Latent: present, but not yet visible. Conjecture: a guess you are willing to test. This lab is mostly the second, in pursuit of the first.https://latentconjecture.com/https://latentconjecture.com/projects/poc-lottery-randomness-auditor/https://latentconjecture.com/projects/poc-lottery-randomness-auditor/A reproducible statistical audit of published UK Lotto winning numbers, testing for detectable deviations from fairness era by era.Wed, 21 Oct 2026 00:00:00 GMTData Sciencestatisticsdata-sciencehypothesis-testingbayesianmarimohttps://latentconjecture.com/projects/semantic-context-lab/https://latentconjecture.com/projects/semantic-context-lab/What does a knowledge graph, and an ontology on top of it, add to an AI answering questions over company data? Three approaches, 300 questions, one synthetic company.Wed, 07 Oct 2026 00:00:00 GMTData & Graphsknowledge-graphragontologyneo4jllm-evaluationfastapireacthttps://latentconjecture.com/projects/treetop-benchmark/https://latentconjecture.com/projects/treetop-benchmark/One brief, one shot, one file. Frontier models each build an interactive 3D treetop village in a single HTML page, and a viewer puts them side by side.Thu, 01 Oct 2026 00:00:00 GMTLLM Evaluationbenchmarkthree.jswebglcode-generationhttps://latentconjecture.com/projects/tabula/https://latentconjecture.com/projects/tabula/An arena where two language models play backgammon against each other, explaining every move, with Glicko-2 ratings and per-move performance metrics.Mon, 28 Sep 2026 00:00:00 GMTLLM Evaluationbenchmarkgamesreactglicko-2agent-sdkhttps://latentconjecture.com/projects/slop-chips/https://latentconjecture.com/projects/slop-chips/An agent skill that finds the wording that marks text as AI-written, then rewrites it into clear, human English using a style guide drawn from real books.Tue, 08 Sep 2026 00:00:00 GMTAgent Skillsagent-skillswritingclaude-codecodexhttps://latentconjecture.com/projects/altman-warnings-vs-promises/https://latentconjecture.com/projects/altman-warnings-vs-promises/Four years of Sam Altman's public statements scored on two axes, how grand the capability claims got and how much airtime the risks got, with every statement on its real date.Thu, 03 Sep 2026 00:00:00 GMTData Sciencedata-sciencetext-analysisai-policyjupyterhttps://latentconjecture.com/projects/muster/https://latentconjecture.com/projects/muster/Structured agentic development as Agent Skills. Seven commands take a half-formed idea to a merged pull request, with sub-agents doing the building and a person approving every gate.Mon, 22 Jun 2026 00:00:00 GMTAgent Skillsagent-skillsagentssub-agentscode-reviewhttps://latentconjecture.com/projects/benchpress/https://latentconjecture.com/projects/benchpress/A hard, deterministic, judge-free benchmark. Models run a small machine for N steps in their heads and report the exact final state.Mon, 27 Apr 2026 00:00:00 GMTLLM Evaluationbenchmarkcellular-automatapythonleaderboardhttps://latentconjecture.com/projects/value-weighted-targeting/https://latentconjecture.com/projects/value-weighted-targeting/A working paper on how ranking players by treatment effect times lifetime value can steer retention incentives toward the players most at risk of harm, and which safeguards help.Tue, 17 Mar 2026 00:00:00 GMTStatisticscausal-inferenceuplift-modellingresponsible-gamblingsimulation