No. If a CEO makes self-serving comments that are in their interests, but not the company’s interests, that could be fiduciary misconduct. With that being said, CEOs have a fair amount of latitude in what they can say, since it’s difficult to tell what the “optimal” thing to say at any point would be (see: Tim Cook telling an activist investor to “get out of the stock” if he didn’t like Apple’s environmental initiatives).
The idea that CEOs’ hands are tied, and that they are kept on some sort of short legal leash, is largely (though not entirely) a fiction.
I think that Jensen's position as the guy selling the proverbial shovels incentivizes him to take a lot of irresponsible positions (namely around safety) but imo he's fundamentally correct here.
He comes off bad in this interview. Out of touch is correct - "I had to pump gas once a few years ago ..." and panicked because he didn't know his address or zip code.
AFAIK he was outright lying when he said he forgot his zip code. It hasn’t been possible to buy gas in Bay Area for decades without inputting your zip code for credit card purchases. That smelled of some bad speechwriting.
>he doesn't think kids should learn their multiplication tables
He said „The majority of the kids shouldn't learn”, not everybody. I listened to that pod and was under the impression that he wants to be portrayed as a "grounded" person and i didn't find his takes to be out of touch.
As sometime who's memorized but also forgotten their timetables while still managing to go through Algebra I. I don't understand the importance for memorizing times tables (or any other tables for that matter).
There are basic skills that everyone should learn, being able to do simple math in your head is a building block for more complex reasoning.
I once had a roommate who I encouraged to divide our grocery bill by 2 in his head for splitting the bill, by the end of the year he was quick and accurate.
That doesn't follow. I'm not especially gifted at mental math yet I've been able to learn various programming languages and libraries and dev tools. I've also been able understand and implement some algorithms.
Quick math. Basic financial literacy. Times tables are for arithmetic, the working man's math. Algebra is an entirely different math which solves different problems and deals more in the abstract.
You can definitely get past algebra without knowing your times tables, but algebra isn't going to help you in the checkout line to make sure you received the correct amount of change and bought the correct number of goods.
Most people on a budget (i.e. the vast majority of people) are constantly using exact or approximate multiplication while shopping. And they are rarely whipping out their calculator for this purpose.
Personally I think multiplication tables aren’t that useful because they’re just rote memorization. I think it’s better to teach kids to memorize a few such as x * 2, x * 5, x * 10 and teach them how to extrapolate from there. Extrapolation is a more useful skill than memorization.
Can someone answer this monumentally dumb question: if anthropic cares so much, why don't they meganerf their thinking summaries (edit: in Claude Code) the way OpenAI does? OpenAI's thinking traces so much that the Codex app doesnt even show them.
In Claude Code (at least inside VSCode) thinking traces are hidden by default. You can reenable them via JSON. They’re not the real thinking traces though, they are something like thinking summaries.
I think it would be totally incoherent to say that copyrighted information is essentially fair game to include in your model but the outputs of your model are privileged against being included in other models.
I'm not sure "competition" is necessarily the right word but I don't think that model creators and their political backers have a leg to stand on when complaining about distillation.
The accusation of theft by anyone is really only as strong as their financial capability to enforce it. The leg that the model companies stand on is their bank account, which is increasingly becoming the only force that moves the needle in this economic and political environment.
> I think it would be totally incoherent to say that copyrighted information is essentially fair game to include in your model but the outputs of your model are privileged against being included in other models.
IANAL, but going from copyrighted texts to an AI model likely constitutes a new original creation, not a slavish reproduction. Whereas going from one AI model to another is more likely to be considered a slavish reproduction reproduction, although arguably AI models are not copyrightable to the extent that the weights are objective facts, like entries in a phone book.
Realistically the companies will organise their data through a network which will involve distributing and storing the copyrighted material without the consent of the copyright holder. This is in and of itself generally ruled to be a breach of copyright irrespective of whether it is fed into a model or not.
If your recommendation algorithm always recommends B when it sees A, and I derive this information from exfiltrating your masked data , is it not theft simply because I didn’t unmask the true identity of A and B?
Stole all other important information embedded in the relationship (the essence of the relationship itself).
This is not seen as theft to only two or three types of people:
1) Technically ignorant
2) Or Technically ignorant and morally bankrupt
3) Truly evil, a combination of technically capable, and morally bankrupt.
(An “Abomination” is also possible, unaware and unable to perceive morality , not just mere ignorance. For example, it is tremendously generous to call an Abomination ignorant or amoral, that’s a compliment to such a thing. Would Jensen Huang know about the words coming out of his mouth, is the true question really.)
Exactly. AI companies distilled the whole internet into AI weights, now other AI companies are distilling their work to generate new AI weights. Nothing wrong with that, and the world gets open weights AI models.
I agree that AI companies shouldn’t be allowed to bar distillation under grounds of fairness and societal benefit.
But I don’t see how copyright has anything to do with anything. Copyright is a legal concept, not an ethical concept. Training an LLM on copyrighted material is legal. Ethically, whether a work that the LLM trained on is copyrighted or not has no relevance
One is allegedly a copyright violation (fair use IMHO but still a grey area legally AIUI), one is a very explicitly a violation of the terms of service you agree to when signing up. And yes I do think that if somebody puts their website behind some barrier where a user has to actively affirm they won't download the contents and use it to train an LLM that should be protected by the same law for the same reason.
Are you arguing that training on copyrighted material is fine but distilling against tos is somehow bad? Lmao what a take. Nah, if copyright cant even be enforced then theres no way on earth violating a tos should ever have repercussions other than they ban you. stop carrying water theyre not gonna pay you
On paper, but the world isn’t constrained by such trivial matters when enough power is behind you. Although, smuggling only goes so far, obviously, it’s not like they’re not available.
Jensen is borderline comical on his instances of high GPU usage. OpenClaw? "let's do it, it's amazing!", selling GPUs to china "it's of uttermost importance, USA can't afford not selling GPUs to china because other will do", chinese firms burning GPU cycles like hell on american datacenters from openai and anthropic? "this is just competition, let's do this, this is amazing".
I wonder how the "Chinese AI is all just distillation" folk are going to cope now that the Chinese are all aboard the post-train via agents in custom training environments wagon?
Here's Xiaomi's discussion of this, plus their open-sourcing of 7000+ RL training environments.
Now when the big LLM companies create their products by accumulating a notable portion of copyrighted creative works (including computer programs) from Internet, it does not count as copyright infringement or competition (or "theft"). It is considered as "fair use". So, why training LLMs on other LLMs is not fair use too?
> If you don’t like that, if you don’t like people to use your products, all you [have to do is] know your customers, and disable the service
Considering the current situation, I understand Mr. Huang point, but that's not how the copyright framework assumed to be working from the beginning. It should protect both small actors (authors) and the big companies from unrestricted use of creative works. Now this mechanism seems to be practically dysfunctional.
And a big portion of this lies on shoulders of proponents of permissive OSS, who defend an idea of (almost) unrestricted use of their source code texts for many years, and long before mass LLM scrapping became a thing.
55 comments
[ 3.3 ms ] story [ 43.0 ms ] threadHis company has been public longer than you’ve been solvent.
The shareholder part is also not correct in they way it's being implied.
The idea that CEOs’ hands are tied, and that they are kept on some sort of short legal leash, is largely (though not entirely) a fiction.
while I generally agree on his open weight stances, I lost all respect in that moment, everyone of these people are so out of touch
He said „The majority of the kids shouldn't learn”, not everybody. I listened to that pod and was under the impression that he wants to be portrayed as a "grounded" person and i didn't find his takes to be out of touch.
I once had a roommate who I encouraged to divide our grocery bill by 2 in his head for splitting the bill, by the end of the year he was quick and accurate.
You can definitely get past algebra without knowing your times tables, but algebra isn't going to help you in the checkout line to make sure you received the correct amount of change and bought the correct number of goods.
I'm not sure "competition" is necessarily the right word but I don't think that model creators and their political backers have a leg to stand on when complaining about distillation.
It's entirely possible for AI output to be copyrighted if it meets the requirements; prompting an AI takes some skill.
IANAL, but going from copyrighted texts to an AI model likely constitutes a new original creation, not a slavish reproduction. Whereas going from one AI model to another is more likely to be considered a slavish reproduction reproduction, although arguably AI models are not copyrightable to the extent that the weights are objective facts, like entries in a phone book.
Stole all other important information embedded in the relationship (the essence of the relationship itself).
This is not seen as theft to only two or three types of people:
1) Technically ignorant
2) Or Technically ignorant and morally bankrupt
3) Truly evil, a combination of technically capable, and morally bankrupt.
(An “Abomination” is also possible, unaware and unable to perceive morality , not just mere ignorance. For example, it is tremendously generous to call an Abomination ignorant or amoral, that’s a compliment to such a thing. Would Jensen Huang know about the words coming out of his mouth, is the true question really.)
But I don’t see how copyright has anything to do with anything. Copyright is a legal concept, not an ethical concept. Training an LLM on copyrighted material is legal. Ethically, whether a work that the LLM trained on is copyrighted or not has no relevance
https://www.reuters.com/world/china/chinas-deepseek-trained-...
Here's Xiaomi's discussion of this, plus their open-sourcing of 7000+ RL training environments.
https://www.alphaxiv.org/abs/2609.mimo-scaling-reinforcement...
https://mimo.mi.com/docs/en-US/news/latest/v2-6
Who needs a few of someone else's "vacation postcards" of their post-training experience when your agents can go on vacation themselves!
> If you don’t like that, if you don’t like people to use your products, all you [have to do is] know your customers, and disable the service
Considering the current situation, I understand Mr. Huang point, but that's not how the copyright framework assumed to be working from the beginning. It should protect both small actors (authors) and the big companies from unrestricted use of creative works. Now this mechanism seems to be practically dysfunctional.
And a big portion of this lies on shoulders of proponents of permissive OSS, who defend an idea of (almost) unrestricted use of their source code texts for many years, and long before mass LLM scrapping became a thing.