Vincent Conitzer can't figure out which model powers Google AI Overviews — even by asking it
conitzer · x · 2026-09-10
AI safety researcher Vincent Conitzer notes on his Substack that it's not transparent which model Google uses to generate AI Overviews. He even tried prompting it to reveal its identity, without success — documenting the opacity as a funny AI fail.
More from Fun
- Drawing World Cup Messi live with tldraw flash — max__drake · 2026-09-10
- Dev mocks paid micro-products: an LLM can build them in 5 minutes — jh3yy · 2026-09-10
- Bengio gives AI safety talk in Tübingen — and yes, shoggoth memes made it into the slides — maksym_andr · 2026-09-10
- The Last Matplotlib Figure: AI Circle Mourns Hand-Made Charts With a Nostalgic Meme — TimDarcet · 2026-09-10
- If AI makes survival cheap, humans will just invent more games — the F1 argument — shauntrennery · 2026-09-10
- DeepSeek Agent Swarm Outcodes K3 on Complex Repos, but Burns ¥70 in 2 Hours — teortaxesTex · 2026-09-10