Future Around & Find Out "Nice Model You Got There — Shame If Something Happened to It" | FAFO Friday
May 9, 2026
A heated take on sudden government moves to vet AI models and whether that’s safety or political leverage. A close look at Anthropic’s compute deal with Elon and what private control of AI infrastructure could mean. A contrasting thread about robot design: fuzzy, non-humanoid companions, dancer-informed robotics, and how design shapes trust and social impact.
AI Snips
Chapters
Transcript
Episode notes
Government Review Is About Leverage Not Just Safety
- Government review of frontier AI is as much about leverage and geopolitics as safety.
- Dan Blumberg argues Trump-era interest likely mixes CYA, leverage over big AI, and desire for early Pentagon access to models.
Compute Contracts Can Become Control Mechanisms
- Private compute deals create de facto private oversight and leverage over model developers.
- Dan highlights Anthropic buying XAI/SpaceX compute while Elon reserves right to reclaim it if he deems a model "harming humanity."
Build Home Robots With Emotional Design Not Human Likeness
- Design home robots to avoid humanoid expectations and instead emphasize EQ and approachable form.
- Dan and Kwaku cite Roomba founder's fuzzy companion and Catie Cuan's dancer-robot work as examples of non-humanoid, emotionally resonant designs.