Mistral CEO Arthur Mensch: powerful AI models will be returned to the commons

victormustar · x · 2026-11-20

Mistral co-founder Arthur Mensch says he is confident humanity will ensure powerful AI models are returned to the commons.

The remark responds to OpenAI researcher ibab's argument that LLM pretraining distills humanity's collective work — writing and open-source code released as a common good — into model weights, which AI companies then access in trillions of tokens, train into powerful models, and sell back to the public for hundreds of billions of dollars. ibab argues this ethical dimension deserves far more attention; Mensch is optimistic the models will ultimately return to the commons.

Original post →

More from AGI Musings

AGI Musings channel →