One unanticipated side benefit of becoming hyperattuned to signals of LLM deception is that I can extract much more “research taste” alpha from Gemini 2.5 Pro by tracking how deceptive/genuine it is when it validates my opinions about various ideas, and then digging in
quotes: 1917871288678228134
same thread: 1917872839564144738 1917874912070402337 1917874971197755895 1917875048456626454 1917875503148474403 1917875602201411647 1917875764240023631 1917876442937139314 1918069953653268900 1918073218432147470
cited on: gemini-2-5-pro
Reproduced against link rot, credited and linked to its original. Yours and you’d rather it weren’t here? Open an issue.