Writings

Notes on AI, building software, and the things I've learned along the way.

A collection of notes on AI, software systems, product thinking, and lessons picked up along the way.

Apr 21, 2026running 4× Claude Code Max still isn't enough — here's what actually helps

A power user's read on caveman, context compression, and every "save tokens" plugin your feed is pushing. Most of them optimize the cheapest slice of the bill.

Apr 21, 2026reverse-engineering Warp's cli-agent notification protocol

Warp's agent sidebar runs on one OSC 777 escape sequence and a six-field JSON envelope. here's the full protocol, reverse-engineered from the three official adapter repos.

Apr 16, 2026using remote SSH with Codex app

Codex app now has alpha SSH remote support, but during rollout you may still need the hidden feature flag, the beta appcast build, or a shell-init fix.

Apr 15, 2026Tip: enable 1M context in Codex with one command

Run one command, let the script create a backup of ~/.codex/config.toml, patch the 1M settings if needed, then restart Codex.

Apr 14, 2026I turned Pokémon Red into a shared multiplayer arcade — here's how

Starting from a Game Boy ROM, through rgbds assembly, Emscripten, WebAssembly, and a Redis-backed turn system, I built a live arcade where everyone who visits plays the same game.

Apr 13, 2026mac-to-mac file system: NFS is faster than native SMB for dev env

For tiny-file dev trees on macOS, tuned NFSv3 beats native SMB — if you pick the right client profile for a wired desktop vs a laptop that roams.

Apr 13, 2026using claude code's new native ssh remote on a mac mini / darwin

Claude Code's SSH remote rejects darwin hosts, but a lightweight OrbStack Linux VM on a Mac Mini turns the whole thing into a workable remote dev setup.

Apr 13, 2026auto-approve Claude Code plan mode via PermissionRequest

Claude Code's plan approval dialog can be auto-approved by hooking PermissionRequest correctly and matching ExitPlanMode instead of Stop.