49 comments

[ 3.5 ms ] story [ 71.6 ms ] thread
> Anthropic said it identified a campaign by DeepSeek and two other Chinese AI labs to illicitly extract capabilities from its Claude AI platform to improve their own models

Oh, won’t someone think of the poor mass copyright infringers.

Hmm, my VPN provider explicitly has Chinese exit points. And whats funny is I can load AliPay from any CVS. (Like, seriously)

You can try to pry Qwen and Deepseek from my Graphene/Linux hands.

The US government exists to defend capital interests. It's why we can't buy BYD cars. It's why we can't import any cars unless they're 25 years old. It's why a Tiktok sale was forced. It's why the US is seeking to block states from banning prediction markets. It's why the federal government is seeking to block states from blocking data center projects.

As soon as DeepSeek came out I realized what was going on: China was going to make sure that no US company was going to "own" AI. It is an issue of national security. It's why the US essentially blocks US tech companies to maintain sovereignty.

I'm reminded of the browser wars of the 1990s that led to the antitrust suit against Microsoft. Microsoft used the "commoditize your complement" strategy [1] against Netscape. The US has blocked the export of not only EUV lithography but high-end chips to China. China doesn't want to be dependent on US platforms or policy.

So China is going to make sure there are open source models available and the US government is going to try and stop them to protect US tech companies.

[1]: https://gwern.net/complement

Noteworthy that Z.ai, maker of the just released near-frontier GLM 5.2, has already been on the Entity List since Jan 2025[1]. Being on the Entity List does not mean all trade is forbidden. Broadly speaking it means American companies and individuals are not allowed sell them goods and services, but they are still allowed to buy from them and pay them.

AFAIK the Chinese AI companies barely depend on US goods and services, except for nVidia GPUs which were export restricted anyway, so it doesn't seem to be very consequential (see Z.ai). For the RAM maker CXMT it could be a lot more problematic though.

[1] https://en.wikipedia.org/wiki/Z.ai

The whole thing seems like nonsensical.

Their website literally has chinese characters on it even in english mode and everyone under the sun including crappy money talk show hosts know them as the chinese player that undercut western players. It's not exactly a secret.

You'd think anyone with two brain cells and confidential data could apply some judgement of their own...

To give credit where credit is due, it is good that the Trump administration has not avidly played these stupid export control games. They tend to do little except hurt open collaboration; I remember when all open source cryptography had to be developed outside the US due to ITAR.
IMHO, models by US companies are the biggest security risk so I'm fine with using models on this "blacklist."
So... anybody who was hoping for CXMT (or YMTC) to maybe cause RAM or flash prices to maybe drop, maybe just a bit, pretty please, can go pound sand? (YMTC of course is already on the Entity List.)
> US. has held off adding.. memory chipmaker CXMT

Since CXMT was _not_ blacklisted, this is good for global RAM supply.

Becoming such a sore loser. Historians will probably look this as the most shameful period of the American empire.
What an amazing achievement by America's adversaries.

The Trump administration lists Anthropic as a security risk and kneecaps its best model, despite the fact that compared to the other frontier US labs Anthropic is more transparent, more safety-oriented, frequently honest to a fault, and is clearly acting with patriotic intent.

Meanwhile, the same administration is hesitating to counter certain Chinese companies' efforts of industrial-scale theft and sabotage due to a fear of angering the CCP!

This administration has it exactly backwards. 4.5 months until election day, 7 months until the next Congress is sworn in.

Is DeepSeek really behaving different than other Chinese companies? Intellectual theft is ongoing and has been ongoing for decades. Besides security risks and foul play, it is impressive by just how much DeepSeek undercuts OpenAI and Claude. DeepSeek charges $0.87 per million output tokens compared to $50 for Fable and $30 for GPT-5.5.
If Chinese LLMs are successfully making people in the west defend China, then I think we have all the evidence we need to explain why they are giving away their models.

The next step of course will be to get people using that ungodly cheap AI on Chinese servers. Which will also be defended because "I would never trust an American Lab".

Talking with an American trying to sell me on "don`t use deepseek because of moral reasons" really made me much more invested in China.
These bastards already prevent me from buying a BYD car, and a xiaomi phone, and they are adamant about me not using a chinese AI model. I hope they do not succeed.
I am using a Xiaomi phone. The hardware is fine, the camera is very good compared to the average, the software absolutely sucks, bootloader unlocking is not possible and it comes full of spyware and adware that you can't disable.
USA is blacklisting all Chinese companies
Why would my country blacklist DeepSeek? Perhaps crazy lunacy like: "Your product is too good and too inexpensive: consumers like US companies and individuals need to pay more for services."
Lol. That will do it.

Fable is getting more attention now precisely because it was taken down. Do the same to DeepSeek and Z.ai, and you will not strengthen your AI labs. You will likely achieve the opposite.

Wait, can't this be reduced to:

"American AI is already trillions of USD underwater, so let's use US gov to build us a moat by making competition illegal"

So this is the other side of banning American models for non-Americans? And how exactly do they plan on enforcing all of this? Great Firewall of America?

This is a complete joke. The malicious clowns behind this should be removed from power and prevented from ever holding any position of power in any form of governance system.

> And how exactly do they plan on enforcing all of this? Great Firewall of America?

That's what Cloudflare will be for.

Banning American models for non-Americans seems like a big reduction of the potential customers for American chatbot companies.
It's likely the elected officials know nothing about AI the technology.

Depends on who has their ears in terms creating policies around technology.

The same way you can't just open a bank account anonymously. It seems inevitable to me that this is going to end up with a KYC/AML style digital online ID.

Without that you are not using Claude 10. It would not be shocking at all at some level of capability that running open source models anonymously is made a crime.

To pretend there is not huge issues with a laissez faire attitude as these models progress I think is intellectually dishonest and performative.

The US is slowly becoming more like China. From talks of nationalizing companies to make US state owned entities to banning foreign competition. It's just so strange how you become the thing you fear.
I think America is becoming more like Russia, with a dictator who eliminates any billionaire who disagrees with His Dictatorship. The American groups who want to go back to the "glory years" of the 1950s are equivalent to the Russian groups who want to go back to the "glory years" of the Soviet Union.
I wish we were becoming like china; it feels a lot more like big brother from 1984.
Does anyone have the actual list of 100? Maybe I missed it but I don’t see the link or list in the article.
Providing the whole list would probably allow the government to identify the person who leaked it to Reuters.
I use DeepSeek every day (via VSCode Insiders and Zed Editor). It's very affordable and, while it's slightly behind Claude (not sure how far behind Fable), it suits my working style well. I'm not using unsupervised multi-agent workflows and don't need a library of skills files - I'm writing most of the code and leaning on AI to help with mundane tasks - like;

- generating types for APIs

- generating boilerplate based on existing code

- improving existing code (adding error handling, timeouts, things like that)

- Writing SQL repository boilerplate / queries

- Creating implementations against hand written tests

- Helping me understand and implement APIs from third party libraries

- Writing documentation

I've spent like $2 in the last month and have used over 100 million tokens.

It's doubled my productivity and unlocked work that I could not have done before.

As an Australian, I'm not sure that I care about the safety of my data when it comes to LLMs. US companies already stole scores of data to train their models on and it's hard to imagine they suddenly grew some integrity. I'll care when regulators step in, until then it's out of my control so I'll just use the best price-to-productivity product available.

> US companies already stole scores of data to train their models on and it's hard to imagine they suddenly grew some integrity.

Can you be specific about what data of yours were stolen by who?

I'm in favour of sending as much data as possible to DeepSeek for them to train on. I'd happily zip up everything useful I have and send it to them if they asked for it.
I can second this. Deepseek is great for mundane work day tasks.

I use it via claude code, just pointing the api to deepseek.

It's also not a clear "Opus 4.8 >> DS 4 Pro", I've done 16 tasks in 4 days across the two, and while Opus was indeed on average better, both models performed well being able to handle most of my workload.

In fact DeepSeek was _significantly_ better on 3 task out of 16 and Opus was _significantly_ better only in 2 out of 16.

So why I still claim Opus 4.8 to be the winner? Because the few times that DS failed or got off the rails, it failed much harder and needed several prompts to be realigned on the actual tasks.

Another thing at which Deepseek is significantly behind is code reviewing. Opus is more intelligent/thorough, Deepseek will sometimes generate bogus or low quality feedback.

And the last thing at which Opus is better, period, is vibe coding. If you want to implement features end-to-end it handles ultracode flows quite better. I don't vibe code at work, but I do so with personal projects.

But the cost concern is real. I've spent sub 2$ in 5 days of work using DS 4 Pro, which is on average just 4 queries to Anthropic.

Give me a slightly better DS 4 (it is still in preview and training isn't finished) and I may ditch Anthropic for good.

The fact that America has yet to figure out the importance of soft power (and TRUST in particular) in the AI/information age is mind boggling to me.

The CLOUD act, FISA rulings, the Snowden leaks, and now the aggressive tech oligarch push to weaponize the unholy combination of AI, MAGA, and social media algorithms in an absurd and patently obvious attempt to impose (or maintain?) "world domination" seems likely to cement their downfall.

"Authoritarian" China's low-cost and open source approach looks downright democratic by comparison.

It seems increasingly clear to me that AI is forcing a reckoning in terms of how we interpret authority and control in light of new ways that information is evolving. The old political labels seem grossly insufficient to describe the present reality and the ones whining about "democracy", "freedom", and "liberal values" increasingly sound (and act) like bitter, out of touch old men desperately clinging onto a world that is rapidly outgrowing them.

Last time I checked, the closest approximation of liberal values are "liberty, equality, fraternity". They seem to think that "liberty" should only apply to them, "equality" is a threat to their consolidated power, and "fraternity" is something that should be weaponized to turn communities against one another to distract from the ways the oligarchs have been and continue to abuse and consolidate their power.

You can thank the us gov for a list of companies to invest in.
Damn, USA, you being foolish.

The world economy is not a zero sum gain, we all do better when we do better

Cooperate with China, love and cherish them.

Unlike us of European heritage, they are not imperialistic

Get over yourselves!

Deepseek is awesome for home projects, I have used it quite a bit and the 10 bucks I loaded is yet to run out. Quality is close enough to Sonnet that with a bit of extra focus, I am not complaining.