Moonshots with Peter Diamandis Google Invests $40B Into Anthropic, GPT 5.5 Drops, and Google Cloud Dominates | EP #252
905 snips
May 4, 2026 A fast tour through the AI arms race, from GPT 5.5 and Kimi K2.6 to the battle over chips, power, and cloud dominance. It also dives into self-hosted models, always-on assistants, deepfake-proof identity, AI reshaping jobs and medicine, and robotics finally moving into daily life.
AI Snips
Chapters
Books
Transcript
Episode notes
Anthropic Is Optimizing For Economic Value Per Token
- Anthropic’s product experiments are less about killing SaaS companies than about finding tasks with the highest economic value per token.
- Alex Wissner-Gross said coding won because useful code earns more per token than cat images or video, and marketplace experiments test other high-value workflows.
Ambient AI Will Be Powerful Because It Sees Everything
- Ambient AI helpers become vastly more useful when they can watch everything on your screen, but current implementations create major privacy and security tradeoffs.
- Peter H. Diamandis compared OpenAI Chronicle to Microsoft Recall, while Alex Wissner-Gross called screenshot-to-server workflows an architectural atrocity that belongs in the OS.
Deepfake Fraud Already Looks Like Everyday Business
- Deepfake fraud is already operational, not hypothetical, and even routine business calls can be convincingly faked.
- Peter H. Diamandis cited a $25 million Hong Kong scam done with AI coworkers, and Dave Blundin recalled staff wiring money after a fake message that sounded like him.
