#248 - Fable 5, Siri AI, IPOs, Policy on the AI Exponential
Summary
Hosts Andrey Kerenkov and Jeremy Harris review the release of Claude Fable 5, noting its benchmark leaps and controversial safety safeguards. They discuss Apple’s new Siri AI powered by Google’s Gemini and Google’s price cuts for AI subscriptions. The episode also covers OpenAI and Anthropic’s confidential IPO filings, highlighting the race for capital, and details Jeff Bezos’s $12 billion raise for physical AI startup Prometheus.
Topics discussed
Intro, Groons sponsor, and episode preview
Box AI context layer and Outshift sponsor
Anthropic Claude 3.5 Sonnet and coding capabilities
AI safety, alignment, and recursive self-improvement
Anthropic's open-source strategy and model performance
Anthropic IPO rumors and safety safeguards
Apple Siri updates and Google Gemini partnership
Google AI Plus price cut and competition
OpenAI IPO filing and valuation
Bezos' physical AI company and funding
HighFlyer and Chinese AI chip investments
DeepSeek training on Huawei Ascend chips
SpaceX orbital data centers and compute demand
IceLLM and diffusion-based language models
AI safety letter on biological risks
AI policy, regulation, and economic impact
Anthropic's post on recursive self-improvement
AI agents, specification failures, and loopholes
US government equity in AI companies
Music industry lawsuits and episode outro
Listen ad-free on Castria