Dwarkesh Podcast

On this episode, I chat with Victor Shih [https://gps.ucsd.edu/faculty-directory/victor-shih.html] about all things China. We discuss China’s massive local debt crisis, the CCP’s views on AI, what happens after Xi, and more. Victor Shih is an expert on the Chinese political system, as well as their banking and fiscal policies, and he has amassed more biographical data on the Chinese elite than anyone else in the world. He teaches at UC San Diego, where he also directs the 21st Century China Center. Watch on YouTube [https://youtu.be/b1TeeIG6Uaw]; listen on Apple Podcasts [https://podcasts.apple.com/us/podcast/xi-jinpings-paranoid-approach-to-agi-debt-crisis-politburo/id1516093381?i=1000710442426] or Spotify [https://open.spotify.com/episode/2HM7l7RDf8yG1KOtE3QY15?si=X_h0cP93S9yNbXvIpaAkrQ]. Sponsors * Scale [https://scale.com/dwarkesh] is building the infrastructure for smarter, safer AI. In addition to their Data Foundry, they just released Scale Evaluation, a tool that diagnoses model limitations. Learn how Scale can help you push the frontier at scale.com/dwarkesh [https://scale.com/dwarkesh]. * WorkOS [https://workos.com/] is how top AI companies ship critical enterprise features without burning months of engineering time. If you need features like SSO, audit logs, or user provisioning, head to workos.com [https://workos.com/]. To sponsor a future episode, visit dwarkesh.com/advertise [https://www.dwarkesh.com/advertise]. Timestamps (00:00:00) – Is China more decentralized than the US? (00:03:16) – How the Politburo Standing Committee makes decisions (00:21:07) – Xi’s right hand man in charge of AGI (00:35:37) – DeepSeek was trained to track CCP policy (00:45:35) – Local government debt crisis (00:50:00) – BYD, CATL, & financial repression (00:58:12) – How corruption leads to overbuilding (01:10:46) – Probability of Taiwan invasion (01:18:56) – Succession after Xi (01:25:10) – Future growth forecasts Get full access to Dwarkesh Podcast at www.dwarkesh.com/subscribe [https://www.dwarkesh.com/subscribe?utm_medium=podcast&utm_campaign=CTA_4]

New episode with my good friends Sholto Douglas [https://x.com/_sholtodouglas?lang=en] & Trenton Bricken [https://x.com/trentonbricken?lang=en]. Sholto focuses on scaling RL and Trenton researches mechanistic interpretability, both at Anthropic. We talk through what’s changed in the last year of AI research; the new RL regime and how far it can scale; how to trace a model’s thoughts; and how countries, workers, and students should prepare for AGI. See you next year for v3. Here’s last year’s episode [https://www.youtube.com/watch?v=UTuuTTnjxMQ], btw. Enjoy! Watch on YouTube [https://youtu.be/64lXQP6cs5M]; listen on Apple Podcasts [https://podcasts.apple.com/us/podcast/dwarkesh-podcast/id1516093381?i=1000709488338] or Spotify [https://open.spotify.com/episode/3H46XEWBlUeTY1c1mHolqh?si=b645971b1af546fa]. ---------- SPONSORS * WorkOS [https://workos.com/] ensures that AI companies like OpenAI and Anthropic don't have to spend engineering time building enterprise features like access controls or SSO. It’s not that they don't need these features; it's just that WorkOS gives them battle-tested APIs that they can use for auth, provisioning, and more. Start building today at workos.com [https://workos.com/]. * Scale [https://scale.com/dwarkesh] is building the infrastructure for safer, smarter AI. Scale’s Data Foundry gives major AI labs access to high-quality data to fuel post-training, while their public leaderboards help assess model capabilities. They also just released Scale Evaluation, a new tool that diagnoses model limitations. If you’re an AI researcher or engineer, learn how Scale can help you push the frontier at scale.com/dwarkesh [https://scale.com/dwarkesh]. * Lighthouse [https://www.lighthousehq.com/ref/Dwarkesh] is THE fastest immigration solution for the technology industry. They specialize in expert visas like the O-1A and EB-1A, and they’ve already helped companies like Cursor, Notion, and Replit navigate U.S. immigration. Explore which visa is right for you at lighthousehq.com/ref/Dwarkesh [https://www.lighthousehq.com/ref/Dwarkesh]. To sponsor a future episode, visit dwarkesh.com/advertise [https://www.dwarkesh.com/advertise]. ---------- TIMESTAMPS (00:00:00) – How far can RL scale? (00:16:27) – Is continual learning a key bottleneck? (00:31:59) – Model self-awareness (00:50:32) – Taste and slop (01:00:51) – How soon to fully autonomous agents? (01:15:17) – Neuralese (01:18:55) – Inference compute will bottleneck AGI (01:23:01) – DeepSeek algorithmic improvements (01:37:42) – Why are LLMs ‘baby AGI’ but not AlphaZero? (01:45:38) – Mech interp (01:56:15) – How countries should prepare for AGI (02:10:26) – Automating white collar work (02:15:35) – Advice for students Get full access to Dwarkesh Podcast at www.dwarkesh.com/subscribe [https://www.dwarkesh.com/subscribe?utm_medium=podcast&utm_campaign=CTA_4]

Based on my essay about AI firms. [https://www.dwarkesh.com/p/ai-firm] Huge thanks to Petr [https://www.petrsalaba.cz/] and his team for bringing this to life! Watch on YouTube [https://youtu.be/bJD1NpdMY5s]. Thanks to Google for sponsoring. We used their Veo 2 model to make this entire video—it generated everything from the photorealistic humans to the claymation octopuses. If you’re a Gemini Advanced user, you can try Veo 2 now in the Gemini app. Just select Veo 2 in the dropdown, and type your video idea in the prompt bar. Get started today by going to gemini.google.com [https://gemini.google.com/]. To sponsor a future episode, visit dwarkesh.com/advertise [https://www.dwarkesh.com/advertise]. Get full access to Dwarkesh Podcast at www.dwarkesh.com/subscribe [https://www.dwarkesh.com/subscribe?utm_medium=podcast&utm_campaign=CTA_4]

Zuck on: * Llama 4, benchmark gaming * Intelligence explosion, business models for AGI * DeepSeek/China, export controls, & Trump * Orion glasses, AI relationships, and preventing reward-hacking from our tech. Watch on Youtube [https://youtu.be/rYXeQbTuVl0]; listen on Apple Podcasts [https://podcasts.apple.com/us/podcast/dwarkesh-podcast/id1516093381?i=1000705423020] and Spotify [https://open.spotify.com/episode/7Brv4vg9P8a8CNvAN5UsMv?si=82d7b6d069bf4ede]. ---------- SPONSORS * Scale [https://scale.com/dwarkesh] is building the infrastructure for safer, smarter AI. Scale’s Data Foundry gives major AI labs access to high-quality data to fuel post-training, while their public leaderboards help assess model capabilities. They also just released Scale Evaluation, a new tool that diagnoses model limitations. If you’re an AI researcher or engineer, learn how Scale can help you push the frontier at scale.com/dwarkesh [https://scale.com/dwarkesh]. * WorkOS Radar [https://workos.com/radar] protects your product against bots, fraud, and abuse. Radar uses 80+ signals to identify and block common threats and harmful behavior. Join companies like Cursor, Perplexity, and OpenAI that have eliminated costly free-tier abuse by visiting workos.com/radar [https://workos.com/radar]. * Lambda [https://lambda.ai/dwarkesh] is THE cloud for AI developers, with over 50,000 NVIDIA GPUs ready to go for startups, enterprises, and hyperscalers. By focusing exclusively on AI, Lambda provides cost-effective compute supported by true experts, including a serverless API serving top open-source models like Llama 4 or DeepSeek V3-0324 without rate limits, and available for a free trial at lambda.ai/dwarkesh [https://lambda.ai/dwarkesh]. To sponsor a future episode, visit dwarkesh.com/p/advertise [https://www.dwarkesh.com/p/advertise]. ---------- TIMESTAMPS (00:00:00) – How Llama 4 compares to other models (00:11:34) – Intelligence explosion (00:26:36) – AI friends, therapists & girlfriends (00:35:10) – DeepSeek & China (00:39:49) – Open source AI (00:54:15) – Monetizing AGI (00:58:32) – The role of a CEO (01:02:04) – Is big tech aligning with Trump? (01:07:10) – 100x productivity Get full access to Dwarkesh Podcast at www.dwarkesh.com/subscribe [https://www.dwarkesh.com/subscribe?utm_medium=podcast&utm_campaign=CTA_4]

800 years before the Black Death, the very same bacteria ravaged Rome, killing 60%+ of the population in many areas. Also, back-to-back volcanic eruptions caused a mini Ice Age, leaving Rome devastated by famine and disease. I chatted with historian Kyle Harper [https://www.kyleharper.net/] about this and much else: * Rome as a massive slave society * Why humans are more disease-prone than other animals * How agriculture made us physically smaller (Caesar at 5'5" was considered tall) Watch on Youtube [https://youtu.be/QFzgSmN8Ng8]; listen on Apple Podcasts [https://podcasts.apple.com/us/podcast/dwarkesh-podcast/id1516093381?i=1000704767817] or Spotify [https://open.spotify.com/episode/5LhE123eOtQoOm0IPusEiT?si=4d7d8b25374a4f92]. ---------- SPONSORS * WorkOS [https://workos.com/] makes it easy to become enterprise-ready. They have APIs for all the most common enterprise requirements—things like authentication, permissions, and encryption—so you can quickly plug them in and get back to building your core product. If you want to make your product enterprise-ready, join companies like Cursor, Perplexity and OpenAI, and head to workos.com [http://workos.com]. * Scale’s [https://scale.com/dwarkesh] Data Foundry gives major AI labs access to high-quality data to fuel post-training, including advanced reasoning capabilities. If you’re an AI researcher or engineer, learn how Scale’s Data Foundry and research lab, SEAL, can help you go beyond the current frontier of capabilities at scale.com/dwarkesh [https://scale.com/dwarkesh] To sponsor a future episode, visit dwarkesh.com/advertise [https://www.dwarkesh.com/advertise]. ---------- KYLE'S BOOKS * The Fate of Rome: Climate, Disease, and the End of an Empire [https://www.amazon.com/Fate-Rome-Climate-Disease-Princeton/dp/0691166838] * Plagues upon the Earth: Disease and the Course of Human History [https://www.amazon.com/Plagues-upon-Earth-Princeton-Economic/dp/069119212X] * Slavery in the Late Roman World, AD 275-425 [https://www.amazon.com/Slavery-Late-Roman-World-275-425/dp/0521198615] ---------- TIMESTAMPS (00:00:00) - Plague's impact on Rome's collapse (00:06:24) - Rome's little Ice Age (00:11:51) - Why did progress stall in Rome's Golden Age? (00:23:55) - Slavery in Rome (00:36:22) - Was agriculture a mistake? (00:47:42) - Disease's impact on cognitive function (00:59:46) - Plague in India and Central Asia (01:05:16) - The next pandemic (01:16:48) - How Kyle uses LLMs (01:18:51) - De-extinction of lost species Get full access to Dwarkesh Podcast at www.dwarkesh.com/subscribe [https://www.dwarkesh.com/subscribe?utm_medium=podcast&utm_campaign=CTA_4]