Nadella criticizes model labs over distillation bans and public-data training

Satya Nadella criticized AI labs that rely on public data under fair-use arguments while prohibiting customers or competitors from distilling their own models. He called the tension a "reverse information paradox" and argued that companies should control their own learning infrastructure.

The criticism lands in a sensitive part of the AI business model. Model providers want to protect their outputs and margins, while customers increasingly want portability, auditability, and the ability to improve systems from their own usage data.