Astral Codex Ten Podcast

Latest Episodes:

Sam Altman Wants $7 Trillion Feb 16, 2024

Show notes

In one sense, this isn't news. Everyone wants $7 trillion. I want $7 trillion. I'm not going to get it, and Sam Altman probably won't either.

Still, the media treats this as worthy of comment, and I agree. It's a useful reminder of what it will take for AI to scale in the coming years.

The basic logic:

https://www.astralcodexten.com/p/sam-altman-wants-7-trillion

ACX Grants Results 2024 Feb 16, 2024

Show notes

Thanks to everyone who participated in ACX Grants, whether as an applicant, an evaluator, or a funder.

The best part of ACX Grants is telling the winners they won, which I'll do in a moment. The worst part of ACX Grants is telling the non-winners they didn't win. If I wasn't able to give you a grant, it doesn't mean I hate your project. Sometimes I couldn't find the right evaluator to confirm that you were legit. Sometimes I sent your project to foundations or VCs who I thought it would be a better match for, or wanted to leave it as a test case for the impact market. Most of the time, I just didn't have enough money1, and I spent what I had according to my own imperfect priorities.

(In particular, I wasn't able to fully evaluate several AI alignment grants and had to pass on them; if this is you, consider applying to OpenAI's Superalignment Fast Grants before February 18.)

If your name is below, you should have received an email with further information. If you didn't, email me at scott@slatestarcodex.com, and include the phrase "this is a genuine non-spam message" in the text. Unless my email specifically mentioned you as an exception, Manifund will be handling payments and you'll hear from them soon.

This year's winners are:

https://www.astralcodexten.com/p/acx-grants-results-2024

Evolution Explains Polygenic Structure Feb 09, 2024

Show notes

We've been gradually working our way through the conversation around E. Fuller Torrey's concerns about schizophrenia genetics - last week we had It's Fair To Describe Schizophrenia As Probably Mostly Genetic, the week before Unintuitive Properties Of Polygenic Disorders. Here are two more arguments Torrey makes that we haven't gotten to:

Studies have failed to find any schizophrenia genes of large effect. If schizophrenia is genetic, it must be caused of thousands of genes, hidden in the most obscure corners of the genome, each with effects too small to detect with current technology. This seems less like the sort of thing that happens naturally, and more like the sort of thing you would claim if you wanted to make your theory untestable.
Schizophrenia is bad for fitness, so if it were genetic, evolution would have eliminated those genes.

In the comments of the Unintuitive Properties post, Michael Roe points out that one of these mysteries solves the other:

https://www.astralcodexten.com/p/evolution-explains-polygenic-structure

You Don't Hate Polyamory, You Hate People Who Write Books Feb 09, 2024

Show notes

Yesterday I criticized The Atlantic's recent invective against polyamory (subscriber-only post, sorry). Today I want to zoom away from the specific bad arguments and examine the overall form of the article.

The overall form was: "I read a memoir about polyamory, everyone involved seemed awful and unhappy, and now I hate polyamorous people." This is a common pattern. Sometimes, if someone's very careful, they read three or four books about polyamory. Everyone in all the books is awful and unhappy. Then they conclude they hate polyamorous people.

But this is an unfair generalization. They should hate people who write books.

https://www.astralcodexten.com/p/you-dont-hate-polyamory-you-hate

It's Fair To Describe Schizophrenia As Probably Mostly Genetic Feb 09, 2024

Show notes

Famous schizophrenia researcher E. Fuller Torrey recently wrote a paper trying to cast doubt on whether schizophrenia is really genetic. His exact argument is complicated, but I feel like it sort of equivocates between "the studies showing that schizophrenia are genetic are wrong" and "the studies are right, but in a philosophical sense we shouldn't describe it as 'mostly genetic'".

Awais Aftab makes a clearer version of the philosophical argument. He's not especially interested in debating the studies. But he says that even if the studies are right and schizophrenia is 80% heritable, we shouldn't call it a genetic disease. He says:

Heritability is "biologically vacuous" (Matthews & Turkheimer, 2022), and I think we would be better off if more of us hesitated to assert that schizophrenia is a "genetic disorder" based predominantly on heritability estimates.

I think about questions like these through the lens of avoiding isolated demands for rigor. There are always complicated ways that any statement is false. So the question is never whether a statement is perfectly true in every sense. It's what happens when we treat it fairly, using the same normal criteria we use for everything else.

https://www.astralcodexten.com/p/its-fair-to-describe-schizophrenia

Seems Like Targeting Feb 09, 2024

Show notes

Recently Claudine Gay resigned as President of Harvard over plagiarism accusations and a fumbled Congressional testimony on anti-Semitism.

The plagiarism was discovered by conservative journalists Chris Rufo and Chris Brunet. It would be quite a coincidence for them to find it at exactly the moment Gay was already under attack for her anti-Semitism testimony. More likely, they either:

Found it a while ago, and kept it in reserve for a time when Gay was in the news
Or were angry about Gay's testimony, looked for dirt on her, and found it.

I think this is obvious to everyone, but I hadn't seen anyone make it explicit, and I think it should be.

I'm not criticizing Rufo and Brunet. Investigative journalism is important, they found a real scandal, and they have every right to bring it to light.

https://www.astralcodexten.com/p/seems-like-targeting

Mantic Monday 1/29/24 Feb 09, 2024

Show notes

Election problems // Trump odds // AI worlds

https://www.astralcodexten.com/p/mantic-monday-12924

The Psychopolitics Of Trauma Jan 27, 2024

Show notes

Everyone knows politics makes people crazy. But what kind of crazy? Which page of the DSM is it on?

I'm only half joking. Psychiatrists have spent decades developing a whole catalog of ways brains can go wrong. Politics makes people's brains go wrong. Shouldn't it be in the catalog? Wouldn't it be weird if 21st century political extremists had discovered a totally new form of mental dysfunction, unrelated even by analogy to all the forms that had come before?

You'll object: politics only metaphorically "makes people crazy"; we just use the word "crazy" here to mean "irrational" or "overly emotional". I'm not sure that's true. Here are some stray findings that I think deserve to be synthesized:

https://www.astralcodexten.com/p/the-psychopolitics-of-trauma

Some Unintuitive Properties Of Polygenic Disorders Jan 27, 2024

Show notes

E. Fuller Torrey recently published a journal article trying to cast doubt on the commonly-accepted claim that schizophrenia is mostly genetic. Most of his points were the usual "if we can't name all of the exact genes, it must not be genetic at all" - but two arguments stood out:

Even though twin studies say schizophrenia is about 80% genetic, surveys of twin pairs show that if one identical twin has schizophrenia, the other one only has a 15% to 50% chance of having it.

The Nazis ran a eugenics program that killed most of the schizophrenics in Germany, eliminating their genes from the gene pool. But the next generation of Germans had a totally normal schizophrenia rate, comparable to pre-Nazi Germany or any other country.

I used to find arguments like these surprising and hard to answer. But after learning more about genetics, they no longer have such a hold on me. I'm going to try to communicate my reasoning with a very simple simulation, then give links to people who do the much more complicated math that it would take to model the real world.

https://www.astralcodexten.com/p/some-unintuitive-properties-of-polygenic

Should The Future Be Human? Jan 27, 2024

Show notes

Machine Alignment Monday 1/22/24

Business Insider: Larry Page Once Called Elon Musk A "Specieist":

Tesla CEO Elon Musk and Google cofounder Larry Page disagree so severely about the dangers of AI it apparently ended their friendship.

At Musk's 44th birthday celebration in 2015, Page accused Musk of being a "specieist" who preferred humans over future digital life forms [...] Musk said to Page at the time, "Well, yes, I am pro-human, I fucking like humanity, dude."

A month later, Business Insider returned to the same question, from a different angle: Effective Accelerationists Don't Care If Humans Are Replaced By AI:

A jargon-filled website spreading the gospel of Effective Accelerationism describes "technocapitalistic progress" as inevitable, lauding e/acc proponents as builders who are "making the future happen […] Rather than fear, we have faith in the adaptation process and wish to accelerate this to the asymptotic limit: the technocapital singularity," the site reads. "We have no affinity for biological humans or even the human mind structure."

I originally thought there was an unbridgeable value gap between Page and e/acc vs. Musk and EA. But I can imagine stories that would put me on either side. For example:

https://www.astralcodexten.com/p/should-the-future-be-human