This is a nice benchmark for AI R&D

LLMs are closing the gap to humans

Details: metr.org/AI_R_D_Evalu...