Kimi: A threat or a threat? | TechCrunch

Chinese company Moonshot AI released a new version of its Kimi model this week, generating another wave of discourse about China and open source AI.

Moonshot said That although the Kimi K3 “still follows the most powerful proprietary models, Claude Fable 5 and GPT 5.6 Sol,” the new open-source model “demonstrated border-level performance across our evaluation suite, consistently outperforming other models tested.” Independent analyzes of Arena.ai and Waltz I It was also suggested that Kimi is competitive with leading frontier models.

The announcement that coincided with Speech by Chinese President Xi Jinping At the World Artificial Intelligence Conference in Shanghai, he appeared to have alarmed Wall Street The Nasdaq falls about 1%. on Friday as investors sold shares in chip companies such as Nvidia.

Many of the resulting posts from tech industry figures will sound familiar to those who remember the debate after another Chinese company, DeepSeek, It released the R1 model open source in January 2025. Except now, everything seems to have increased after that The Trump administration’s tariff war with ChinaThe repeated battles are over The national security threat that anthropologists supposedly poseAnd as Major AI companies are preparing to finally go public.

For example, David Sachs—a former AI czar in the Trump administration who now serves as co-chair of the President’s Council of Advisors on Science and Technology— Contrasted with Kimi’s progress With the US “tying itself up in knots: politicians and bureaucrats ban new data centers, impose regulatory restrictions on states, and pressure new federal agencies to pre-approve frontier models. This is how you lose the AI ​​race.” (The news also gave him an alibi To take a dig in Anthropic, describing Claude as an example of “awakened lobotomist archetypes.”)

And former CEO of Uber Travis Kalanick echoed the complaints that the Chinese “distill” (i.e., Training is based on outcomes) American artificial intelligence models.

“If drip isn’t enforced against everyone, then everyone should be able to drip from everyone else. Otherwise, one arm (will be) tied behind the backs of American models,” Kalanick wrote. (Of course, the American models were also built on top of the Chinese models, Specifically Kimmy.)

Meanwhile, OpenAI’s Head of Strategic Futures, Dean Paul He said That Kimi is a “very good model” whose performance probably “cannot be explained by distillation or anything like that,” adding that he is “personally surprised that the Chinese state continues to allow open sourcing of good models, given the potential risks.”

In fact, Paul has suggested that “the likely outcome of a world dominated by the open weight model is full-blown AI communism,” where AI is treated as a “public good” that will eventually be provided by the state as a kind of “digital public infrastructure.”

“This future looks to me like a dystopian hellscape, but I’ve never met an open-weight model advocate who doesn’t ultimately acknowledge that this is where things end,” Paul said. He even suggested that the Trump administration (Which he was working onIt will eventually realize that it needs to “create significant amounts of regulatory risk around the use of Chinese open-weight models.”

“You don’t need to ban open source (one of the dumber ideas in the AI ​​policy debate),” Paul said. “You just need to direct every agency to issue soft law that will create fear, uncertainty and doubt.” A Federal Reserve advisory found there may be backdoors in Chinese AI models. “It doesn’t have to be well-justified. You’re just creating enough regulatory risk that every regulated institution backs off.

However, Shakeel Hashim, editor of the AI-focused Transformer magazine, said so Much of the worry is overblownThis is because Kimi is “unlikely to have serious cyber capabilities,” and because the Chinese government would face “very similar incentives” to restrict open Chinese models once those capabilities are developed.