David Silver

Their best minutes

• 4:45

  Training Data

  Dropping human data let AlphaZero beat two years of work in hours

  David Silver & Lasse Espeholt · Ineffable Intelligence

  27:19of 1:08:03 · Oct 13
• 3:58

  Training Data

  Simple rewards in complex worlds beat hand-crafted RL objectives

  David Silver & Lasse Espeholt · Ineffable Intelligence

  16:09of 1:08:03 · Oct 13
• 3:27

  Training Data

  Chain-of-thought monitoring will fail as agents grow more capable

  David Silver & Lasse Espeholt · Ineffable Intelligence

  47:38of 1:08:03 · Oct 13