are competing for the market share and fighting each other. Could Mexico..... quietly tolerate the largest cartels (I mean to some degree, not officially) in exchange for their tacit agreement to tamp down violence? If not, has someone or some group of people written an intelligent thesis on what exactly Mexico should be doing? US media is quite European-focused, I feel like I don't read a great deal about Latin American problems, even those that are literally nextdoor

Expand full comment

How might the average person in a developed country apply the ideas behind Effective Altruism to improve their own life? What is the low-hanging fruit? Is this a harder question than how to help people in developing countries? Apologies if this is well-worn ground; I'm fairly new to these ideas.

Expand full comment

In his new book, Bryan Caplan says the most intellectually respectable feminists are left wing labor economists. Do you know who exactly he is talking about ? I would love to read them

Expand full comment

Hi ACX readers!

I'm looking for recommendations for novels about heroes.

By "heroes," I mean people who have admirable qualities of character and who are struggling against great odds toward a worthy goal.

Most stories that meet these criteria are a) for children or b) set in a fantasy or sci-fi setting. I love fantasy and sci-fi stories - The Lord of the Rings, Star Wars, Ken Liu's Dandelion Dynasty saga, Pierce Brown's Red Rising series - but I'm looking for some hero stories set in our world (either the present day or historical fiction).

As an example, I consider Ken Follett's "The Pillars of the Earth" and its sequels to be a hero story, broadly defined.

Does anyone have other recommendations? Thank you in advance!

Expand full comment
Sep 18, 2022·edited Sep 18, 2022

Why isn't anti-aging research the number one or two EA funding topic?

I can understand prioritizing AI safety above it, but not anything else. I think it's self-evident that aging causes more human suffering than any other single cause. I'd be surprised if it caused less than an order of magnitude more suffering than the next-largest cause. And I think it's the only major basic research area that the U.S. government has explicitly declared it will not spend money on (although that may change; the National Institute of Aging hasn't announced any change of policy, but has recently begun sneaking non-disease-specific topics into its grant solicitations.)

If you don't think aging is the largest cause of human suffering, you're young and all your friends are young.

Expand full comment

When did the American right start using Marxist terms in their cultural critique? Eg Thiel describing the “superstructure” of California

Expand full comment

Hello everyone,

I'm re-posting a survey proposal, as the number of responses I received a couple of weeks ago is just a tad too low for me to do the statistical analyses. If you have a few minutes to spare, would you like to take a short survey about your memory and thought process?

A while back I was extremely interested in an ACX book review (one of the non-winners but a finalist!), and it made me wonder if a specific hypothesis about memory and consciousness of self was true. I made a very short survey to try to answer the question and posted it on the ACX subreddit. The responses were very interesting and suggested, upon analysis, that my hypothesis was... more or less true, but that there was something else more important going on.

I also received some great suggestions and another related hypothesis to test. With these suggestions and taking into account the results of the first survey, I have created a new version of the survey, which should take about 7 minutes to complete. Here is the link, thank you so much if you choose to participate!

https://docs.google.com/forms/d/e/1FAIpQLSfj2nV0JvmBC0hkDPaR4oP-mDoMg4tJEq4dj0RlTwFVlPf65w/viewform

PS: If you already took the previous survey on the ACX subreddit, it would be wonderful if you took this one too, as there are many new questions.

PPS I plan to publish the results of both surveys on the subreddit in a few weeks.

Expand full comment
Sep 18, 2022·edited Sep 18, 2022

Also: Scott, for next year have you considered running an essay contest instead of a book review contest, seeing how even among finalists, some were clearly just opinion pieces disguised as book reviews?

I recognize that the book review thing is obviously an intellectual effort filter, but it can cut both ways - it could prevent people who aren't deep enough to have anything interesting to say from submitting, but OTOH it could also filter out people with genuinely interesting ideas who are too lazy to look for books they could "review" to excuse their submissions. IDK which way this balances out, but it seems like a worthwhile and cheap enough experiment to run.

Expand full comment
Sep 18, 2022·edited Sep 18, 2022

It occurred to me that plot holes and nitpicks may essentially be failures of compositionality in human reasoning: by the time you about Chekhov's gun in chapter 30, you've forgotten that you established Chekhov's dislike of sudden loud noises in chapter 5, or the whole archery subplot in the last book.

This seems to be a recurring pattern in AI-scaling debates: claiming that scaling can't get much further given the nature of the errors being made, without noticing that humans routinely make the same errors, further into the details of a composition. Sure, maybe no amount of parameters will ever truly overcome and defeat that error. So what? Once the errors become long -range/subtle enough, you'll fail to notice them anymore, just like your favourite iron-man/pop-song/scott-essay.

Expand full comment

You all may enjoy my interview of Charles Mann, author of The Wizard and the Prophet, 1491, & 1493. https://www.dwarkeshpatel.com/p/charles-mann

We discuss, why Native American civilizations collapsed and why they failed to make more technological progress, why he disagrees with Will MacAskill about longtermism, why there aren’t any successful slave revolts, how geoengineering can help us solve climate change, why Bitcoin is like the Chinese Silver Trade, and much much more!

Btw, The Wizard and the Prophet was reviewed during last year's book review contest: https://astralcodexten.substack.com/p/your-book-review-the-wizard-and-the

Expand full comment

Hi all! Trevor Klee here. Thanks to Scott for signal-boosting the crowdfunding campaign. A couple things:

1. I wouldn't call this a charitable endeavor exactly. We've already attracted significant investment, have had licensing negotiations with large pharmaceutical companies, and have a reasonably clear path towards becoming a highly profitable ($100+ million in revenue) endeavor in the next 5 years or so. Still, I do believe that this is important work that could have significant benefits for the world, which is why I'm doing it.

2. We have a full investing pitch on the crowdfunding page: https://wefunder.com/highway.pharmaceuticals/ . We are in the "testing the waters"/precommitment phase until September 27, at which point all commitments need to be final. Full legal disclosures at the link.

3. If you're an accredited investor and would like to invest $25k+, please contact me at trevor [at] highwaypharm.com . We can do a SAFE directly, instead of the crowdfunding.

Expand full comment

I know Brasilia is generally not well-liked by actual Brasilians and is frequently touted as an example of failed modernistic urban planning, but I read one specific detail about it - that its roads were built in a way that deliberately avoids intersections. To me this sounds like a great idea, because I really don't like stopping at red lights. Did that specific aspect of Brasilia also fail? If yes, how?

Expand full comment

Alexandros’s entire series is worth the read. I’m not sure the one linked above is the crux though. Of those published to date I might point to this one as key: https://doyourownresearch.substack.com/p/the-potemkin-argument-part-14-achilles?r=7ezsm&utm_medium=ios

Expand full comment

C'mon, you also wrote, "I won my bet after three months." That's a hell of a lot stronger than, "without wanting to claim that Imagen has fully mastered compositionality, I think it represents a significant enough improvement to win the bet, and to provide some evidence that simple scaling and normal progress are enough for compositionality gains."

Expand full comment

It seems to me that the lack of development of any new cities or towns of any significance in the western world in recent decades is a staggering sign of decline of community in the physical sense. Regardless of where the metaverse takes us, I think we can pretty much call it for the types of mid- to large-scale communities seen in the past few millennia.

Expand full comment
founding
Sep 18, 2022·edited Sep 18, 2022

So I started reading Why We Sleep, which is a wildly popular pop-science book about what happens in our bodies and brains when we sleep.

I've been hearing though that "[the author] is a grifter" or that the book is bunk, and people keep linking me to this supposed debunking: https://guzey.com/books/why-we-sleep/ . I find the points there interesting and indeed suspicious, but largely nit-picky. I'm planning to do some personal spot-checks of my own on some specific factual points in the book, but until I do, this all makes me wonder whether I'm wasting my time.

I'm wondering if anyone else has any significant knowledge to share about this; genuinely just want to know people's thoughts, to the extent they're backed by evidence and not just hearsay. Is the book worth reading or should I believe the skeptics on this one?

Expand full comment

I’m organizing several conferences for entrepreneurs, scientists and innovators in Próspera, the startup city.

Scholarships for flights and accommodation options are available.

More here: https://infinitafund.com/scholarship

And here:

Prospera Healthtech Summit, September 23-25, 2022: https://infinitafund.com/healthtech2022

Prospera Edtech Summit, October 28-30, 2022: https://infinitafund.com/edtech2022

Prospera Fintech Summit, November 18-20, 2022: https://infinitafund.com/fintech2022

Expand full comment
Comment deleted
Expand full comment
Comment deleted
Expand full comment
Comment deleted
Expand full comment