Perfect is the enemy of good
Chasing a bug-free Forge with Claude – 209 issues, a fuzzer, two AI code audits, and learning when to stop.
Posts
Solvent Streets
Building a tool with Claude to measure my city’s pavement and forecast its maintenance cost — a look at the fiscal toll of car-dependent design.
Forge end-to-end testing
Building out Forge’s end-to-end test suite with Claude, beads, and a sandboxed feedback loop.
From Bash Loops to a Ralph FSM
Evolving a Ralph-style Claude loop from a one-shot bash script into a generic Go tool driven by a finite state machine and reusable Go-CLI design scaffolding.
Towards a Rust Symbolic Execution Engine
After months of iteration with Claude, a Rust port of Angr’s symbolic execution engine that matches or beats Python on most CTF benchmarks – up to 12.7x on the largest.
Hand tools versus Power tools
What barrel-making teaches us about AI and the future of software development.
Can AI Test All Code Like SQLite?
SQLite has 590x more test code than library code. Can AI help us achieve that level of testing for all software?
Forge, meet Claude, part 4
Final progress on Forge bug fixes and thoughts on the challenge of upstreaming AI-assisted contributions.
Forge, meet Claude, part 3
Forging ahead with new features and code cleanups, including a keybinding cheatsheet overlay and quality-of-life improvements.
Forge, meet Claude, part 2
Continuing to tackle the Forge backlog with Claude: 43 bug fixes, 19 new features, and 728 unit tests.