Kevin Keller·Jun 1An honest introduction to LQ.AIOpen-source AI for legal teams. Bring your own keys, run it where you want, own your data.
Kevin Keller·May 22When the verifier knows everything except how to winAn empirical look at where Rodionov-style verifier-driven coding agents fail — and where they keep failing across model scales.
Kevin Keller·Apr 22How We Named ADEThe Path from Idle Speculation to Artificial Doubt EngineA response icon1A response icon1
Kevin Keller·Apr 21The 70% Rule: Why Prompt Libraries Don’t Work for Real Knowledge WorkPrompting isn’t incantation. It’s delegation. And delegation to experts doesn’t scale the way people keep trying to make it scale.
Kevin Keller·Apr 17The Cognitive Sheffer Stroke: Generating Structure of Human Reasoning and Its Relationship to…Kevin Keller, Tucuxi, Inc.
Kevin Keller·Apr 10RSH Cosmological Corollaries — Complete Mathematical FrameworkKevin Keller Tucuxi Inc. April 2026
Kevin Keller·Apr 9Vijoy Pandey’s Internet of Cognition whitepaper is the clearest articulation I’ve seen of a problem…The proposed architecture has three layers: Cognition State Protocols (LSTP / CSTP / SSTP) as the wire format at three fidelity tiers, a…
Kevin Keller·Apr 8The Harness Is the Floor, Not the CeilingWhy the Great Convergence in AI Agents Leaves the Most Important Layer UnbuiltA response icon1A response icon1
Kevin Keller·Mar 19We Tested the Evolutionary Science of Project Hail Mary in a Cognitive Simulation.Our AI agents accidentally recreated Idiocracy before they recreated Project Hail Mary. That turned out to be the most important finding of…
Kevin Keller·Mar 5Your NPCs Are Lying to You. Ours Are Lying to Themselves.How we’re building game characters that actually think — and what happens when they do