Astral Codex Ten

Vice

12 min read

Vice

Article

Vice is a recurring publication in the Astral Codex Ten archive, appearing 4 times across 4 issues between January 21, 2021 and September 11, 2025. The archive places it in contexts such as “She granted an interview to a Vice reporter under the condition that he not reveal some sensitive details”; “And Vice (why is it always Vice?) is somehow even more melodramatic”; “Whenever British tabloids, Vice, and the FDA all hate a thing”. It most often appears alongside Balaji Srinivasan, Congress, Eliezer Yudkowsky.

Metadata

  • Category: Publications
  • Mention count: 4
  • Issue count: 4
  • First seen: January 21, 2021
  • Last seen: September 11, 2025

Appears In

Source Context

Recovered passages from the original issue text. When the raw archive preserved outbound links inside the source passage, they are listed directly under the quote.

Still Alive
January 21, 2021 · Original source
513,000 people read my blog post complaining about the New York Times' attempt to dox me (for comparison, there are 366,000 people in Iceland). So many people cancelled their subscription that the Times' exasperated customer service agents started pre-empting callers with "Is this about that blog thing?" A friend of a friend reports her grandmother in Slovakia heard a story about me on Slovak-language radio.
I got an email from a former member of the GamerGate movement, offering advice on managing PR. It was very thorough and they had obviously put a lot of effort into it, but it was all premised on this idea that GamerGate was some kind of shining PR success, even though as I remember it they managed to take a complaint about a video game review and mishandle it so badly that they literally got condemned by the UN General Assembly. But it's the thought that counts, and I am humbled by their support.
As for the Times' mistakes: I think they just didn't expect me to care about anonymity as much as I did. In fact, most of my supporters, and most of the savvy people giving me advice, didn't expect me to care as much as I did. Maybe I should explain more of my history here: back in the early 2010s I blogged under my real name. When I interviewed for my dream job in psychiatry, the interviewer had Googled my name, found my blog, and asked me some really pointed questions about whether having a blog meant I was irresponsible and unprofessional. There wasn't even anything controversial on the blog - this was back in the early 2010s, before they invented controversy. They were just old-school pre-social-media-era people who thought having a blog was fundamentally incompatible with the dignity of being a psychiatrist. I didn't get that job, nor several others I thought I was a shoo-in for. I actually failed my entire first year of ACGME match and was pretty close to having to give up on a medical career. At the time I felt like that would mean my life was over.
Shilling For Big Mitochondria
March 02, 2021 · Original source
The British tabloids had a field day with this. The Daily Mail talks about being BOILED ALIVE BY INTERNET SLIMMING PILLS! And Vice (why is it always Vice?) is somehow even more melodramatic:
Inline links: talks about, even more melodramatic
Whenever British tabloids, Vice, and the FDA all hate a thing, I’m inclined to feel at least a little fondness towards it. So is there a case for 2,4-dinitrophenol? I think the case would look like: sure, it has a very low therapeutic index. Sure, if you take just a few times the recommended dose, you could die. But if you very carefully take exactly the recommended dose, you probably won’t. You could give it out like methadone, in a way that makes it impossible to overdose. Patients might still get cataracts. But cataracts are treatable, or at least more treatable than some of the complications of obesity.
Prospectus On Próspera
April 14, 2021 · Original source
What’s to stop this from getting really weird? Can I have my house be in Próspera while my neighbor’s house is in regular Honduras? My kitchen in Próspera but my dining room in regular Honduras? I asked Trey, who said HPI has to guarantee provision of services to everywhere within Próspera; they wouldn’t expect providing services to somebody’s dining room to be worth it, and so they would turn down those kinds of frivolous applicants. Their plan is for a network of hubs, each city-sized and relatively convenient to provide for. They’ve already selected an area near the city of La Ceiba as their first satellite, and eventually want to expand across Honduras’ north coast.
Oliver Porter is the architect of the Sandy Springs model, named after the city of Sandy Springs, Georgia. This is an interesting rabbit hole in itself - Sandy Springs incorporated in a hurry, didn’t have time to create its own city services, hired private corporations to do everything, then liked it so much that they advertised themselves as a libertarian paradise and a model for everyone else to follow (though see here for another perspective).
Inline links: a model for everyone else to follow, here
They’re especially focusing on Honduran professionals and remote workers - the more remote, the easier a time they’ll have moving to a new city on an island. But they’re also interested in poorer Hondurans looking for construction and service jobs, and expats after a slice of paradise on a tropical beach.
Book Review: If Anyone Builds It, Everyone Dies
September 11, 2025 · Original source
Is this convergent evolution? IABIED has three sections. The first explains the basic case for why AI is dangerous. The second tells a specific sci-fi story about how disaster might happen, with appropriate caveats about how it’s just an example and nobody can know for sure. The third discusses where to go from here. II. Does the world really need another ‘The Case For Why AI Could Be Dangerous’ essay? On the one hand, definitely yes. If you’re an “infovore”, you have no idea how information-starved the general public is (did you know 66% of Americans have never used ChatGPT, and 20% of Americans have never even heard of it?). Probably a large majority of people don’t know anything about this. Even people who think they know the case have probably just heard a few stray sentences here or there, the same way “everyone knows” about the Odyssey but only a few percent of people have so much as read one line of its text. So yes, exposing tens of thousands of people to a several-chapter-length presentation of the key arguments is certainly valuable. Even many of you readers are probably in this category, and if I were a better person I would review it all here in depth. Still, I find I can’t bring myself to do this, on the grounds that it feels boring and pointless. Why? The basic case for AI danger is simple. We don’t really understand how to give AI specific goals yet; so far we’ve just been sort of adding superficial tendencies towards compliance as we go along, trusting that it is too dumb for mistakes to really matter. But AI is getting smarter quickly. At some point maybe it will be smarter than humans. Since our intelligence advantage let us replace chimps and other dumber animals, maybe AI will eventually replace us. There’s a reasonable answer to this case. It objects to chaining many assumptions, each of which has a certain probability of failure, or at least of taking a very long time. If there’s an X% chance that getting smarter-than-human AI takes N years, and a Y% chance that it takes P years for the smart AI to diffuse across the economy, and a Z% chance that it takes Q years before the AI overcomes humans’ legacy advantage and becomes more powerful than us - then maybe you can find good odds that the danger point is a century plus away. And in a century, maybe we’ll have better alignment tech, or at least a clearer view of the problem. Why worry about vague things that might or might not happen a century from now? The problem with this is that it’s hard to make the probabilities work out in a way that doesn’t leave at least a 5-10% chance on the full nightmare scenario happening in the next decade. You’d have to be a weird combination of really good at probability (to know how to deploy enough epicycles to defuse the argument) and really bad at probability (to want to do this). There aren’t that many people who are in this exact sweet spot of probabilistic (in)competence. So everyone else just deploys insane moon epistemology. Some people give an example of a past prediction failing, as if this were proof that all predictions must always fail, and get flabbergasted and confused if you remind them that other past predictions have succeeded. Some people say “This one complicated mathematical result I know of says that true intelligence is impossible,” then have no explanation for why the complicated mathematical result doesn’t rule out the existence of humans. Some people say “You’re not allowed to propose that a catastrophe might destroy the human race, because this has never happened before, and nothing can ever happen for the first time”. Then these people turn around and panic about global warming or the fertility decline or whatever. Some people say “The real danger isn’t superintelligent AI, it’s X!” even though the danger could easily be both superintelligent AI and X. X could be anything from near-term AI, to humans misusing AI, to tech oligarchs getting rich and powerful off AI, to totally unrelated things like climate change or racism. Drunk on the excitement of using a cheap rhetorical device, they become convinced that providing enough evidence that X is dangerous frees them of the need to establish that superintelligent AI isn’t. Some people say “You’re not allowed to propose that something bad might happen unless you have a precise mathematical model that says exactly when and why”. Then these people turn around and say they’re concerned about AI entrenching biases or eroding social trust or doing something else they don’t have a precise mathematical model for. There are only a few good arguments against any given thesis. But there are an infinite number of insane moon arguments. “Calvin Coolidge was the Pope, therefore your position is invalid” - how do you pre-emptively defend against this? You can’t. Since you can never predict which insane moon argument a given person will make, and listing/countering every possible insane moon argument makes you sound like an insane moon person yourself, you just sort of give up - or, in Eliezer’s case, take a several year break to teach people epistemology 101. Why do these discussions go so badly? I am usually against psychoanalyzing my opponents, but I will ask forgiveness of the rationalist saints and present a theory. I think it’s because, if it’s true, it changes everything. But it’s not obviously true, and it would be inconvenient for it to change everything. Therefore, it must not be true. And since most people refuse to use this snappy and elegant formulation, they search for the closest thing in reasoning-space that feels like it gets at this justification, and end up with things like “well you need to prove all of your statements mathematically”. Lest I sound too dismissive, I notice myself reasoning this way all the time. The easiest examples I can think of right now: Some people claim that human sperm count is declining, and in ~20 years it will be so low that people cannot conceive naturally. If this were true it would change everything and we should stop what we’re doing and deal with it right now (see here for more). But this would be inconvenient. So we assume it’s probably false, or at least that we can deal with it later.
Inline links: https://substackcdn.com/image/fetch/$s_!Mf3D!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1905faaf-ea4f-4e84-9fb0-25a48170f7ff_534x153.png, take a several year break to teach people epistemology 101, usually against psychoanalyzing my opponents, rationalist saints, see here for more
Eliezer Yudkowsky, at his best, has leaps of genius nobody else can match. Fifteen years ago, he decided that the best way to something something AI safety was to write a Harry Potter fanfiction. Many people at the time (including me) gingerly suggested that maybe this was not optimal time management for someone who was approximately the only person working full-time on humanity’s most pressing problem. He totally demolished us and proved us wronger than anyone has ever been wrong before. Hundreds of thousands of people read Harry Potter and the Methods of Rationality, it got lavish positive reviews in Syfy, Vice, and The Atlantic, and it basically one-shotted a substantial percent of the world’s smartest STEM undergrads. Fifteen years later, I still meet bright young MIT students who tell me they’re working on AI safety, and when I ask them why in public they say something about their advisor, and then later in private they admit it was the fanfic. Valuing the time of the average AI genius at the rate set by Sam Altman (let alone Mark Zuckerberg), HPMOR probably bought Eliezer a few billion dollars in free labor. Just a totally inconceivable level of victory.