
All Else Equal: Making Better Decisions Ep27 "The Future of AI" with Michal Kosinski
6 snips
May 31, 2023 How can AI read what people believe, and how quickly is it moving from clumsy chatbot to superhuman reasoner? Stanford professor Michal Kosinski discusses theory of mind experiments, GPT’s performance on cognitive tests, the nature of machine intelligence, and the challenge of regulating a technology that could reshape everyday life.
Chapters
Transcript
Episode notes
1 2 3 4 5 6 7 8 9
Introduction
00:00 • 3min
The Unexpected Transfer
02:54 • 2min
How AI Solves the Unexpected Transfer Task
04:25 • 3min
GPT-4 Tasks: How Much Has Changed Since the First Generation?
07:12 • 3min
The Evolution of Cognitive Reasoning Tasks
10:38 • 5min
How a Language Program Gets Intelligence
15:14 • 5min
The Importance of Human Brains
19:46 • 5min
The Language Predictor Model and Human Intelligence
24:18 • 3min
AI and the Disruptors of Our Lives
26:51 • 3min
