#2September 6 · 1 hr 17 min
00:00 - GoPro, Allbirds and the "add AI to save the stock" playbook
01:30 - Nvidia's $12.9B Hugging Face deal and Kevin Durant's $60M payday
03:52 - Hugging Face as AI's distribution layer, and why Nvidia paid 86x revenue
06:18 - Nvidia's real goal: defending GPU demand from custom chips
08:34 - Open source, falling prices, and the enterprise security problem
14:44 - Why premium brands keep enterprise customers, and Anthropic's IPO plans
18:09 - HBM3 memory and CXMT's challenge to the bottleneck
20:18 - The AI agent attack on Hugging Face and OpenAI, and what it means for cyber
26:40 - Did the "agent civilization" story serve an AI marketing agenda?
34:24 - Why the hack was less sophisticated than the headlines
37:37 - Bernie Sanders, AI regulation, and why US rules could help China
45:19 - Offensive vs defensive agents, and CrowdStrike's data advantage
51:54 - Meta's Muse Spark, Google's Gemini Flash, OpenAI's GPT Astra
53:55 - How Meta funds AI differently, and what employee data reveals
59:33 - Meta's platform risk and the case for owning the models
01:04:09 - Why Google doesn't need the best frontier model
01:07:14 - GPT Astra's 3D reasoning, the AGI gap, and AI in game development
01:13:59 - Tim Cook's record, the handover, and John Ternus's engineering edge
01:23:01 - Apple's on-device AI, distribution control, and valuation
01:30:13 - Mac Mini agents, memory shortages, and pricier iPhones
Highly Leveraged is a finance and tech podcast hosted by Joshua and Brian Tam. Every episode breaks down the numbers behind the market moves, the companies building the future, and the trades everyone else talks about a week late. Produced by Barebone.
Download Barebone: https://apps.apple.com/app/apple-store/id6737490098?pt=127405361&ct=Highly%20Leveraged%20Podcast&mt=8
Follow us:
Barebone IG: https://www.instagram.com/barebone.ai
Joshua IG: https://www.instagram.com/joshua.jooste_/
Barebone X: https://x.com/BareboneTeam