Scaling Laws Live from Ashby: Taking a Long View on AI Governance with Austin Carson and Caleb Watney
Mar 10, 2026
Austin Carson, founder of Seed AI and regional AI builder, discusses public compute access and place-based adoption. Caleb Watney, policy researcher at the Institute for Progress, highlights meta‑science and team-focused funding. They talk about public compute infrastructure, immigration’s role in innovation, diversifying science funding, and making long-term institutional reforms politically viable.
AI Snips
Chapters
Transcript
Episode notes
Treat Science Funding As An Experiment
- Meta-science applies the scientific method to how science is funded and organized to test new institutional forms empirically.
- Caleb Watney recommends experimenting beyond project grants toward models that support interdisciplinary, infrastructure-intensive, team-based research.
Fund Team Based Science Not Lone Genius Projects
- Fund large, team-based research and create roles for specialized technicians to match modern interdisciplinary science needs.
- Caleb suggests grants and institutions that support 20–50 person teams and dedicated specialist roles like electron-microscope operators.
Diversify Science Funding Like An Investment Portfolio
- Science funding should be a diversified portfolio: big-team, small-hobbyist, and flexible grants to capture varied innovation modes.
- Watney warns current grant sizes are awkwardly between too small for big teams and too big for hobbyist experimentation.