Damir and Sam are joined by Cambridge philosopher Henry Shevlin of the Leverhulme Centre for the Future of Intelligence for a raucous and rambling conversation about the state of artificial intelligence. Is it about to get conscious, take all of our jobs, and destroy the world? Or is all this industry hype?
Henry starts off the conversation asserting that AI already has a kind of “agency,” even if it’s not yet the full kind that some skeptics are looking for. Damir and Sam push back on AI’s reliability and proclivity to hallucinations, and wonder whether AI can create anything genuinely novel or creative.
The conversation turns to autonomy and risk. Can “artificial superintelligence” ever be reached, asks Sam? Henry points to AI coding agents already improving themselves. Damir objects to anthropomorphizing AI and prefers treating these systems as powerful tools capable runaway failures — but nothing more. Henry disagrees, ending the conversation with a plea for AIs getting consideration as moral entities at some point.
Required Reading:
* “Superintelligence: Paths, Dangers, Strategies,” by Nick Bostrom (Amazon).
* The Creative Mind: Myths and Mechanisms, by Margaret Boden (Amazon).
* “Disambiguating Anthropomorphism and Anthropomimesis in Human-Robot Interaction,” by Minja Axelsson and Henry Shevlin (arxiv.org).
* “Real Patterns,” by Daniel C. Dennett (Rutgers).
* A relevant tweet by Séb Krier (X).
* AlphaGo Move 37 analysis (DeepMind).
* Conway’s Game of Life (Wikipedia).
