Jensen Huang on Distillation: AI Learning from AI is Fundamental, Smarter Means Safer

FinanceYF5 · x · 2026-08-06

When asked whether open-source model companies should be allowed to distill closed-source models, Jensen Huang gave a clear response.

He drew an analogy to everyday human learning, emphasizing that distillation is not plagiarism, but a fundamental method of intelligence. He pointed out that it is an inevitable trend for AI systems to continuously distill knowledge from other AI systems, as AI-generated content on the internet is rapidly surpassing human-created content.

Huang directly linked this distillation learning to 'safety,' arguing that it is good for AI to learn. The industry should want AI systems to become more intelligent because smarter AI can also be safer AI.

Related event: Jensen Huang Defends Model Distillation as Learning, Not Plagiarism(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →