Canales actualizados recientemente

Elige uno para ver solo ese canalVer todos los canales →
20VC with Harry Stebbingsayer +173JSJane Streetayer +119AI Engineerayer +1硅谷101anteayerMBMatthew Bermanhace 3 días +111PragerUhace 3 díasY Combinatorhace 3 díasa16zhace 4 días +1SemiAnalysishace 4 díasGOTO Conferenceshace 4 días

Resultados

2 artículos
Ordenar
TemaPrompt injectionVoces2 voces
1:03:46EN
20VC with Harry Stebbings· 4 sept 2024Pódcast · 2 vocesZico Kolter on AI Safety, Scaling and the Limits of Open ModelsInvitado: Zico Kolter · Presenta: Harry StebbingsThe Carnegie Mellon researcher and OpenAI board member explains why prompt injection worries him, how he thinks about open weights, and why AI progress remains uncertain.AI SafetyPrompt injectionOpen-weight models+2
1:37:47EN
Lenny's PodcastPódcast · 2 vocesYou Can Patch a Bug, but You Can't Patch a Brain: Why AI Guardrails FailInvitado: Sander Schulhoff · Presenta: Lenny RachitskyAdversarial-robustness researcher Sander Schulhoff explains why prompt injection and jailbreaks remain unsolved, why guardrail products don't deliver, and what teams can do instead.SeguridadPrompt injectionAgent red teaming+3