Interview with Alan Kay
summary
Argues that a change of point of view is worth 80 IQ points, shows how Xerox PARC’s success came from integrated hardware and software (microcode, B5000 lineage, Simula ideas, Smalltalk/Squeak prototypes) and warns that current LLM hype repeats past fads unless paired with vetting systems; cites real failures (airline autopilot crashes, major site outages) to demand software safety checks—an engineering do no harm principle—and champions McCarthy/Simula-style pseudo-time functional models and “worlds” for undo as concrete paths toward verifiable, scalable systems.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 1:19Points of view
- 7:06AI and scientific methods
- 14:15Smalltalk and hardware
- 19:11Research and development
- 27:02Software integrity
- 36:15Functional programming
- 44:24Learning and critique
- 52:55Limits of progress
