https://x.com/commiepommie/status/2122240808229507283
The US says it’s protecting the world from Chinese AI: However, the world is already choosing China’s open models
OpenAI was founded in 2015 as a nonprofit to “benefit humanity as a whole, unconstrained by a need to generate financial return.”
Today it’s a US$500 billion company that keeps its best models behind a paywall.
Meanwhile, China wrote open-source AI into national policy.
The AI+ plan, the 15th Five-Year Plan and the Global AI Governance Action Plan all call AI an “international public good” and promise to help the Global South build its own capacity.
The results speak for themselves.
By March 2026, Chinese open models had 1.15 billion downloads to America’s 723 million.
70 percent of new spin-off models are now built on Chinese models, while Meta’s share fell from 44 percent to 11 percent.
Alibaba’s Qwen alone has over 151,000 spin-off models, 4.7 times Llama’s footprint.
DeepSeek trained its V3 model for US$5.6 million and gave away the weights under an MIT licence. Anyone can download them, run them locally and send no data anywhere.
So what is the US doing?
It’s taking a 25 percent cut on Nvidia chip sales to China, banning Chinese models on government devices and pushing allies onto the “full US AI stack”.
OpenAI even urged Washington to consider banning Chinese models in allied countries.
They call it safety and national security.
However, every rule they ask for happens to protect their profits.
The US still has a narrow lead at the very top, in closed and expensive models protected by chip bans.
China is building what the rest of the world can actually use.
That isn’t a technology gap. It’s a difference in values.
If American AI is really for humanity, why does it need a paywall, a chip ban and a ban on the competition?
https://x.com/commiepommie/status/2122240808229507283
0 Comments:
Post a Comment
Subscribe to Post Comments [Atom]
<< Home