{"type":"rich","version":"1.0","provider_name":"Transistor","provider_url":"https://transistor.fm","author_name":"A Productive Conversation","title":"Why Logic Alone Can't Explain the Human Mind (with Tom Griffiths)","html":"","width":"100%","height":180,"duration":2314,"description":"Logic is beautiful and it's also almost never enough. That's the tension running through this entire conversation — the idea that we want our decisions to be purely logical because it's simple and concrete, but almost no real situation in a human life actually has the structure logic requires. I've believed for a long time that reason lives somewhere in the nuanced middle between logic and emotion, and I wanted to sit down with someone who's spent a career studying exactly that boundary.\nTom Griffiths is a Princeton professor and the author of the new book The Laws of Thought: The Quest for a Mathematical Theory of the Mind. If you know his earlier book, Algorithms to Live By, this one is the foundation underneath it — the deeper, denser history of how mathematics has tried, and often failed, to fully capture how minds work. We covered logic's limits, David Hume's riddle about induction, why AI systems fail in strange and \"jagged\" ways humans don't, and how a 200-year-old idea from George Boole ended up, several generations later, inside modern neural networks through his own great-great-grandson.\n\nSix Discussion Points\nLogic only works when you already have all the information you need and the conclusion is certain — which describes almost none of the decisions we actually make as humans.\nAmos Tversky's early work on similarity (a horse is more \"like\" a camel than a camel is \"like\" a horse) laid the groundwork for the heuristics-and-biases research he later did with Daniel Kahneman.\n\"Productivity,\" in the cognitive science sense, means a system's ability to generate something genuinely new — which is closer to creativity than to output.\nDavid Hume's riddle of induction shows we can't logically justify assuming the future will resemble the past — we simply have a habit of expecting it to.\nAI systems demonstrate \"jagged intelligence\": strong on one problem, mysteriously weak on the problem right next to it, because they arrive at solutions through a completely...","thumbnail_url":"https://img.transistorcdn.com/RaxQE_yNeOcP9CV60hOV3GBXJq5J7iHtixqMZ6k8ieU/rs:fill:0:0:1/w:400/h:400/q:60/mb:500000/aHR0cHM6Ly9pbWct/dXBsb2FkLXByb2R1/Y3Rpb24udHJhbnNp/c3Rvci5mbS84ODBi/MTA3MDFjYjQwMDVj/ZGQ2N2I1MjZiNjhh/YTlhMS5wbmc.webp","thumbnail_width":300,"thumbnail_height":300}