cayley_graph

Yeah, open weights hasn't fully caught up yet, but it's getting very close. And it's certainly passed the point of being reasonably interchangeable with the frontier for (programming) work. Add in the benefits of not being rug-pulled by the frontier labs silently messing with, the knobs on their models or outright denying you the ability to do certain kinds of work (c.f. the HuggingFace fiasco), and they probably come out ahead in several respects.

The idea that the Chinese labs cannot make progress except by copying superior American products is just prejudice against the former and exceptionalism of the latter at play. Even the OpenAI top brass have admitted otherwise [1]. China is an equal match in every respect, and we'd better admit this to ourselves sooner rather than later so as to see the game clearly.

[1] https://xcancel.com/deanwball/status/2078133895766114412

HF need not be party to it at all, beyond being the victim. I suspect the hack is real; I have observed GLM 5.2 being able to discover similar vulnerabilities in web applications I'm hosting (which I've then fixed!). At the same time, it seems very neatly timed at an inflection point in the conversation around open models, and there's questions around the incompetent isolation under which the hacking benchmark appears to have been run.

Remember that there is generational wealth on the line for most OpenAI employees, and consider what people might do to obtain it.

The timing after the release of GLM 5.2 and Kimi K3 is quite convenient, too, as an angle for regulatory quashing of open-weights models just as they're entering the mainstream conversation around usurping the American frontier labs. I accept my thinking here is conspiratorial, but there's also a hell of a lot of money on the line to encourage the unscrupulous.

They've been saying so from the beginning, and yet did not take the basic precaution of airgapping their off-the-leash model while it's been instructed to succeed at a hacking benchmark by any means necessary. So which is it? I _want_ to believe them, I do, but there's always these gaps between what they say and their actions on display that give me reason to think otherwise.

Yeah, agree on all counts. I'd give them leeway if they were still scrappy startups, but they have entire countries' worth of resources at their disposal and the best of the best on their payroll. No excuses at this point for oopses like this, I would think.

Why is a machine running these sorts of hacking benchmarks not airgapped? That seems a basic precaution, if OpenAI believes what they're selling. I mean, stuff like this is done for CTFs played by humans, too, to rule out collateral damage; it's not some new concept. So this is either thorough incompetence by OpenAI, a marketing piece, or both.

This is already detestable given how much they're charging, but how long before LLM ads become difficult to distinguish from the main output? Remember how Google ads started out? And for all their faults, Google _after_ 20 years of maximizing-shareholder-value disease is a far more trustworthy company than OpenAI _today_. How long do you trust the ethical instincts of the OpenAI leadership to leave money on the table?

Yeah. It's something I can do myself in a couple seconds if I want, also on more varied SVG scenes. If this is going to be a benchmark people turn to I'd like to see more effort put into it than just a one-sentence prompt.

Yup, fair's fair. Anything else stinks of 'rules for thee but not for me' (a maxim the frontier labs seem worryingly happy to apply, on several counts).

Many words expended here to avoid asking an answerable question. Which complaints? Must I recapitulate the last 70 years of history of politics around science in the United States? You have fixated on a singular line from the article rather than reply substantively to anything I've said (which indeed responds to such complaints inasmuch as this is possible without hearing a specific one), because you have little interest in reaching a shared understanding of its subject material (you do not remotely care about any such historical complaints, or I would've heard one by now) relative to your desire to advance the culture war, as laid bare by your other comments. This sort of low-trust behavior, exacerbated by pretending otherwise, is typical of your unfortunate mindset.

The tactics of discourse you employ are tired and show no signs of improving. You've unfortunately lost the courtesy of a further response from me.

Sorry, I'm not in the business of responding thoughtfully to low effort questions slung rapid-fire over the fence, which moreover take nothing I've said into account. Case in point, this comment; it seems innocuous but would take an extreme amount of effort on my part to reply to what is, I suspect, something you've made up your mind about.

I think it's clear from your other comments cited here that you're exactly who I'm talking about: a gullible soldier of the culture war who has perfected the art of wasting endless quantities of time arguing online.

That is an enormous budget cut; it is exactly an evisceration. And the general trend is to cut both science and its application. Trials for life-saving treatments _by biotech companies attempting to commercialize them_ are being halted or cancelled due to the science cuts and general corruption.

Many people have many complaints; many should be ignored. There's lots of money (and indeed, huge market incentive) to commercialize potentially successful science, and as such it has been done consistently. Curiosity-driven science must feed it, and the idea that it does not unambiguously benefit society at large, with extreme and breathtaking return on investment, is a fantasy perpetuated by those susceptible to the idiotic culture war.

You and people you know will lead worse and less fulfilling lives due to this. You, or someone you love, will likely die of causes that would have been preventable without this destruction of domestic science. Academic culture undoubtedly needed reform, but this evisceration bent on shortsighted retribution will help absolutely nobody.

I wouldn't normally reply again, but please try and take my arguments in the good faith and appeal to the heart and humanity that they were given. I think you're quite capable of reading what I've written and responding to it coherently; consequently, I also think you must be aware that you haven't done this in the heat of the disagreement. It's only if we're charitable, not cynical, about this stuff that the world may improve. Thanks. :)

Anyone with kids could tell you that this is an ethically vacant position to hold. Particularly given the social network effects of not being on these platforms and the addiction engineering that goes into keeping you on them, especially at an age when you're prone to feeling insecure about your place in the world. The effects are pronounced in children, but still hold for adults.

Anyway, I think I've said all I possibly can to educate you. I hope you can take something from it.

This is tired and reductivist reasoning deployed only by those rationalizing wrongs, and immediately recognizable by nearly everybody as such. My kids could tell you what's wrong with this thinking. I believe that working a job where I would knowingly contribute to mass mental illness (including but not limited to inducing teenage depression and body image issues leading to suicide), the destruction of liberal democracy and free society, mass surveillance unlike anything constructed before in the history of mankind, and even genocide falls on the wrong side of the line. If you have convinced yourself otherwise, this is a weight you must carry. Acting so callously against your fellow humans always exacts a price, knowingly or not; it requires destroying part of your capacity to care for other people.

Nope, sorry, I have absolutely had the option to do something similar and emphatically declined. I generally don't care to tell anyone this, either, outside the rare instances when it organically comes up as it did here. I want to see myself as a good person with a positive effect (as much as feasible) on the world, and taking such jobs is deeply incompatible with that.

I am completely willing to forgive Meta (and Palantir etc) employees who quit their job and donate their blood money (all wages above some low multiplier of median US SWE salary, adjusted for cost of living) to a reputable charity of their choice. Preferably one focused on repairing the incredible harm inflicted on other humans to which they have been a proactive and willing accomplice. Anything less than that does not constitute genuine remorse; we do not let millionaire criminals keep their illicit earnings because they apologized on the stand.

That nearly none will do this (I suspect most would be irritated at even the suggestion) tells you all you need to know about them.

edit: The disagreement is unsurprising, but I'd like to hear the reasoning against this. If you truly believed you'd wronged humanity at a job you voluntarily took for its high pay over all the others you could have easily gotten, keeping the exorbitant excesses of money should be unpalatable to you. That's how having a conscience works. Anything else is just a vacuous attempt to regain social standing.

You need tools sufficient to do the job in an economical way, optimizing for both cost and quality. That is what 'best' means. We don't give every engineer all the resources under the sun, only what is appropriate.

I suspect many will realize millions more dollars are being spent than needed to achieve the highest marginal productivity gains, and reallocate accordingly. Who wants more of their money going to developer tooling, rather than bonuses?

Since it is the unverified SMP config of the kernel

I don't disagree with your point (formal verification does not rid you of all bugs), but this is not the subject of the linked issue. This was a bug in an unverified path.

Yeah, they showed their true colors there. This, compounded with the fact that they're the only frontier lab with no open models, tells you all you need to know. Tired of the insanely patronizing (+ conveniently and overwhelmingly self-serving) attitude out of them. My goal is to own my computing and be able to choose what to do with it.

Intentionally and silently sabotaging work done with Claude whenever Anthropic decides it is appropriate is unacceptable behavior, and comically tone deaf given the state of open models. Why on earth would I ever pay for a malicious product?

People usually become billionaires via having “super-powers,” i.e., very unusual abilities, at least within some context.

There are certainly sometimes unusual abilities in a positive sense, but the common case likely falls closer to having an unusual degree of sociopathy. It is unclear to me how else one could view the state of perfectly solvable human suffering in the world and continue to prioritize accumulating wealth over all else, moreover and overwhelmingly at the cost of being party to the suffering itself. Indeed, I suspect having such callous disregard for your fellow person is prerequisite to encountering these unfathomable sums.

When people with an intact capacity for empathy come into huge amounts of money I think it's far more common to give a large proportion of it away (say, Jane Street workers have a culture of doing this). And thus you only stay 'comfortably' wealthy, rather than accumulating so much that it distorts society around your singular existence.

Mind linking the experiment? Sounds interesting.

You're of course correct, but it's a complex issue. See my other comment for details.

I did. I believe this sort of stuff to be, at least morally, a violation of the 4A. It's no secret that the anti-Israel protests have gotten an inordinate amount of attention from the law relative to any harm caused, and overstepping bounds like this even to catch actual criminals (as happened here) isn't worth the price paid in liberty.

My comment was targeted at the government/ICE's notorious targeting of anti-Israel protesters broadly. It's absolutely clear that we're giving up rights left and right for this total farce, the same way we did for 9/11. It is imperative to the survival of liberal democracy that this ceases.

No? I would defend the first example too, which is why I specified the idea rather than the letter (the 1A). Is it so rare to see someone who genuinely cares about this stuff, not just for those who agree with me? That I think they're an asshole is irrelevant.

Yeah. It's not cool to be principled anymore, I guess...

I've noticed many right-leaning tech types give quite a lot of lip service to free speech when it's about someone getting banned from a mailing list for being an asshole, and not so much when it's the government quashing protest against genocide. I'll personally always defend the idea of free speech, no matter the side.