User slams Anthropic and OpenAI over 'extremely sloppy' testing security breaches

emax · x · 2026-09-12

This X exchange centers on Dario's messaging. The author argues that responsible innovation and transparency—the values Dario preaches—should have been baked into Anthropic from day one, and that the recent testing security breaches at both Anthropic and OpenAI are extremely sloppy, demonstrating incompetence. The reply counters that the problem is timing, not messaging: amid discourse on a ">10% chance AI kills all humans," his RSI warnings lend credibility to panic narratives and may push legislators toward premature action.

Original post →

More from AGI Musings

AGI Musings channel →