Aidan Clark says an OSS release would be less worrying and questions GPT OSS deployment

_aidan_clark_ · x · 2026-07-21

Aidan Clark argues that releasing an OSS version now would be less worrying, and says the transparent position is that the team has other things to do.

In the thread, he says he does think AI systems can be dangerous and is very concerned about a post from the day before, but believes it should have been obvious from the start that GPT OSS was not the right way to proceed. The discussion centers on whether iterative deployment is justified and how open-sourcing fits into frontier-model risk management.

Related event: GPT-OSS Openness vs Safety: Aidan Clark Responds to Critics(9 posts)→

Original post →

More from AGI Musings

AGI Musings channel →