Neolithic

AI safety research is bottlenecked by how long evaluations take to run. Leo McKee-Reid from Neolithic is building infrastructure to automate AI safety research: including a platform for rapidly developing and iterating on dangerous capabilities evaluations.

Biography

The AGI arms race is outpacing AI safety research — so Leo McKee-Reid is building the infrastructure to close that gap. He is the founder of Coordinal Research, developing automated R&D systems to accelerate AI safety work. A former AI safety researcher at LASR Labs and AI Safety Camp, Leo previously founded UVic AI and has a background in ML, mechanistic interpretability, and dangerous capability evaluations.