Into the Impossible With Brian Keating Will AI Replace Scientists?
47 snips
Mar 6, 2026 A wide-ranging conversation about how AI is reshaping physics research and the limits of machine creativity. They tackle whether silicon intelligence can match human insight and where predictive models fall short. The data deluge from modern observatories and practical AI tools for triage and discovery get attention. Ethical risks, academic disruption, and strategies for orchestrating AI in real research are explored.
AI Snips
Chapters
Books
Transcript
Episode notes
Orchestrate Models Instead Of Relying On One LLM
- Use AI as an orchestra conductor rather than a replacement: orchestrate models, humans, and tools to get useful results.
- Keating runs multiple models (Claude, Gemini) and orchestrators to automate daily tasks while retaining human oversight.
No Proven Computational Ceiling For Silicon Intellect
- There is no proven hard computational ceiling preventing silicon-based systems from understanding discoverable physics.
- Keating notes theoretical universal computers can in principle compute any discoverable result given enough resources and time.
Artificial Wisdom Matters More Than Raw Intelligence
- Intelligence alone is not wisdom; AI could amplify raw intelligence without ethical judgment.
- Keating emphasizes we need artificial wisdom and formal ethics training in science, which currently lacks structured programs.
