Pre‑training vs Post‑training and Compute Constraints

Join Pondering Durian and José Macedo as they dive deep into the US-China AI competition with Alex Lee, co-founder of TrueNorth and former VP at Enflame (a $3 billion AI chip startup). With his unique perspective spanning both ecosystems - from his PhD in electrical engineering to roles at Temasek and McKinsey - Alex breaks down China's dominance in open source AI, export controls on semiconductors, what a potential Taiwan conflict would mean for global chip supply chains, and insider insights on Chinese tech culture, state-led industrial policy, and why the innovation gap may be closing faster than expected.TrueNorth: https://www.true-north.xyz🎯 Key Highlights▸ China's open source AI leadership: sustainable or temporary advantage?▸ How Chinese tech giants (Alibaba, ByteDance, Tencent) leverage broad business models▸ Export controls reality: are Chinese companies actually GPU-constrained?▸ Inside Chinese AI labs: Deep Seek vs Moonshot vs Zhipu vs the big players▸ The shift from pre-training to post-training and what it means for US-China competition▸ Why the "9-9-6" work culture narrative is more myth than reality▸ Industrial policy lessons: EVs and solar success vs semiconductor struggles▸ Taiwan semiconductor crisis: what would actually happen in a conflict?▸ The future of AI accelerators and system-level innovation▸ Context engineering as the new bottleneck in AI development▸ China's robotics supply chain advantage vs US physical intelligence models▸ 2035 predictions: will the US-China AI gap close or widen?💡 Subscribe for more AI, crypto & tech insights! 🔔🧠 Follow the Alpha▸ Alex Lee's Twitter: @moonshot6666▸ TrueNorth Twitter: @get_truenorth🔗 Connect with Delphi🌐 Portal: https://delphidigital.io/🧠 Intelligence: https://www.delphiintelligence.io🐦 Twitter: @delphi_intel🎧 Listen onSpotify: https://open.spotify.com/show/0Zp0R78f6wFQoPaVr99CCPApple Podcasts: https://podcasts.apple.com/us/podcast/2035/id1832694022YouTube: https://www.youtube.com/channel/UC9Yy99ZlQIX9-PdG_xHj43QTimestamps00:00 – Intro: Alex Lee and US-China AI dynamics01:30 – Is China's open source dominance sustainable?05:15 – The economics behind open source AI development09:30 – Post-training shift: advantage US or China?11:30 – Cultural differences: Silicon Valley vs Chinese tech16:30 – Innovation cycles: zero-to-one vs one-to-x19:45 – Chinese AI lab landscape: who's winning?25:30 – Are AI startups in China "cooked"?28:00 – GPU constraints: export controls reality check31:00 – China's industrial policy: successes and failures34:45 – Land financing model breakdown38:45 – Youth unemployment and "lying flat" culture42:45 – Taiwan semiconductor supply chain crisis scenarios49:45 – Leading edge vs trailing edge chip competition55:00 – The future of AI accelerators and system innovation58:30 – Context engineering as the new frontier1:03:30 – TrueNorth: AI-powered crypto discovery platform1:09:30 – Why build in crypto vs traditional fintech1:12:30 – Robotics: China's supply chain vs US intelligence1:18:00 – 2035 predictions: leveling the playing fieldDisclaimerThis podcast is strictly informational and educational and is not investment advice or a solicitation to buy or sell any tokens or securities or to make any financial decisions. Do not trade or invest in any project, tokens, or securities based upon this podcast episode. The host and members at Delphi Ventures may personally own tokens or art that are mentioned on the podcast. Our current show features paid sponsorships which may be featured at the start, middle, and/or the end of the episode. These sponsorships are for informational purposes only and are not a solicitation to use any product, service or token.