Adaption Labs: Gradient-Free Continual Learning — Sara Hooker, Adaption
AI research frontier access requires an 'unreasonably narrow path' that excludes most talent
“the unreasonably narrow path”
Sara Hooker of Adaption Labs argues that frontier AI research has been gated behind a rigid credentialing pipeline — the right PhD, the right lab, the right advisor — filtering out vast amounts of potential talent. The talk frames this as a structural problem in how scientific discovery works, not just an AI-specific issue. The technical content on gradient-free continual learning was not captured in the available transcript, limiting the signal value of this excerpt.