What OpenAI and Anthropic Think Happens Next With AI
Summary
This episode covers reports of the US government negotiating equity stakes in major AI labs to fund public dividends, alongside OpenAI’s new efficient “Dreaming” memory system and TSMC’s warnings about persistent chip shortages. The main analysis contrasts Anthropic’s essay on recursive self-improvement with OpenAI’s policy blueprint for democratic AI governance. It also examines upcoming model releases from both companies and the political dynamics surrounding new bipartisan AI regulation efforts in Congress.
Topics discussed
Intro, sponsors, and show announcements
US government reportedly seeking equity in AI firms
OpenAI launches 'Dreaming' memory system
TSMC CEO on chip shortages and pricing
Sam Altman's new AI lab and leadership
Anthropic's Mythos launch and GPT-5.6 rumors
Sponsor reads: KPMG, Section, Zencoder, OutSystems
Anthropic's RSI paper: AI building AI
Three future scenarios for AI development
Criticism of Anthropic's safety messaging
OpenAI's policy blueprint and US AI regulation
Listen ad-free on Castria