From Entropy to AI: The Hidden Mathematics of the Modern World

From Entropy to AI: The Hidden Mathematics of the Modern World What connects a steam engine, a thermostat, a telegraph wire, a chess game, the internet, recommendation algorithms, and ChatGPT? In this episode, we trace the genealogy of systems mathematics,: the hidden mathematical language behind uncertainty, entropy, information, feedback, computation, game theory, networks, data, and artificial intelligence. We begin with Boltzmann’s entropy and the problem of disorder, move through Shannon’s theory of communication, Wiener’s cybernetics, Turing’s computation, von Neumann and Nash’s strategic systems, Euler’s graph theory, Markov chains, causal inference, neural networks, embeddings, transformers, and finally the philosophical question at the heart of modern AI: Is intelligence just prediction at scale — or is something still missing? This is not just a history of AI. It is the story of how mathematics learned to describe chaos, communication, control, strategy, networks, learning, and perhaps even thought itself. 0:00 — The hidden mathematical DNA of modern life 1:14 — What is systems mathematics? 2:43 — Why AI begins with disorder 3:01 — Entropy, steam engines, and the arrow of time 3:39 — Boltzmann and the mathematics of microscopic chaos 5:22 — The playing-card thought experiment 6:12 — Entropy moves beyond closed physical systems 6:45 — Prigogine, chaos, and far-from-equilibrium systems 7:47 — Jaynes and entropy as a measure of ignorance 10:18 — Information as the reduction of uncertainty 10:48 — Claude Shannon and the birth of information theory 11:50 — Shannon’s model of communication 13:05 — Shannon entropy and the uncertainty of a source 14:03 — Compression, noise, and channel capacity 14:45 — Why Shannon separated information from meaning 15:35 — The noisy phone-call thought experiment 17:22 — From messages to actions 18:00 — Norbert Wiener and cybernetics 19:19 — Feedback loops and error signals 19:36 — The thermostat as a cybernetic machine 20:46 — Feedback in bodies, machines, and markets 22:06 — The Macy Conferences and universal systems theory 22:31 — The danger of treating society like a machine 24:28 — Computation: formalising thought itself 25:02 — Boole, logic, Church, and Turing 25:33 — The Turing machine explained 27:15 — Von Neumann and the architecture of modern computers 27:44 — Chess as a model of computation 28:43 — Syntax, semantics, and the problem of understanding 30:28 — When the environment fights back 30:50 — Game theory and strategic systems 31:56 — Penalty kicks and mixed strategies 33:14 — Nash equilibrium explained 34:01 — Game theory, economics, war, and evolution 34:27 — The limits of rational-agent models 35:49 — From two-player games to massive networks 36:28 — Euler and the Seven Bridges of Königsberg 37:44 — Nodes, edges, and the birth of graph theory 38:23 — Emergence and complex systems 38:44 — Traffic jams as network behaviour 39:39 — Small-world networks and six degrees of separation 40:15 — Scale-free networks and the “rich get richer” 41:36 — PageRank and the mathematics of the web 42:21 — From networks to probabilistic learning systems 43:01 — Markov chains and memoryless prediction 43:27 — A simple Markov weather model 44:26 — From statistics to prediction 44:46 — Judea Pearl and causal inference 45:43 — Recommendation algorithms as feedback systems 46:50 — Machine learning and artificial intelligence 47:08 — McCulloch, Pitts, and neural logic 48:56 — Rosenblatt, Hinton, Bengio, LeCun, and learning networks 49:15 — Parameters, error signals, and backpropagation 50:26 — Transformers and attention 51:35 — Embeddings and high-dimensional meaning 52:07 — King, queen, and vector space analogy 53:02 — ChatGPT as systems mathematics made interactive 53:59 — Is AI actually thinking? 54:58 — The grand synthesis: Boltzmann to Shannon to Turing to AI 55:40 — What is lost when the world becomes data? 56:03 — Is AI intelligence, or a mathematical mirror?