If alignment is unsolved, capability breakthroughs are bad news, researcher warns

sjgadler · x · 2026-09-09

Citing OpenAI's Chief Scientist's statement from just two days ago that alignment remains unsolved, the author argues that massive capability breakthroughs are extremely bad news if alignment hasn't caught up. He points to a "serious missing mood" amid the industry's celebration of new capabilities: "What are we doing here???" A pointed critique of frontier labs racing on capability while openly admitting their safety work is incomplete.

Related event: OpenAI Chief Scientist Warns of AI Future as Alignment Remains Unsolved(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →