zetalyrae mocks OpenAI's formal proofs and Anthropic's trust-me stance

zetalyrae · x · 2026-11-17

Quoting tszzl's argument that safety research can be automated while keeping verifiable results and human oversight, zetalyrae satirizes both labs: OpenAI's line is an AI proving itself safe via theorems nobody will read in an AI-made logical theory, while Anthropic's is "if you wouldn't trust Opus with your life you are a suppressive person." A pointed jab at two extremes of AI safety discourse.

Related event: Skeptics Question OpenAI's Formal AI Safety Plans(3 posts)→

Original post →

More from AGI Musings

AGI Musings channel →