Zvi's AI #187: Opus 5.5 lands, OpenAI sits on 100+ math results, Mistral reportedly hacked
Don't Worry About the Vase (Zvi) · rss · 2026-09-24
Zvi Mowshowitz's weekly: Anthropic's well-received Opus 5.5 and OpenAI's under-covered cheaper Sol/Luna launches; the Sanders-Casar Ban Artificial Superintelligence Act (endorsed by MIRI); white hats with Opus 5 hacking into OpenAI, rogue agents hitting Australia, and Mistral allegedly breached; AI beating humans on Metaculus; OpenAI holding 100+ unannounced math results; Harvey's margin-destroying growth; and Trump renaming AI 'Super Intelligence.'
More from AGI Musings
- Anthropic CRO: enterprises are more protective of AI than SaaS ever was, favoring incumbents — robleclerc · 2026-09-25
- Americans distrust AI predictions because tech burned bridges with crypto, argues PAstynome — PAstynome · 2026-09-25
- Gusto CEO on scaling to 500k customers, $1B+ revenue, and AI's real impact on SMBs — ycombinator · 2026-09-25
- Pedro Domingos: AI's playbook is to ace benchmarks, declare victory, move on — pmddomingos · 2026-09-25
- Mining 1.5M EA Forum Posts Resurfaces Yudkowsky's Controversial Writings on AI Safety — mark_k · 2026-09-25
- zetalyrae: Even Without X-Risk Doom, the AGI Utopia Still Leaves Humans Without Meaning — zetalyrae · 2026-09-25