Luca Righetti
@lucafrighetti
AI risks are hard to study — we need more transparency and rigor. Senior Researcher @GovAI_, @METR_Evals. Podcast @HearThisIdea.
We ran a randomized controlled trial to see how much AI coding tools speed up experienced open-source developers. The results surprised us: Developers thought they were 20% faster with AI tools, but they were actually 19% slower when they had access to AI than when they didn't.
I am *unbelievably* excited for all 5 hours of this! 💊🤙
LAUNCH DAY 🚀 Today I’m launching a new podcast, Hard Drugs, with Jacob Trefethen (@JacobTref) Our first episode is about lenacapavir — a new HIV drug that blocks infections with an efficacy rate of nearly 100%, and which could completely change the fight against HIV worldwide.…
OpenAI and Anthropic *both* warn there's a sig. chance that their next models might hit ChemBio risk thresholds -- and are investing in safeguards to prepare. Kudos to OpenAI for consistently publishing these eval results, and great to see Anthropic now sharing a lot more too.
