AI Companies Still Haven’t Delivered on Their Biggest Promises
Summary
The episode analyzes the fallout from claims that Anthropic leaders believe they could become the sole private AI company, prompting rare public responses from CEO Dario Amodei on regulation and trust. It also covers ZAI’s release of the GLM-5.3 model, highlighting its competitive performance and cost efficiency against Western frontier models. Finally, the show discusses Anthropic’s upcoming IPO, detailing its massive revenue growth and the debate surrounding its projected $2 trillion valuation.
Topics discussed
Intro: Anthropic's 'only company left' controversy and show sponsors
ZAI GLM-5.3 release: benchmarks, cyber focus, and performance reviews
Analysis: Chinese AI labs' capabilities and Wall Street's infrastructure view
Anthropic's unreleased internal models and OpenAI's Astra rumors
Anthropic's IPO plans, $2T valuation, and financial skepticism
Sponsor segments: KPMG, Blitzy, Robots & Pencils, and HyperAgent
The Gavin Newsom vs. Dario Amodei debate on AI concentration of power
Dario's defense: Regulation, trust, and the 'false choice' argument
Dario's detailed post on regulation, scaling laws, and power concentration
Dario on curing cancer, public trust, and the need for honest risk communication
Public reaction: Critics question Dario's tone and PR strategy
Conclusion: Show, don't tell, and the value of public discourse
Listen ad-free on Castria