Joon Hee KimEssays on how AI changes what it means to know, build, and decide.https://joonhee.kim/Shape of thoughthttps://joonhee.kim/essays/2026/06/shape-of-thought/https://joonhee.kim/essays/2026/06/shape-of-thought/The scarce thing in discovery is not a stronger generator or a sharper judge, but the structure that brings the right move within reach. The shape of the library in the head.Mon, 01 Jun 2026 00:00:00 GMTNone of them chose the gamehttps://joonhee.kim/essays/2026/05/none-of-them-chose-the-game/https://joonhee.kim/essays/2026/05/none-of-them-chose-the-game/Machines climb the measures we hand them. None has set its own, and maybe we have not either.Sun, 31 May 2026 00:00:00 GMTThe layer that remembers peoplehttps://joonhee.kim/essays/2026/05/the-layer-that-remembers-people/https://joonhee.kim/essays/2026/05/the-layer-that-remembers-people/The living part of taste is per-person and sits outside the weights. The densest map of it already exists, and we threw it away by reading it as a score.Sun, 31 May 2026 00:00:00 GMTCognition went publichttps://joonhee.kim/essays/2026/05/cognition-went-public/https://joonhee.kim/essays/2026/05/cognition-went-public/Reflexivity has moved one level deeper, and most participants have not noticed.Sun, 10 May 2026 00:00:00 GMTThe coupled recursionhttps://joonhee.kim/essays/2026/05/the-coupled-recursion/https://joonhee.kim/essays/2026/05/the-coupled-recursion/The corpus does not just accumulate. It shapes the cognition that produces the next corpus.Sun, 10 May 2026 00:00:00 GMTThe May 2026 Joint Ventures Between AI Labs and Private Equityhttps://joonhee.kim/essays/2026/05/the-may-2026-ai-labprivate-equity/https://joonhee.kim/essays/2026/05/the-may-2026-ai-labprivate-equity/Main essay: The Substrate InversionFri, 08 May 2026 00:00:00 GMTThe Substrate Inversionhttps://joonhee.kim/essays/2026/05/the-substrate-inversion/https://joonhee.kim/essays/2026/05/the-substrate-inversion/What the labs are actually buyingFri, 08 May 2026 00:00:00 GMTWhy code compiles and law doesn'thttps://joonhee.kim/essays/2026/04/why-code-compiles-and-law-doesnt/https://joonhee.kim/essays/2026/04/why-code-compiles-and-law-doesnt/The next AI transition is gated on who builds the verifier, not who trains the biggest model.Mon, 20 Apr 2026 00:00:00 GMTBuilding AI Producthttps://joonhee.kim/essays/2026/03/building-ai-product/https://joonhee.kim/essays/2026/03/building-ai-product/I Spent Two Days Speccing a Product I’ve Never Used. Here’s What I Should Have Done Instead.Fri, 20 Mar 2026 00:00:00 GMTThe Architecture of Discoveryhttps://joonhee.kim/essays/2026/03/the-architecture-of-discovery/https://joonhee.kim/essays/2026/03/the-architecture-of-discovery/What AlphaGo revealed about thinkingTue, 10 Mar 2026 00:00:00 GMTCompiled Awayhttps://joonhee.kim/essays/2026/03/compiled-away/https://joonhee.kim/essays/2026/03/compiled-away/A Map of What Gets Automated NextThu, 05 Mar 2026 00:00:00 GMTThe Taste Compilerhttps://joonhee.kim/essays/2026/03/the-taste-compiler/https://joonhee.kim/essays/2026/03/the-taste-compiler/How to build real-time taste inference for generative systemsSun, 01 Mar 2026 00:00:00 GMTThe Missing Taste Modelhttps://joonhee.kim/essays/2026/02/the-missing-taste-model/https://joonhee.kim/essays/2026/02/the-missing-taste-model/Create with AI needs a compass, not a generatorFri, 27 Feb 2026 00:00:00 GMT(note to self) From Punch Cards to Python, From Python to Agentshttps://joonhee.kim/essays/2026/02/note-to-self-from-punch-cards-to/https://joonhee.kim/essays/2026/02/note-to-self-from-punch-cards-to/How the agentic shift will redefine the best builderThu, 26 Feb 2026 00:00:00 GMTA Mesh Is Not a Doorhttps://joonhee.kim/essays/2026/02/playable-worlds/https://joonhee.kim/essays/2026/02/playable-worlds/Why world models will be won in game engines, then move into Physical AISun, 22 Feb 2026 00:00:00 GMTPixels → Protocolshttps://joonhee.kim/essays/2026/02/pixels-protocols/https://joonhee.kim/essays/2026/02/pixels-protocols/A2A turns SaaS into an execution layer. Operability becomes distribution.Fri, 20 Feb 2026 00:00:00 GMTLanguage-Conditioned Knowledge Inequalityhttps://joonhee.kim/essays/2026/02/language-conditioned-knowledge-inequality/https://joonhee.kim/essays/2026/02/language-conditioned-knowledge-inequality/Measuring quality-adjusted knowledge diffusion across language communities, and diagnosing language-dependent performance gaps in agentic knowledge work.Wed, 18 Feb 2026 00:00:00 GMTWhat Enchantment Means When You Don’t Believehttps://joonhee.kim/essays/2026/02/enchantment-as-a-model-update/https://joonhee.kim/essays/2026/02/enchantment-as-a-model-update/A rationalist’s attempt to name awe without overreach.Tue, 17 Feb 2026 00:00:00 GMTThe Inference Lead in the LLM Erahttps://joonhee.kim/essays/2026/02/the-inference-lead-in-the-llm-era/https://joonhee.kim/essays/2026/02/the-inference-lead-in-the-llm-era/What data science becomes after LLMsTue, 17 Feb 2026 00:00:00 GMTThe New Information Supply Chainhttps://joonhee.kim/essays/2026/02/the-new-information-supply-chain/https://joonhee.kim/essays/2026/02/the-new-information-supply-chain/When the cost of knowledge goes to 0, humans become the bottleneck and value moves downstream.Fri, 13 Feb 2026 00:00:00 GMTBy Logic We Prove; By Intuition We Discoverhttps://joonhee.kim/essays/2026/02/by-logic-we-prove-by-intuition-we/https://joonhee.kim/essays/2026/02/by-logic-we-prove-by-intuition-we/Why today’s LLMs are great proof engines, and why the next 2-3 years belong to “silent” discovery architecturesSun, 08 Feb 2026 00:00:00 GMTThe Evolution of Artificial Cognitionhttps://joonhee.kim/essays/2026/01/the-evolution-of-artificial-cognition/https://joonhee.kim/essays/2026/01/the-evolution-of-artificial-cognition/Five Layers of AbstractionFri, 30 Jan 2026 00:00:00 GMT