Regulated-domain AI, world models, and policy-driven agents. Built in the open.https://blog.hammer.ai/en-ushttps://blog.hammer.ai/blog/what-a-world-model-is-worth-to-a-swarm/https://blog.hammer.ai/blog/what-a-world-model-is-worth-to-a-swarm/You cannot get it right the first time. Nobody can. So we built a world you can rerun a thousand times, and used it to test which swarms find the best path rather than the one that sounds good. The answer took twenty experiments and is not the one we expected. Playable in your browser.Wed, 19 Aug 2026 00:00:00 GMTworld-modelsswarmdronesaisimulationdecisionsgame-theoryhttps://blog.hammer.ai/blog/the-sign-test/https://blog.hammer.ai/blog/the-sign-test/Banking ran our definition-change test for us, better documented than we could have managed, and the answer was not the one anybody expects. On one day in January 2020 a single accounting rule moved reserves up at three banks and down at a fourth, and up and down inside the same bank. That gives a coherence test costing one group-by, and it also breaks a category we had been treating as one thing.Thu, 30 Jul 2026 00:00:00 GMTaiworld-modelsregulated-aibankingsimulationhttps://blog.hammer.ai/blog/the-rules-are-half-the-physics/https://blog.hammer.ai/blog/the-rules-are-half-the-physics/Regulated industries break world models in four specific ways: colliding clocks, feedback loops, legal fences, and your own output becoming your next input. They also hand you the one thing world-model builders everywhere else lack, which is a supply of free, dated, published interventions to test against.Mon, 27 Jul 2026 00:00:00 GMTaiworld-modelsregulated-aisimulationhealthcarehttps://blog.hammer.ai/blog/03-the-llm-learned-to-stop-planning/https://blog.hammer.ai/blog/03-the-llm-learned-to-stop-planning/We let an LLM design competing species, run them headless in the living world, and refine them from the results. Over three rounds it revived an extinct lineage into the winner and abandoned its own planner. Going deeper on Part 1.Mon, 06 Jul 2026 00:00:00 GMTaiagentsgoapworld-modelssimulationhttps://blog.hammer.ai/blog/01-hello-world-models/https://blog.hammer.ai/blog/01-hello-world-models/A handful of procedural rules can grow a whole living world. First a planet from a seed, then life that adapts to it, procedurally, and with an LLM. This is how games can teach our agents.Mon, 29 Jun 2026 00:00:00 GMTaigamesprocgenworld-modelshttps://blog.hammer.ai/blog/start-here/https://blog.hammer.ai/blog/start-here/Why we build practice worlds for regulated AI, learned from games. The story behind the series, and what it explores: costs, tradeoffs, and worlds grown from the law itself.Mon, 22 Jun 2026 00:00:00 GMTaigamesregulated-aiworld-modelsintrohttps://blog.hammer.ai/blog/implementing-turboquant-in-llamacpp-cuda-scars-and-what-actually-ships/https://blog.hammer.ai/blog/implementing-turboquant-in-llamacpp-cuda-scars-and-what-actually-ships/Part 1 of 2. Why We Did This Hammer.ai runs a industrial research lab hyper focused on regulated domain document understand at extremely efficient margins. Private equity self funded companies like fMon, 06 Apr 2026 00:00:00 GMTcudaGPUllmquantization