5 Comments
User's avatar
Paul Triolo's avatar

Nice. One quibble. Hugging Face did not turn to GLM-5.2 to defend itself but to just run his forensics on the rogue attack. Both US models read the request to do forensics as interest in offensive cyber ops and refused to assist. GLM helped process reams of incident logs to assist Hugging Face in determining what had happened...important distinction..

Alberto Romero's avatar

Thanks Paul, will edit that part

Think AI's avatar

Strong argument. The US should investigate genuine IP theft, but accusations without public evidence cannot replace a competitive strategy. If Chinese open models are winning users through cost, access, and flexibility, restrictions may only make American AI look defensive. The better response is to build stronger open models at home. Confidence should be demonstrated through innovation, not blocking competition.

Lotus Rose's avatar

Good article. I borrowed a quote for my own post today, but corrected the wording slightly. Was this what you meant?

“The world using Chinese AI because American AI is expensive, or close-capped, or lobotomized, or safeguarded, or not allowed overseas is a total civilizational victory.”

Evan Gelders's avatar

Fantastic rationale