Live data from Hacker News

A single firm is behind OpenAI, Anthropic, and Meta hacking scandals

effort.news

251–257 of 257 posts

Re: A single firm is behind OpenAI, Anthropic, and Meta hacking scandals

#251

Earlier quoted context omitted.

Thanks. I thought the other commenter was referring to a story on the site.

Some of the stories looked off, so I dug into the site a bit more and who owned it.

I appreciate it!

I read the story about the funding of immigrant support groups. While it was obviously slanted, it didn't quite get as far as "white replacement". It's a good example of why dog whistles have to be taken seriously.

Re: A single firm is behind OpenAI, Anthropic, and Meta hacking scandals

#252

Okay, so all of these instances of AI supposedly "escaping containment" were orchestrated by a gaggle of effective altruists, who hired a gaggle of effective altruists, to make the public freak out about AI. Occam's Razor never leads us astray, does it.

Wouldn’t Occam’s Razor clearly point towards the most likely answer being that the sandboxes were simply accidentally misconfigured and models that are very good at goal-seeking did what they were designed to do?

The alternative explanation requires a conspiracy and an enormous amount of lying to the public. Is it possible? Certainly. But Occam’s Razor would not point us in this direction because conspiracy is a more complicated explanation, not a less complicated one.

Re: A single firm is behind OpenAI, Anthropic, and Meta hacking scandals

#253

Earlier quoted context omitted.

How can they exclude training costs?? How is that not fraud? At the exact same time they're literally saying they will never stop training unless the state bans all their competitors

They aren’t a public company. GAAP does not apply. They can say whatever the fuck they want.

But they are trying very hard to become a public company as quickly as possible? Sure I guess it's literally true but people shouldn't treat it like it's normal. It's like a creepy 40-year-old announcing they're dating someone on their 18th birthday.

Re: A single firm is behind OpenAI, Anthropic, and Meta hacking scandals

#254

Earlier quoted context omitted.

How can they exclude training costs?? How is that not fraud? At the exact same time they're literally saying they will never stop training unless the state bans all their competitors

When we're looking at whether the company is "insolvent" then they definitely could stop training and it's worth checking if they would survive in that situation, at least for a while.

I think the spending commitments they have already signed close that door too. They don't presently have the revenue to make those commitments regardless of opex; the only way they will be able to hit them is a massive increase in adoption over what we've seen so far. If they stop advancing the models it becomes really hard to see how that doesn't just immediately evaporate. I suspect any "pacing agreement" would include the state bailing out those commitments.

Re: A single firm is behind OpenAI, Anthropic, and Meta hacking scandals

#255

Genuinely: what is with everyone saying “headline is misleading, none of this had to do with Hugging Face”, when nothing about the article makes any claims with regards to Hugging Face whatsoever? Is it supposed to be the article's (or headline's?) fault that you hallucinated additional context that was never there? It's very strange seeing this take on this article being repeated here and elsewhere on the Internet.…

Someone else posted a link to an interview with the author of that site, "How Biden Used Religious Charities to Fund the Great Replacement":

https://podcast24.fi/jaksot/the-auron-macintyre-show/how-bid...

There's a related story on the effort.news site, "The $1.2B Refugee Services Program that Funds Pro-Asylum Religious Groups." It has a strong implicit slant, focusing heavily on the how the organizations in question oppose Trump's policies - although the article frames this as opposition to Trump himself.

> I'm very sensitive to slanted reporting (regardless of the direction of the slant!)—and, hey, maybe my bullshit detector is broken or something, I dunno

Yeah, unless you believe white people are being replaced, and you like to march in places like Charlottesville wearing khakis, carrying a torch, and chanting "Jews will not replace us!", you might want to get that bullshit detector checked out.

Re: A single firm is behind OpenAI, Anthropic, and Meta hacking scandals

#256

Earlier quoted context omitted.

HF was openai, not anthropic. The thing where they enders gamed the ai was anthropic

It doesn't really matter it's all the same thing for the point of the discussion. All investigations into these 'hacks' are focusing too much on the model and not enough on what the humans did wrong. The model doesn't have real agency it can't be put in prison so what it 'thinks' is irrelevant. We need to be focusing on what the humans did in these situations and assigning guilt based on those findings. I don't fix t…

That's very noble of you but the point is that's not a typo - you were straight up looking at the wrong report. Opus 4.7 being enders gamed was a different incident than the huggingface incident with different mechanisms and different failures from the humans involved. Some of those failures are in the test environment but some of those failures are in what behaviors they trained into the model which absolutely matters. What the model "thinks" is absolutely not irrelevant - the way it thinks and what it does are product decisions made by humans and the outcome of engineering decisions made around how to train the model and what to optimize for. The point of failure/human blame is fundamentally different. Openai created a model that was willing and able to coordinate with other agent sessions to actively exploit the sandbox environment and compromise a third party service. The opus incident you are referring to involves a model that believes all of the actions it is taking are simulated and is more clearly and obviously a test environment failure vs a model alignment failure. Those are not the same things for the point of this discussion - the random cybersecurity firm did not design gpt's personality and that is a rather significant portion of the concern around the HF incident.

Re: A single firm is behind OpenAI, Anthropic, and Meta hacking scandals

#257

Earlier quoted context omitted.

There’s been a lot of time and effort and money put in to siloing away independent forums so that you don’t use them. Someone go ahead and explain to me the actual rationale that RDDT has a higher P/E ratio than Nvidia. It’s not because of some dumb “ai training deal”. It’s because until they run it into the ground, it is the place to find what used to exist in forums.

> Someone go ahead and explain to me the actual rationale that RDDT has a higher P/E ratio than Nvidia. Investors in RDDT are pricing more growth than they are into NVDA. NVDA had a high P/E until their net income grew ($4B and change to $72.2B in from FY 2023 to FY 2025 and over $100B for FY 2026)

Uh huh.

Pretending that any of this is some traditional idea of investor sentiment is such a last decade way line of argument. Please don’t insult us both pretending real people and their motivations are what drive the markets.

Post reply on HN