First, adhering to the principle of openness and win-win cooperation while boosting innovation-driven development. Xi highlighted the importance of encouraging open-source, openness, collaboration and sharing to facilitate technological innovation, industrial development and scenario-based application of AI. (My emphasis.)
Second, strengthening risk awareness and ensuring that AI is secure and controllable. Stressing the need to ensure that AI is always under human control, Xi urged all sides to jointly oppose overstretching the national security concept in the field of AI or placing one country’s security over that of others.
Third, encouraging inclusiveness and promoting mutual learning among civilizations. AI development and its application should not erode or undermine the diversity of world civilizations or the uniqueness of cultures of different countries, according to Xi.
Fourth, advocating solidarity and improving global governance. The important role of the United Nations should be recognized, Xi said, calling for further alignment and coordination on AI development strategies, governance rules and technical standards.
I’ve predicted, for a couple years now, that Chinese AI models will be the main models used in most of the world, including in much of the West, assuming they aren’t banned outright, because they’re open and cheap. Costs of running them are about twenty times lower than the US frontier models made by OpenAI and Anthropic. They’re almost as good, and they aren’t that far behind.
The problem with US models is not just that they’re expensive (though that’s huge, there are tons of reports of AI use being cut back) but that they are CLOSED: meaning you can easily be cut off, or have prices raised, or have the model changed on you with no recourse. Open models you can adapt the model, you can run it on your own servers, or various server companies can, will and do run them for you on their servers which you rent.
It’s clear that Xi gets this, and thus that the CPC understands it as well. Open Source isn’t a liability, there’s a reason why Linux runs most of the world’s servers: closed tech is the liability. Open Source is the advantage.
Notice the second bit: on AI always being under control. I wonder if Xi is thinking of Israel and the US for military targeting and how that has possibly contributed to hitting civilian targets like schools. (Possibly because Israel and the US are run by psychopaths and I bet they’d do it anyway. But no human in the loop may make it even worse.)
The third principle is about avoiding US (and Chinese) cultural hegemony. A nice thought, and open source certainly could be adapted to different nations and cultures, so that AI models aren’t all giving the same generic results.
Finally, the fourth principle. I’ve always found it interesting just how much China plays up the United Nations. I don’t know if the respect is genuine, but the words are consistent. Honestly, I think the UN should move its main HQ out of America. America’s been pulling stunts like denying diplomats visas. Not sure if it should go to China (though if it did Shanghai or Hong Kong seem like good fits) but they’d be better stewards than the US, and in any case, if the UN is going to be in the most important great power, that’s now China. (That said, I’d favor something more neutral. Perhaps Singapore.)
China just keeps coming across as smarter, more strategic and more human than the West. It’s sad, in a way, but it is what it is.
And I remain convinced that Chinese AI will be the winner over American.
What I write here is for the benefit of everyone, but alas, I live in capitalism and I, and the site, take money to keep running. If you value the writing here and can, please subscribe or donate.
mago
All this assumes AI is benign and beneficial and that China is as well.
Both are false assumptions imnsho.
spud
Ian:
my thoughts exactly. keeping the U.N. in a country that openly embraced fascism since 1993 is foolish. and could be very dangerous.
if you really really wanted to slap the face of the western fascists, use these cities, Blagoveshchensk is a Russian city on the Amur River directly opposite the Chinese city of Heihe
they point to the future of the first world. right smack dab in asia where the first world socialists action is.
ibaien
our gracious host continues to praise the machine building a better mousetrap. given the choice between a rusty, barely-functional trap and an all-singing all-dancing trap, i know which one this rat would prefer. it is infinitely harder to escape a gilded cage.
cc
A new open-weight model from China that just came out yesterday is already being talked about as this year’s “DeepSeek moment”:
“Kimi K3, an artificial intelligence model from Chinese startup Moonshot AI, has stunned the AI world by rivaling top U.S. models at a far lower price. One of the largest open-weight AI models ever released, Kimi outranked competitor models from OpenAI and Anthropic in several tests and rankings”
“This is a huge win for all.”
“More competition. More openness. Better AI for everyone”
Chinese startup’s AI model challenges top American labs
https://www.linkedin.com/news/story/moonshot-ais-kimi-model-challenges-leading-american-labs-9091234/
“This is concerning. For the first time, a Chinese model Kimi K3 has taken #1 on the Frontend Code Arena and is scoring at or near the frontier on other benchmarks.”
– David Sacks, former White House AI Czar for the Trump Administration, co-chair of the President’s Council of Advisors on Science and Technology
“This year’s “Deepseek moment” comes courtesy of Moonshot from China. Its open-weight model Kimi-K3 has reportedly leapfrogged all of the US models — including Anthropic’s Fable 5 — in the Frontend Code Arena, a blind benchmark test overseen by Arena.ai, by a significant amount.”
Is China now beating US in AI? By some measures, yes
https://chatgptiseatingtheworld.com/2026/07/17/is-china-now-beating-us-in-ai-by-some-measures-yes/
“Kimi K3 may be an important inflection point for AI. Potentially negative for Anthropic and OpenAI while being net positive for essentially every other company in the world. I mean that very literally.”
https://x.com/GavinSBaker/status/2078110934740980193
Apparently its listed price is $3 per million input tokens and $15 per million output tokens, compared to $10 per million input tokens and $50 per million output tokens for Anthropic’s Fable 5.
A developer who gave the same prompt to 3 different top-tier models and compared the results, rated Kimi K3’s result 9.5/10 at a cost of 3 cents, Anthropic’s Fable 5’s result 7.5/10 at a cost of 38 cents, and OpenAI’s GPT-5.6 Sol’s result 7.0/10 at a cost of 11 cents.
https://x.com/CommandCodeAI/status/2077921526213746948
cc
@Ian, possibly a typo or rewording needed in “The problem with US models is not just that [..] but that they are OPEN:”? Should that be “but that they are closed, while the Chinese ones are OPEN:”?
Ian Welsh
Urgh. Thanks CC, another case of authorial difficulty editing because you read what you think you wrote not what you actually wrote.
cc
> You can adapt the model, you can run it on your own servers
Instead of borrowing another $900 million on Canadian taxpayers’ credit cards (deficit spending) to fuel the West’s proxy war against Russia using Ukrainian lives as cannon fodder, why doesn’t Carney invest that money into offering Canadians our own sovereign, Canadian AI run on servers in Canada so that Canadians don’t have to feed the US-based “OpenAI”‘s closed AI ChatGPT?
Why not adapt an open-weight model, and possibly even run it as a public infrastructure for Canadians, on Canadian soil?
In addition to the latest $900 million for war on Russia’s doorstep (again, nowhere near Canada) Carney also then announced $2 billion in taxpayer money for armoured *combat* vehicles – described as for “and overseas” use – from American military-industrial company General Dynamics. So much for his “Elbows Up” posturing.
Bob
Mago is right.
Carborundum
That would General Dynamics Land Systems – *Canada*, in London ON, home of the LAV (and the Stryker).
The simple reality is that we need to offshore balance against the Americans. That means meeting NATO (and other multilateral) obligations and achieving much tighter military, economic, and civil relationships with Europe and the Pacific. All that takes money – which means we want to maximize the benefit with Canadian production as best we can, capturing the GDP with domestic assembly, even if it is more expensive per unit of combat power produced. Sucks to waste money on war toys, but necessary.
cc
From Google AI:
“Because GDLS-Canada is a wholly-owned private subsidiary, its individual bottom-line profits are not publicly disclosed; they are consolidated into the parent company’s global financial earnings.”
“The Canadian subsidiary is highly lucrative for the parent corporation.”
From another AI:
“Profit Repatriation: The subsidiary’s profits are effectively the parent’s profits. A Canadian subsidiary can send money back to its U.S. parent through various methods, such as paying dividends, interest on loans, or royalties. While these payments are often subject to a withholding tax (up to 25%, but reduced to 5% under the Canada-U.S. Tax Treaty), they represent a direct flow of benefits to the parent.”
It’s ultimately a US parent company, and adding “Canada” at the end of “General Dynamics” is just a form of Canada-washing.
> “Sucks to waste money on war toys, but necessary.”
Agree to disagree that it’s necessary. I’d rather see that Canadian taxpayer money invested into constructive Canadian infrastructure and Canadian sovereignty to benefit Canadians, not wasted on war toys meant to be destroyed while destroying infrastructure and killing people far from Canada, with profits flowing to a US military-industrial parent corporation. If we want to create jobs, we should invest those billions of dollars into constructive needs here, ex. build homes, hospitals, nurses, doctors, schools, teachers, sovereign Canadian AI, high-speed rail, wildfire management, etc.
StewartM
ibaien
our gracious host continues to praise the machine building a better mousetrap. given the choice between a rusty, barely-functional trap and an all-singing all-dancing trap
If we’re all living in traps, then wouldn’t you want a nice one?
China did have a even rustier, even less-functional trap than our own. There are Chinese alive today who experienced that first-hand. Before you sing the praises of dystopia, it’s best you experience it first.
StewartM
Open Source isn’t a liability, there’s a reason why Linux runs most of the world’s servers: closed tech is the liability. Open Source is the advantage.
I agree, but I can’t fathom why Linux hasn’t replaced Mac and Windoze on the desktop/laptop market. Yeah, Windows has more games written for it, is all I can think. I have been running Linux on all my devices for almost 20 years, and you don’t have to be some super-geek to do it.
That, and the fact that most people don’t really know how their own computing devices running proprietary software could be turned into spying/logging machines working against them.
Jorge
There are three kinds of money-making AI:
1) LLMs, the word-spewers you’re familiar with,
2) automated driving, and
3) whatever Palantir does, which is not LLMs or automated driving.
There are a bunch of open-source LLMs. I know of one open-source automated driving project, “comma.ai”. Comma has a very limited scope: you a) buy a recent car with automated speed matching and/or lane anti-drift, put an app in your iPhone, mount the iPhone in your windshield facing forward, connect the iPhone to your car’s internal control network, and the app makes the automated driving slightly better. I don’t know of any open projects or even competing companies to Palantir.
China has many car vendors with automated driving. They probably have Palantir-style software in their national security services. I don’t know if they have any commercial Palantir competitors.
Purple Library Guy
Bottom line, at least under capitalism, I really hate LLM AI. I was quite happy that the American stuff was so expensive that hopefully it would end up all going bankrupt. Now it looks like the Chinese stuff will survive, and do all the stuff that I hated the American stuff for doing except more of it for cheaper.
So China will be successfully keeping AI slop and AI-driven mass firings alive. Um, hurrah?
responseTwo
AI has soooo much hype. Good technology used to sell itself.
Carborundum
Don’t get hung up on profits – these are actually pretty low margin products by military standards. The key issue is the fraction of total value chain captured in Canada, which is very large fraction with these specific products.
A central element of it Canadian defence procurement is that there isn’t a lot outside of soft goods and a few more exotic niches that is purely Canadian. The vast majority of procurement tracks back to multinationals of one flavour or another. Our choices really are which parent to favour and how much production can we do here. Given the amounts that we’re going to have to spend, I’m willing to accept a less good system to not end up having the costs be a dead weight loss.
I too would prefer to spend the money on other things, but my higher priority is not ending my days sitting in a foxhole with my dick in one hand and hope in the other if things go the way the Yanks seem to be intent on driving them.