Smale’s 18th Problem and Intelligence Science
summary
Smale’s 18th frames a concrete program: compare discrete Turing models with the real number machine (BSS), quantify complexity via polynomial time, and test universality by whether a single form of intelligence can solve general polynomial systems. Empirical anchors include 86 billion neurons as a biological constraint and milestones like AlphaGo (2016), ChatGPT and AlphaFold showing engineering gains. Specific tools discussed are Cucker–Smale models, Hutter’s optimal agent theory, randomness, approximate arithmetic, and learning via repeated games, neural nets and genetic algorithms.
AI-written summary · about
chapters (10)
- 0:00Smale’s 18th Problem intro
- 1:51Digital twin of Smale’s 18th
- 5:46Why Smale proposed limits
- 7:35Human intelligence limits
- 9:23AI progress over 20 years
- 13:09Intelligence definitions
- 15:18Machine vs human intelligence
- 16:50Intelligence and consciousness
- 17:38AI research growth
- 18:01Math for AI and AGI
related talks
58mStrong Generalization from Small Brains and No Training Data
40mWhere Machine Intelligence Ends and Human Creativity Begins
27mCan computers become intelligent through playing games?
21mLearning from History: The Past as a Compass for Systems Engineers in the Age of AI
42mReasonable AI
21mWhen the machines out-smarts us
24mThe Rise of Reasoning Models
48mThe Perils and Promise of Artificial Intelligence