Catherine Thorbecke writes:

In addition to being an ardent advocate of selling hardware to China, [Jensen] Huang was also one of Silicon Valley’s most powerful champions of open models, where Chinese firms have concentrated their bets.

In July, he used his first-ever X post to rally America’s tech industry behind open-weight AI after fears mounted of a US ban on Chinese offerings. He has defended distillation, or training one model on the outputs of another, as a “widely used technique” and “competition.” The distillation debate has become central to efforts to restrict Chinese open models in the US amid accusations from companies and lawmakers that Chinese firms are shoplifting American AI progress.

Without one of the industry’s most influential voices speaking up for distillation and open models more broadly, pressure for a crackdown will grow. This would hurt the US startups and developers who rely on these Chinese models for cheap, customizable tools and even cybersecurity defenses. Over time, it could further concentrate power among elite Silicon Valley AI labs. But the immediate pain would be felt by Chinese AI outfits that just lost their biggest cheerleader.

And remember when we were mostly preoccupied with AI killing us all? This scandal broke mere weeks before US-China AI safety talks were set to take place, extinguishing already faint hopes for cooperation on that front. The Nvidia crisis will deepen mistrust and reduce visibility on both sides. The result is likely a more blind and reckless race for dominance. [Continue reading…]