The Fabric of Knowledge - David Spivak

MLST is sponsored by Brave: The Brave Search API covers over 20 billion webpages, built from scratch without Big Tech biases or the recent extortionate price hikes on search API access. Perfect for AI model training and retrieval augmentated generation. Try it now - get 2,000 free queries monthly at http://brave.com/api. David Spivak, a mathematician at MIT's Topos Institute known for his work in applied category theory, talks with Tim Scarfe about the nature of intelligence, creativity, and knowledge itself. Spivak explains category theory in surprisingly concrete terms. Categories are about systems of relationships -- not just collections of things, but how those things relate to each other. Functors map one system of relationships to another, like how "counting" connects the world of sets to the world of numbers. He argues category theory's value lies in making the obvious things mathematically precise, which sounds trivial until you realize how much of mathematics depends on shared but unspoken assumptions. The conversation moves to collective intelligence and sense-making. Drawing on Mike Levin's claim that all intelligence is collective intelligence, Spivak describes sense-making as a process of "accounting" -- like balancing a checkbook, where different perspectives contribute until the books settle and understanding stabilizes. This applies at every scale, from neurons communicating in a shared language to two people trying to agree on what category theory means for machine learning. Where things get genuinely interesting is Spivak's take on creativity and open-endedness. He pushes back on the idea that Karl Friston's prediction error minimization framework captures everything interesting about intelligence. His counterexample: a kid shoveling sand in a sandbox who cries when pulled away. There's something about care and engagement that doesn't obviously reduce to prediction error. Questions, he argues, are more important than answers -- the act of questioning creates a spaciousness where real insight can arise. On AI, Spivak is measured but candid. He thinks current approaches are "kicking the ball really hard" without thinking about where the ball needs to go. He worries about optimization without understanding what we're optimizing for. The discussion covers embodiment and how physical experience shapes abstract thought, the role of written language in transmitting knowledge across generations, and whether the intelligence explosion is an extension of evolutionary processes or something genuinely new. --- TIMESTAMPS: 00:00:00 Introduction to Category Theory 00:04:40 Collective Intelligence and Sense-Making 00:09:54 Embodiment and Physical Concepts in Knowledge 00:16:23 Creativity, Open-Endedness, and Care 00:25:46 Modeling Creativity and the Role of Questioning 00:36:04 Evolution, Optimization, and AI 00:44:14 Written Language and Knowledge Transmission --- REFERENCES: person: [00:00:00] David Spivak - Personal Page http://www.dspivak.net/ [00:04:40] Mike Levin - Google Scholar https://scholar.google.com/citations?... reference: [00:00:00] MIT Category Theory Lectures by David Spivak    • Applied Category Theory. Chapter 1, lectur...   [00:00:00] Spotify Podcast Version https://podcasters.spotify.com/pod/sh... [00:25:46] Herbert Simon - Satisficing and Bounded Rationality https://plato.stanford.edu/entries/bo... [00:36:04] Eric Smith - Complexity and Early Life    • 5th ELSI Symposium: Eric Smith   book: [00:09:54] Karl Friston - Active Inference https://direct.mit.edu/books/oa-monog... [00:36:04] Richard Dawkins - The Selfish Gene https://amzn.to/3X73X8w [00:36:04] Carl Sagan - The Cosmos Knowing Itself https://amzn.to/3XhPruK paper: [00:36:04] DeepMind - Open-Ended Systems Paper https://arxiv.org/abs/2406.04268 --- LINKS: Full Transcript: https://app.rescript.info/share/3e498... Download PDF transcript: https://app.rescript.info/api/public/...