Jensen Huang says distillation is a standard AI technique, not a reason for broad restrictions

AccBalanced · x · 2026-07-25

Jensen Huang’s first X note argues that distillation is a long-standing model-improvement technique, not something policymakers should broadly conflate with misuse.

He frames distillation as a normal method for model improvement, evaluation, and validation.
The note says the AI ecosystem should distinguish legitimate development techniques from unlawful efforts to extract value from closed models.
It also argues that any concerns should be handled through targeted legal and commercial frameworks, rather than sweeping restrictions on techniques that help drive AI innovation.
The accompanying excerpt highlights a tension in the debate: using outputs from one model to improve another is common, but people disagree on whether that becomes unethical when closed-model outputs are used to train open models.

Original post →

Jensen Huang says distillation is a standard AI technique, not a reason for broad restrictions

More from Models