Sep 24, 2025 · 28m · allin
Eric Schmidt on AI, the Battle with China, and the Future of America
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
At the All-In Summit, former Google CEO Dr. Eric Schmidt engages in a wide-ranging conversation with the hosts covering the global AI race, geopolitical competition with China, modern drone warfare in Ukraine, and the future trajectory of artificial general intelligence. He highlights how American leadership must combine scientific innovation, work ethic, and high-skilled immigration to maintain its competitive technological advantage globally.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. The hosts hold 26% of the talking time here. How this is scored →
speaking balance: gold is the hosts, purple is the guest (3 minute bins)
Schmidt directly rejects Sacks's assertion that Washington and industry are pulling back from AGI, stating flatly that he disagrees with Sacks on the topic.
Hardest push from the hosts ▶ 26:08 Sacks challenges doomerism with 'middle to middle' frameworkSacks refuses to accept the narrative that AI will replace humans wholesale, arguing that AI operates middle-to-middle and requires human framing and prompt iteration.
Biggest teaching moment ▶ 6:40 Schmidt details technical training precision of DeepSeekSchmidt educates the hosts on DeepSeek's technical breakthrough in reinforcement learning using 8-bit and 4-bit numeric precision compared to Western 16-bit standards.
The host holds their own ▶ 26:08 Sacks articulates original conceptual AI frameworkSacks demonstrates host expertise by contrasting human end-to-end biological task execution with current AI middle-to-middle capability.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The hosts as informed peer | Guest teaching | Guest disagreement | The hosts pushing back | Why |
|---|---|---|---|---|---|---|
| Reconnecting and Stories from Google Days | 2 | 3 | 2 | 1 | The episode opens with light banter and reminiscing about Google and flying private jets. Schmidt offers a defense of in-office work over remote work for young professionals and highlights China's intense 996 work culture. | |
| U.S.-China AI Rivalry and Open vs. Closed Source Models | 3 | 6 | 2 | 2 | Sacks frames a White House policy question on China, and Chamath raises Meta's open source strategy. Schmidt educates the hosts on China's open-weights strategy and details DeepSeek's technical lower numeric precision (FP16 vs FP8/FP4). | |
| Relativity Space and the Physics of Rocket Science | 2 | 5 | 2 | 1 | Chamath asks about Schmidt taking over Relativity Space. Schmidt schools the hosts on rocket physics, explaining that propellant makes up 80 percent of a rocket's total weight. | |
| Ukrainian Innovation and the Evolution of Drone Warfare | 3 | 6 | 3 | 2 | Schmidt shares insights on drone warfare from Ukraine, contrasting a five thousand dollar drone with a thirty million dollar American tank. When Friedberg suggests the last remaining drone inventory wins, Schmidt firmly shuts it down, stating that war has no winners and is entirely lose-lose. | |
| AI Deterrence, Autonomous Fleets, and Naval Warfare | 2 | 5 | 2 | 2 | Chamath asks if drone automation creates detente, but Schmidt corrects him that it creates deterrence through penalty inflation. Schmidt also dismisses Jason's question on humanoid robots, pointing out that fast aerial and naval USVs are the present reality. | |
| American Exceptionalism and Global Technological Leadership | 4 | 4 | 1 | 2 | Schmidt expresses optimism for American capital markets and university innovation. Friedberg brings up demographic decline and cultural issues raised by Alex Karp and Tucker Carlson, prompting Schmidt to share birth rate stats in Korea and China. | |
| AI Progress, AGI Realities, and Human-AI Collaboration | 6 | 5 | 4 | 5 | Schmidt explicitly disagrees with Sacks's premise that the AGI narrative has pulled back, estimating superintelligence savants in six to seven years. Sacks pushes back with an elaborate framework arguing AI is 'middle to middle' whereas humans are 'end to end'. |