Summary
This episode of The Week from Prof G Media, hosted by George Hahn, covers the accelerating US-China AI competition through three distinct lenses: model capabilities and pricing, the financial sustainability of American AI labs, and global public sentiment toward AI. The central news hook is the release of Kimi K3 from Moonshot AI — the largest open-weight model ever released — which outperforms some OpenAI and Anthropic models on benchmarks while running at roughly one-third the price of Claude.
Scott Galloway argues the dynamic resembles industrial dumping at scale: free Chinese AI models have grown from under one-third of global traffic in late 2025 to roughly two-thirds recently, subsidized by local governments and cheaper compute. Ed provides specific pricing data — DeepSeek’s output token pricing sits at $0.87 per million versus $45 for GPT-5.6 and $50 for Claude Fable 5 — and argues OpenAI and Anthropic cannot sustain a price war indefinitely without venture capital support, while Meta explicitly states its mission to initiate one. Base 10 co-head of model training Charlie O’Neil reframes the story as open-source versus closed-source competition rather than a national one.
Former Bloomberg China reporter Selina Shu explains a striking data point from the episode: 84% of Chinese survey respondents say they are more excited than worried about AI, versus only 10% of Americans — a divide she attributes to pragmatic, tool-first AI adoption in China versus the “doom trolling” that dominates US discourse. The episode draws on multiple expert guests and specific market data to make it a useful digest of the geopolitical and economic forces reshaping the AI industry.
📺 Source: The Prof G Pod – Scott Galloway · Published July 24, 2026
🏷️ Format: News Analysis
