AI safety skeptic says more deployment is how we engineer bad behavior out of models

ctjlewis · x · 2026-07-22

A quoted reply argues that even if the worst-case doom scenario were true, it would not change the answer: more deployment is more safety.

The claim is that wider real-world use helps engineers discover and remove undesirable model behaviors, much like how early steam engines were improved after many boilers exploded.

Original post →

More from AGI Musings

AGI Musings channel →