meowface

It's actually the opposite now. I have never used Svelte until LLMs came around. LLMs are better at writing Svelte than React, and then you also get a faster frontend by default.

It's like Rust. Way more Python than Rust in the training data, but with LLMs you're much better off defaulting to a Rust backend and Svelte frontend than any other combination. I'd say "unless you have a specific reason", but besides legacy requirements, there is actually no reason.

...Could you not ask this of anyone who works on pure math problems? What on Earth are you talking about??

"Academics are paid paltry salaries, university departments are cash-strapped, why are they studying these abstruse math problems instead of building rockets?"

Successfully advancing fields of math and science is a good thing. Why would they not do it. Why would they not want to prove their increasingly intelligent computers are able to solve unsolved, complex math problems humans have failed to solve for centuries and then release the solutions to the public. What do you want them to do, not try to solve them?

Also, people would get mad at them (rightly) if they sat on a proof of a major unsolved problem for months while bringing in human mathematicians to convert it into a more human-friendly form

(They definitely should spend SOME time working with human mathematicians before putting out a press release, but that's exactly what they've done every time)

Am I the odd one, or is this, like... absolutely insane logic?

"The use of proprietary internal models by AI labs to do mathematical research risks creating a two-tier system where labs outrun the rest of the field, effectively alienating the mathematical community from its own discipline."

Are mathematicians' private brains unfair if they're way better at solving open math problems than anyone else? Are they required to share everything they're working on, as they're working on it?

Should Andrew Wiles have been condemned? Wouldn't these arguments all have applied to him? Besides him working at more human speeds.

I'm overall pro-Anthropic and pro-banning open-weights AI, but I agree with the parent commenter; distilling Claude models is not that different from pretraining on web data. It's all basically the same sort of thing.

I think a good litmus test here would be if Anthropic were to not care about distilling their models when the distillers keep the resulting models closed-source and sell tokens via an API. If they cared only about security concerns and not about people profiting off of their work, then they should be publicly fine with this and only protest against it going into open-weights models.

The discussion was originally about Altman. If we're going to discuss Musk: sure, he is an absolutely horrible person and political figure in myriad ways. Arguably one of the worst people alive, in some ways. (Despite his laudable business achievements and some public goods like Starlink.)

One can't lump together Altman and Musk. (Plus they despise one another.) Or lump together all tech CEOs or all billionaires.

The rest of your post is incorrect, zero-sum thinking. There is not a fixed pie which people take proportional slices out of. That's not how wealth or the economy work. Nor are relative wealth differences the most important thing people care about. People somewhat care about making more or less money than their neighbor, but they care even more about their absolute standard of living and material conditions.

Some thoughts on this topic that I agree with, by the creator of this website:

https://www.paulgraham.com/wealth.html

https://www.paulgraham.com/sim.html

https://www.paulgraham.com/ineq.html

https://x.com/paulg/status/2089044855158956533

https://x.com/paulg/status/2052510098035998846

As for concentration of power in societies: so long as there is still democracy and some laws preventing ultra-wealthy people from influencing politics through completely corrupt means (directly bribing politicians), it's not necessarily a big deal, if everyone already has great lives. The issue with people like Peter Thiel is they publicly say they dislike democracy, whereas someone like Sam Altman publicly (and I suspect privately) supports democracy.

Even if they had no power at all, they'd still be forced to suck up to the authoritarian Trump regime. They have to, because he's petty and vindictive and they rationally don't want to be wrecked by a narcissistic authoritarian.

>However, even in a literalist reading, the trends are obvious and alarming: Twelve people own more of the world's wealth than the bottom 50% of the population (!).

This alone implies nothing. What matters is the floor. A world where there was almost no wealth disparity but everyone was poor would be awful, and a world where 5 people had 10 billion times more wealth than everyone else combined would be good if all those other people lived in amazing prosperity. And if you check the correlations on standard of living and wealth inequality, you'll see it not only implies it but to some degree implies the opposite. (Obviously, it is more complicated than that alone, and increasing the floor would at this time require raising taxes in a way that evens things out more.)

I might not have been clear enough in my post, but I tried to include those exact caveats there and in other posts in the thread.

In an ideal world, that would be true.

I definitely do not know what I'm talking about, but that individual doesn't know what they're talking about even more than I don't know what I'm talking about.

Correct: none of that is correct.

The poster is not claiming Bel is a secret AGI. Just that it exists and only exists internally at the moment.

It's rumored to be over 10T parameters. When released it'll probably be very good at certain tasks, albeit slow and expensive and not necessarily "wiser". You don't have to make this a binary.

Also, Mythos was in fact a significant step-up in several ways. It fits the trend line, but only because the trend line for LLMs is quite steep. Plus Astra is still in many ways less intelligent than Fable/Mythos despite being released much later.

With all due respect, you do not have a single clue what you're talking about.

Conspiracist nonsense lapped up by base populists.

(I say this as someone very pro-immigration. Though, I guess I would be, wouldn't I, since anti-immigration people will readily adopt both right and left conspiracy theories.)

You are empirically correct, but one could likely design a prediction market that resolves this while still making insider trading an informational public good. Unfairness to counterparties is a different axis from non-monetary public forecasting signals.

Gambling is for games of chance, while one can "beat the house" in prediction markets simply by being more intelligent, informed, and better at forecasting.

At least for the non-random-like bets.

I think the thing is insider trading genuinely is exactly part of how prediction markets contribute something that was not offered before: a more accurate way for the public to gauge the probability of future outcomes. I, as a rationalist, think there is a lot of good that results from that.

It just happens to almost always be the case that the bad greatly outweighs the good, and it usually just leads to things like brazen war profiteering (+ plan-leaking that jeopardizes operations, in the event the military operations were perhaps for a morally net good cause).

Prediction markets seemed like a possible exception to the rule that anything to do with betting is always a scourge on the commons, but those companies quickly realized sports gambling is the best way to rake in cash, so they all converge on it.

I am one of the very few people in the United States who actually does still earnestly believe prediction markets do in principle serve an immense social good that can't be provided in any other way: https://www.astralcodexten.com/p/prediction-market-faq

Just, in practice, Polymarket and Kalshi are (unsurprisingl)y run by people who solely care about how much revenue they could ever possibly inhale.

And everything else about them is so unethical and so destructive to the commons I would support them being metaphorically razed to the ground.

Like, just the Kalshi and Polymarket Twitter accounts alone contribute more to disinformation (disinformation, not misinformation) than Alex Jones could ever dream of. And then all the accounts affiliated with them make it 1000x worse. And then they have so many other unbelievable scandals I'd have to write a wall of text to list them all. Tons of egregiously machiavellian, intelligence agency-tier tactics and operations.

The only prediction markets that seem to be mostly upside with minimal downside are the ones that do not accept real money.

I do not support the US's support for Israel. I also do not support Russia's support for Iran or China's support for North Korea.

Hopefully we get a 2029 president who completely alters the alliance with Israel and makes further close association contingent on many serious changes in how they handle Gaza, the West Bank, and prosecution of terrorists and war criminals.

I don't support the Iran War but I do support global US hegemony because I believe the other two possible candidates, China and Russia, would be worse. And I think retaliation against Iran for their massacre of protesters was morally warranted, even if that was only a pretext for the war by Trump rather than an actual moral consideration, and even if moral merit alone does not justify major intervention or war.

"Slightly geopolitically neoconservative" is the short way to describe that.

There's no evidence the US's motive or achieved outcome of the war was oil.

Seems "he tried to kill my father, now there's an excuse to take him out and turn his country into a democracy in the region, and a US-aligned entity" explains most of it.

Much like Trump, it's pettiness and prestige and a notion of more points on the board. (Though Trump did in fact attack Venezuela for oil, or possibly more so just to publicly boast about taking their oil and seem strong to his fans.)

It does, but it feels like every paragraph is inserted to fit only one thesis.

Why is this article written like an op-ed, despite being Reuters?

I am not even necessarily disagreeing with it (I lean very anti-Iran and slightly geopolitically neoconservative) but the tone is not at all neutral and makes no mention of the possible downsides of troops staying in Iraq after decades.

That may contribute to hesitance to report such incidents upon their discovery, yes. They've been caught trying to cover them up repeatedly.

(As opposed to the bizarre bullshit where people are claiming them having or talking about such incidents helps their IPO.)

My post doesn't have any LLM tells, nor have I ever used one to write a social media comment.

That would actually be interesting if LLMs were to ever refer to LLM trust in the way I did. Kind of a scary thought.

Same. I spend tokens on so much more than just the initial implementation of a feature I have an idea for.

I am actually going to go out on a limb and guess the opposite of this is true, and that on average vibecoders burn tokens less readily than veteran software engineers. I could list several reasons why I think this would be likely. No idea which of us is empirically right, though.

(With exceptions for what I can only call the "manic vibecoders" with like 10 simultaneous weird slopprojects they're spewing out at once. Generally with each project itself being something related to vibecoding. Steve Yegge being an example of a "manic vibecoder-actual programmer" hybrid.)

Sure, but their revenue growth numbers are also crazy.

I am sure this poignantly describes some CEOs, but my guess is this is probably empirically false on average.

(I'd ask some LLM to research it but the people who would be doubtful it's false significantly overlap with the people who distrust LLMs, so I'll just leave this as a random guess and nothing more.)

That would require mens rea. This may be criminal negligence, but it wouldn't be more than that. (I am basically 100% sure none of the employees or execs are intending or desiring any of these outcomes, regardless of the very large number of people who believe in conspiracy theories about regulatory capture and other sinister motives.)

I'm totally on board with treating it as gross negligence requiring hundreds of millions or billions of dollars paid in fines and compensation to victims, but don't act like this is more than what it is.