dash2

It's a subliminal warning. "On the trolley problem, this model will try to kill you and make it look like an 'accident'."

I think lots of Catholics don't take that bit too literally though, eh? Like they might practice contraception despite what various Popes have said.

It's easier to play with if the documentation and initial selling page is easy for humans to read!

> equivalent

Do you think it would be worth writing the landing page yourself? The concepts seem interesting but it's painful to wade through stuff like "One memory model: stacks, and scope exit is the free".

I'm sorry, I've read about thirty paragraphs of this, and I still would like to know: what does Raven do? Also btw, I don't know what Bottles is, or Eagle, or a PE file. Maybe this is just for a very specialized target audience, and if so, my bad.

The US and China are not the same country and do not do the exact same stuff. To take some easy examples, the US government never imposed forced abortions on large numbers of women, nor has it invaded its neighbours (in the past century at least), nor has it starved many of its own citizens due to misguided economic policies, nor does it regularly and systematically violate habeas corpus (though perhaps some ICE policies might come close), nor does it censor its citizens.

Aren't llms now outperforming humans in prediction markets? If so, that seems as if they would probably be quite good at confidence scoring.

> I expect it to cost somewhere in the neighborhood of an experienced dev because that’s the alternative.

That's only true if a single company owns the frontier, otherwise the alternative is "a rival frontier model" and the cost will be the marginal price of the compute + amortized training costs.

>For example, “You can’t beat the market, dollar-cost average into index funds!” is the kind of viral story that will earn you nods of approval from millennials and upvotes on Reddit.

>Pattern matching, on the other hand, can drive you to make moves like an advanced chess engine. Moves that others find alien and incomprehensible.

Sure. Index funds is a viral story which has a lot of rigorous data behind it. As for the advanced chess engine of intuition... ever heard of Fooled By Randomness?

So it seems worth vibecoding a survival analysis for bugs introduced. But not quite a survival analysis because we don't even know the bugs exist until the event of their discovery.... Is there a publicly available dataset for this, with a URL?

I would not start using this until it is mature enough for the developers to write their own homepage.

What's missing from the OSI definition in this case? How do you think it is limited compare to the standard usage?

The UK has the opposite problem. You don't get overtreated, but you may get undertreated - sometimes it is impossible to find a NHS dentist at all. Indeed, I went private because I literally couldn't find one after my old one sacked me. The private dentists probably have the same problems as described here. But I luckily found a small, nice, ethical one who was recommended by people in town I trust. Literal word-of-mouth still beats any algorithm I know for this kind of question.

If you chose at random and it guessed at random, it would get 32% about 1% of the time. So worse than chance, but not surprising that it happened to at least one HN reader?

I'm talking about now not 100 years ago. Your last point doesn't address social welfare, which is not the same as maximizing unionized wages.

I never saw anything remotely close to the Trump 2 level of corruption in any previous administration, Democrat or Republican.

Unions, like other monopolies, are good for their members and bad for everyone else. Whether unionizing is unselfish or not is debatable, but it makes the world worse.

I am not trying to deny it, but I won't believe it without some evidence beyond bald assertion.

> For 90 percent of refugees from Russia if you participate in totally peaceful protest in front of Russian embassies, you will get yourself arrested

Sorry, in which EU countries does participation in a peaceful protest get you arrested?

Lots of Americans were against the Iraq war, and I'd think expats were particularly likely to be (and the US did not kill a million Iraqis).

That's fine, because you don't need infinite compute to run or train current models. If the "big boys" are buying more compute that's because they hope to provide a better service.

I misspoke when I said "zero". I should have said "down to marginal cost". It's fine for AI to be priced at the cost of its compute! That's the efficient outcome.

What reason do you expect to provide a moat for a monopoly? Training on existing customer interactions? That might get you an oligopoly, but there are already two serious competitors in the US alone and several potential Chinese rivals.

Were these problems just beautiful but useless? If so why did we pay people to work on them? My impression was that many mathematicians receive large research grants.

I sympathise with a lot of the details of mathematicians' complaints, but at the same time I still see very few people stepping beyond "what will this do for the discipline" to think "what will this do for social welfare". It's like hearing scribes complaining about the printing press. Yes, I'm sorry that you won't get paid to do beautiful illuminated manuscripts, but I think the benefit to society outweighs that, and if it didn't, then why were you being paid to do the illumination in the first place?

It's not that there aren't possible arguments here (AI is destroying the basis for further progress). But this perspective is hardly even being addressed by most of the reactions.

Competition will drive down prices to zero, so we'll all be demigods.

Why does it matter to hn? Is it because Dan Brown wrote it?

I think you need to have overall population shares in the denominator.