Dario Amodei (Anthropic CEO) - Scaling, Alignment, & AI Progress

<p>Here is my conversation with Dario Amodei, CEO of Anthropic.</p><p>Dario is hilarious and has fascinating takes on what these models are doing, why they scale so well, and what it will take to align them.</p><p>Watch on <a target="_blank" href="https://youtu.be/Nlkk3glap_U">YouTube</a>. Listen on <a target="_blank" href="https://apple.co/3rZOzPA">Apple Podcasts</a>, <a target="_blank" href="https://spoti.fi/3QwMXXU">Spotify</a>, or any other podcast platform. Read the full transcript <a target="_blank" href="https://www.dwarkeshpatel.com/dario-amodei">here</a>. Follow <a target="_blank" href="https://twitter.com/dwarkesh_sp">me on Twitter</a> for updates on future episodes.</p><p><strong>Timestamps</strong></p><p>(00:00:00) - Introduction</p><p>(00:01:00) - Scaling</p><p>(00:15:46) - Language</p><p>(00:22:58) - Economic Usefulness</p><p>(00:38:05) - Bioterrorism</p><p>(00:43:35) - Cybersecurity</p><p>(00:47:19) - Alignment & mechanistic interpretability</p><p>(00:57:43) - Does alignment research require scale?</p><p>(01:05:30) - Misuse vs misalignment</p><p>(01:09:06) - What if AI goes well?</p><p>(01:11:05) - China</p><p>(01:15:11) - How to think about alignment</p><p>(01:31:31) - Is modern security good enough?</p><p>(01:36:09) - Inefficiencies in training</p><p>(01:45:53) - Anthropic’s Long Term Benefit Trust</p><p>(01:51:18) - Is Claude conscious?</p><p>(01:56:14) - Keeping a low profile</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>