Anthropic’s C.E.O. Dario Amodei on Surviving the A.I. Endgame
Hard Fork
2025/02/28
Anthropic’s C.E.O. Dario Amodei on Surviving the A.I. Endgame
Anthropic’s C.E.O. Dario Amodei on Surviving the A.I. Endgame

Hard Fork
2025/02/28
In this episode, Dario Amodei, CEO of Anthropic, joins the show to delve into the latest advancements in AI technology. The discussion centers around the newly launched Claude 3.7 Sonnet model and its capabilities, alongside broader concerns about global AI development and societal impacts.
Amodei discusses how Claude 3.7 Sonnet excels in coding and complex tasks while addressing safety concerns and misuse risks. The podcast also explores China's rapid AI advancements, emphasizing the need for liberal democracies to maintain a technological edge safely. It warns against polarized discussions on AI risks, highlighting misinformation challenges and the importance of nuanced conversations. Furthermore, it examines the emotional and societal implications of AI, such as identity shifts among coders and professionals whose work is impacted by automation. Lastly, the episode touches on AI's influence on employment, encryption issues, and creativity in design fields, urging critical thinking amidst growing AI-generated content.
15:37
15:37
Substantial probability of medium-risk stage in next three to six months
21:07
21:07
Wants liberal democracies to have a tech advantage over China.
37:21
37:21
Poor evidence of risk is harmful and undermines AI safety efforts
46:56
46:56
Assessment of civilization derailment risk remains unchanged despite technical improvements.
52:49
52:49
AI coding will augment coders' productivity but may replace lower-level ones in 18-24 months.