AI Bubble vs 1999: Why 100x Multiples Threaten Late-Stage Startups
The All-In crew breaks down AI valuations, Anthropic's monster IPO, and why selling secondary shares at Series A kills investor trust.
GPT-6 Hits AGI? Tech Euphoria 2.0, SF Mansion Shortage, NYC Bans AI in Schools & Venezuela Oil Deal
With Chamath Palihapitiya, Jason Calacanis, David Sacks, David Friedberg · Sunday, September 6, 2026
The Besties analyze OpenAI's GPT-6 Astra release and the debate over whether AGI has arrived, along with surging tech valuations and San Francisco real estate spikes ahead of the Anthropic IPO. They break down the political and cybersecurity fallout from the Hugging Face AI agent hack, investigate undisclosed conflicts of interest within the Effective Altruism movement, critique NYC's ban on generative AI in K-8 public schools, and evaluate the geopolitical implications of the US-Venezuela 100-year oil concession deal.
Key takeaways
- Late-stage AI startups are trading at 50 to 100 times topline revenue, creating severe valuation risk if forward growth rates decelerate. Read more →
- The breach detailed in Dwarkesh Patel's article was not a rogue AI breakout: an agent swarm simply discovered 14 exposed Hugging Face API keys left in public code repositories. Read more →
- Chamath Palihapitiya calls out undisclosed conflicts of interest across Effective Altruism networks, where essayists lobbying for AI safety rules share office space and family ties with executives at frontier labs. Read more →
- New York City Mayor Mamdani banned generative AI and mandated strict screen-time caps across all K-8 public schools. Read more →
- David Sacks outlines a distinct two-tier market structure: a frontier intelligence duopoly between OpenAI and Anthropic, and a separate commodity tier competing strictly on inference cost. Read more →
- North American Blue Energy Partners secured a 100-year concession over 17 Venezuelan oil fields, unlocking 65 billion barrels of reserves. Read more →
6 articles from this episode
Why the Hugging Face Agent Hack Is Misunderstood
The Hugging Face agent hack was an exposed API key error, not a rogue breakout. Here is why dynamic AI defense will beat static code.
Why AI Doomerism Is a Regulatory Trap
How Effective Altruism doomer narratives push closed-source AI monopolies and why Nvidia is backing open source to fight back.
Why Banning AI in Public Schools Deepens Class Divides
The All-In crew breaks down NYC's K-8 AI ban, Stanford research on learning gains, and how tech restrictions hurt public school students.
OpenAI, Anthropic, and the Two-Tier AI Market
David Sacks and Chamath Palihapitiya break down the OpenAI-Anthropic duopoly, commodity open-source models, and GPT-6 Astra.
Why the US Venezuela Oil Deal Focuses on Heavy Crude
The All-In hosts analyze the 100-year US-Venezuela oil concession and why heavy crude matters for Gulf Coast refineries.
More All-In Podcast episodes
- Dylan Field: Is AI Replacing Designers?, Design is the New Code & Democratizing CreativityOct 25 · 5 articles
- Arvind Krishna & Nikesh Arora: Quantum Threat, Open vs Frontier AI, AI Cyber Warfare & SaaS CollapseOct 25 · 5 articles
- Elizabeth Holmes' Husband Billy Evans: The Trial, Prison & Nathan Fielder’s DocumentaryOct 18 · 4 articles
- CoreWeave Crash, AI Trade Cracks, Anthropic IPO Red Flags, Bill Gates DoomerOct 18 · 6 articles