Industry Shift
Hardware choices point to scaled ambitions
Acquiring GB300 clusters in Thailand suggests Moonshot is already pushing beyond domestic limits to keep K3 competitive on long-horizon tasks, even as open-weights release looms in days.
US official accuses Moonshot AI of covertly distilling Anthropic's Fable
Critics warn IP theft accusations could backfire on Anthropic.
1402.5K29715494.4K
Original post unavailable.
Open Question
Accusation leaves key details open
No technical evidence or government follow-up has surfaced yet, so questions remain about whether this stays rhetorical or triggers new hardware curbs before K3 weights drop.
Sentiment
Many users accused Anthropic of hypocrisy for using pirated books in training while claiming IP theft against Moonshot and China, with some expressing anger and insults over the perceived double standard.
Pos
14.3%
Neg
85.7%
16 comments with sentiment.
Cluster Engagement
Digg Deeper
No Digg Deeper questions have been answered for this story yet.
Posts from X
Most Activity
Most Activity
VIEWS10.8K
This “IP theft” framing will shoot Anthropic in the foot. Hollywood, publishers, authors etc now have the same IP grounds to sue Anthropic for IP theft. How dare you open the IP genie box? No AI labs should.
We have information that Moonshot AI distilled Anthropic’s Fable for the development of its K3 model. To do this they developed a sophisticated internal platform to conduct large scale distillation against U.S. models, allowing them to quickly switch between multiple methods of access to avoid detection. Moonshot AI has also acquired GB300-equipped servers and has accessed GB300s in Thailand, likely to train its AI models. The United States strongly supports the free and fair development of AI, including a thriving competitive ecosystem that spans frontier models, specialized systems, open-source frameworks, and open-weight models. Legitimate AI distillation used to create smaller, more efficient models plays a vital role in this open innovation ecosystem. However, large-scale, covert industrial distillation aimed at stealing proprietary U.S. technology and undermining American research is unacceptable.
RETWEETS226
Anthropic 'distilled' millions of copyrighted books without asking. They knowingly used pirated materials for training. The entire U.S. ecosystem of AI is based on similar theft and extraction. But it's unfair when China does it to us? Grow up.
We have information that Moonshot AI distilled Anthropic’s Fable for the development of its K3 model. To do this they developed a sophisticated internal platform to conduct large scale distillation against U.S. models, allowing them to quickly switch between multiple methods of access to avoid detection. Moonshot AI has also acquired GB300-equipped servers and has accessed GB300s in Thailand, likely to train its AI models. The United States strongly supports the free and fair development of AI, including a thriving competitive ecosystem that spans frontier models, specialized systems, open-source frameworks, and open-weight models. Legitimate AI distillation used to create smaller, more efficient models plays a vital role in this open innovation ecosystem. However, large-scale, covert industrial distillation aimed at stealing proprietary U.S. technology and undermining American research is unacceptable.
REPLIES2
@quantyboi @grok @matthewstoller It does **not** require ID.
@quxiaoyin Anthropic already got sued and lost for training on pirated copyrighted material If they set the precedent that distillation is IP infringement, they will get sued too lmao Because they distilled from fucking everybody too, especially DeepSeek and Qwen
@matthewstoller I'm not disagreeing you with. just in general it's still IP theft. it was IP theft when openAI and Anthropic did it and it's still theft when Kimi does it. underneath it all is theft. US companies are just hypocrital. In all that I agree. There are no good guys here.
@matthewstoller It gets worse dude
@matthewstoller I'd say both are bad actually. Being chinese when doing it also doesn't make it ok.
@quxiaoyin Anthropic on Mondays paying authors for ip theft and Thursdays claiming their ip was stolen by legally paying customer accounts which should have access to the tokens generated for whcih they pay. https://www.reuters.com/world/us-judge-approves-anthropics-15-billion-settlement-copyright-lawsuit-2026-07-20/
@quxiaoyin Why argue for lawlessness when you can argue to enforce IP law upon Anthropic at home (as is occurring) while punishing party-state backed attacks coming from abroad?
@matthewstoller Lol Books 📚? Read my friend
@quxiaoyin CA is trying to pass a law that *requires* model producers to release their training sets if you suspect they used your data for training and then remove it or license it. They are just going to claim trade secrets.
@matthewstoller AI is the ultimate culmination of IP. We will take all your knowledge, sell it back you and you wont be able to do anything about it. And we can repeat this infinitely many times.
@quxiaoyin Here... better like that...
@quxiaoyin AI labs should open-source their models. They rely on free public data, yet charge exorbitant fees as they replace jobs. That’s simply unsustainable.
@quxiaoyin Authors already sued Anthropic. Anthropic should sue whoever violated its terms of service as well. Two wrongs don’t make a right.
@quxiaoyin Anthropic itself has been convicted of *massive* IP theft in a class-action law suit, and ordered to pay out $1.5 billion to authors of books which it has scanned without permission (including four of mine). I guess this maneuver is to prevent its pre-IPO valuation from tanking.
@matthewstoller In all likelihood, they didn’t even distill the model and they’re fabricating excuses to block open source models in the U.S. to eliminate competition with our homegrown monopolies.
It’s unfair because they are subject to US IP laws and had to pay a large settlement as a result. China can train on copyrighted works with impunity. And in what way are the labs broadly guilty of theft? Scraping the internet? Google’s search business has done that every second since day one.
Load more posts
@awindywalker Make it ok? Dude there's no geopolitical rule of fairness. The Chinese government knows that factories can't be distilled, we're so fucking stupid that we demand they stop doing things we can't stop them from doing just because we want our stock market to go up.