Blog

Who benchmarks the benchmark? 26.2% to 100% without touching the model. Six ways a benchmark measures itself instead of the agent. August 2026 What's in the box? Seventeen years of papers saying "assume oracle access," circled in red pen. Plus a 12-lesson course that opens the box. August 2026 Spatial languages Writing code in 2D: vertical operands, quantum uncomputation puzzles, and a 3-bit adder in two expressions. July 2026 Bring-Your-Own-Agent No backend, no install: a webpage that borrows your local coding agent over a localhost WebSocket. June 2026 On Planet Quantos, Programmers Flee The quantum rank-select primitive, described through an interactive story. May 2026 Entangled Spaghetti Code Quantum computing has its own class of bugs. b01t is a Python DSL that catches them at build time. April 2026 Game of Sway Quantum computing for the whole family! March 2026 Grammar models are back, baby! A framework for grammar generation, inference, and scoring with applications to MCTS and grammar synthesis. February 2026 Nano Governments Mathematical formulations of nanoscale governments. January 2026 Algebra Notes Diabolical charades of algebra. November 2021