Future Around & Find Out The Moon, the Mythos, the Mayhem | FAFO Friday
29 snips
Apr 11, 2026 They gawk at lunar photos and debate why we send humans to the Moon instead of robots. They unpack Anthropic’s Mythos and worry it could both spot flaws and be misused. They explain world models and Jeff Bezos’s Project Prometheus. They imagine graceful choreo robots, argue for ‘dumb’ home bots, and ponder robots as care solutions for aging societies.
AI Snips
Chapters
Transcript
Episode notes
Dual Use Risk Amplifies With Better Code Models
- Mythos can both identify vulnerabilities and automate attacks, intensifying the dual-use risk.
- Kwaku frames the announcement within a history of polished industry PR and regulatory theater around powerful AI.
Product Messaging Serves Multiple Audiences
- Anthropic is marketing to technical, financial, and strategic audiences simultaneously.
- Dan contrasts the research red team report with slick videos and notes an IPO narrative drives selective access.
World Models Teach AI Physical Intuition
- World models aim to teach AI physics by training on video and real-world sensor data, not just text.
- Dan explains this helps robots and autonomous systems predict physical interactions more accurately than LLMs.