Shane Legg (DeepMind Founder) - 2028 AGI, New Architectures, Aligning Superhuman Models

<p>I had a lot of fun chatting with Shane Legg - Founder and Chief AGI Scientist, Google DeepMind!</p><p>We discuss:</p><p>* Why he expects AGI around 2028</p><p>* How to align superhuman models</p><p>* What new architectures needed for AGI</p><p>* Has Deepmind sped up capabilities or safety more?</p><p>* Why multimodality will be next big landmark</p><p>* and much more</p><p>Watch full episode on <a target="_blank" href="https://youtu.be/Kc1atfJkiJU">YouTube</a>, <a target="_blank" href="https://podcasts.apple.com/us/podcast/the-lunar-society/id1516093381">Apple Podcasts</a>, <a target="_blank" href="https://open.spotify.com/show/4JH4tybY1zX6e5hjCwU6gF">Spotify</a>, or any other podcast platform. Read full transcript <a target="_blank" href="https://www.dwarkeshpatel.com/p/shane-legg">here</a>.</p><p><strong>Timestamps</strong></p><p>(0:00:00) - Measuring AGI</p><p>(0:11:41) - Do we need new architectures?</p><p>(0:16:26) - Is search needed for creativity?</p><p>(0:19:19) - Superhuman alignment</p><p>(0:29:58) - Impact of Deepmind on safety vs capabilities</p><p>(0:34:03) - Timelines</p><p>(0:41:24) - Multimodality</p> <br/><br/>This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit <a href="https://www.dwarkeshpatel.com?utm_medium=podcast&#38;utm_campaign=CTA_1">www.dwarkeshpatel.com</a>