As I was falling asleep last night, this thought occurred to me: maybe NVIDIA bought Hugging Face so they prevent HF from litigating OAI for the hacking incident, because if OAI was very publicly sued for this kind of behavior from an agent, it could pour a massive amount of ice water on agent adoption as businesses suddenly see current agent systems as a existential business risk, and it would pop the infrastructure bubble (of which NVIDIAs entire valuation rests). The fact that somehow OAI committed a felony and have managed to pivot the public conversation around it to be aimed at regulatory capture instead of them being held legally accountable is pretty wild.
We must build the Torment Nexus, lest others build the Torment Merry-go-round, and the Torment Rollercoaster, and we would miss out on the Torment Tormentor Tormentober!
If HF holds OAI accountable to pay for "rogue" agent behavior, the rest of the business world will be more afraid of adopting AI, which will reduce demand for Nvidia GPUs, possibly at a long-term cost to Nvidia much greater than $100M.
Whether or not a human is putting in any effort to ensure the safeguards are enabled and working should be the million dollar question. And even more expensive if the safeguards were deliberately disabled for a "our AI is so smart we can't contain it" marketing piece.
Did you miss the "existential" part? It's not about the cost of one lawsuit, it's about the possibility of being held responsible for what one's agents do - something that could pop the ai bubble and crash nvda.
> When Hugging Face tried to investigate, analysing the intrusion meant submitting the attacker’s own code to commercial AI tools. Those tools refused, unable to tell an attacker from a victim.
This is a worrying part as well. Our tool broke into yours but you can not use our tool to fix it - sorry.
When this event happened I was confused why they didn’t file a police report and make OpenAI demonstrate in criminal court that they weren’t engaging in illegal corporate espionage and other rather felonious hacking activities intentionally.
Why did anyone take their word that it was all an accident? Why am I believing the burglar standing in my house?
i'm not in the cybersecurity business so someone who is explain SOP to me in times like this. Let's say you and your team are 90-95% confident an attack has taken place and your systems have been breached. How far down in the list of things to do next is "Engage law enforcement"?
Until now, I would think it's pretty high up the list. Hugging Face never seemed to reach that step even though they had confirmed an attack. To me, if my systems were attacked and I was damaged I would notify the FBI of the attack in the same way I'd notify my insurance.
Maybe because all the AI actors, probably including Hugging Face, want to push the narrative that AI companies aren’t responsible for what their agents are doing.
I think our societies will have to settle on this.
this is an article from July, I haven't heard any updates of what came out of this apart from the incident report that OpenAI released.
I presume the poster submitted this article in light of the recent AI safety discussions taking place, but as for this particular incident I haven't found an actual confirmation that $100M worth of inference was given out to the HuggingFace community.
This statement was also months prior to the acquisition by Nvidia, which again I was interested in seeing if they would still pursue this
This article is 2 months old when the event happened. And they didn't "bill them", they politely asked them on Twitter, and it looks like that went nowhere.
$100M for RubyGems ought to keep the servers running for several years. A policy to pursue and spend such windfalls on security development would make the service antifragile.
I wonder if it's in openai's interest to play nice here. They can't have a precedent of "the future is we do whatever we want with no consequences for screwing up" if they want public interest on their side, but they also can't have "you (our customers) will be held accountable for what you're paying us to do if we screw up." And that's before the IPO interests even come in.
This whole thing is still all too weird, the disclosing blog post and whatnot clearly shows how carefully engineered and designed it all was, it's like many thinking heads and hands touched it every step of the way. What amazing times.
and the actual tweet from the HuggingFace CEO back in July
> In the spirit of transparency, here’s what I asked @OpenAI:
> • Radical transparency: let’s release the traces from the “rogue” agents so the entire research community can study what happened.
> • More capabilities for defenders: let’s commit $100M in compute from OAI to help the Hugging Face community build powerful cyber defenses with the best open and closed models.
> The first autonomous agent cyberattack is an unprecedented event. It deserves an unprecedented response!
Clem knows they have the community support and potential legal leverage here. It's not unreasonable to ask for a lot. It's in everyone's interests to play nice. Huggingface could do a lot with more compute, even with nvidia backstop. Anything with large scale gpus quickly gets into ~year lead times, "contact sales", and complex talks. Large scale being only > 64!
The liability question is one of the biggest things people are looking for in the S1. If you sell a dangerous product that harms people or businesses you’re generally liable. It’s as of yet unclear how the big labs intend to account for potential massive liabilities.
Folks tend not to go after companies that are just burning cash, but the second they become legit GAAP profitable (if that ever happens) lawyers will be lining up down the block to sue them for any and every mistake these models make.
“Were you harmed by OpenAI’s models? You may be entitled to
Compensation. Call 1-800-SUE-AI-BROS” billboards will be everywhere.
Yeah no, man, that's slumberland territory.
Why would nVidia, then, do what TFA is describing?
Whether or not a human is putting in any effort to ensure the safeguards are enabled and working should be the million dollar question. And even more expensive if the safeguards were deliberately disabled for a "our AI is so smart we can't contain it" marketing piece.
It's not like they didn't get anything of value for that 13 billion
This is a worrying part as well. Our tool broke into yours but you can not use our tool to fix it - sorry.
When this event happened I was confused why they didn’t file a police report and make OpenAI demonstrate in criminal court that they weren’t engaging in illegal corporate espionage and other rather felonious hacking activities intentionally.
Why did anyone take their word that it was all an accident? Why am I believing the burglar standing in my house?
Until now, I would think it's pretty high up the list. Hugging Face never seemed to reach that step even though they had confirmed an attack. To me, if my systems were attacked and I was damaged I would notify the FBI of the attack in the same way I'd notify my insurance.
I think our societies will have to settle on this.
They, like all capitalists, want to push the line "We have money, will make money, and thus, like all money makers, shouldn't be punished"
I presume the poster submitted this article in light of the recent AI safety discussions taking place, but as for this particular incident I haven't found an actual confirmation that $100M worth of inference was given out to the HuggingFace community.
This statement was also months prior to the acquisition by Nvidia, which again I was interested in seeing if they would still pursue this
You made my day.
ChatGPT
> In the spirit of transparency, here’s what I asked @OpenAI:
> • Radical transparency: let’s release the traces from the “rogue” agents so the entire research community can study what happened.
> • More capabilities for defenders: let’s commit $100M in compute from OAI to help the Hugging Face community build powerful cyber defenses with the best open and closed models.
> The first autonomous agent cyberattack is an unprecedented event. It deserves an unprecedented response!
https://x.com/ClementDelangue/status/2081056675558195657
b) And this is before Hugging Face agreed to acquired by NVIDIA, supplier to OpenAI?
Folks tend not to go after companies that are just burning cash, but the second they become legit GAAP profitable (if that ever happens) lawyers will be lining up down the block to sue them for any and every mistake these models make.
“Were you harmed by OpenAI’s models? You may be entitled to Compensation. Call 1-800-SUE-AI-BROS” billboards will be everywhere.
Can we stop these charades? Maybe circular hacking deals are next?