Guides
Longer pieces on choosing between tools, written for people who have to pick rather than people collecting options.
Terminal coding agents, and how to actually choose one
Nine tools, three families, and the one question that decides between them. Written for people who have to pick, not people collecting options.
Updated 2026-09-04Pi vs opencode — a product and a toolkit, wearing the same clothes
Both are MIT-licensed TypeScript terminal agents you point at your own model. One ships plan mode and permissions; the other deliberately ships neither. That single choice decides which one you want.
Updated 2026-09-08herdr vs tmux and Zellij — the terminal problem that starts at your third agent
A multiplexer keeps terminals alive. It cannot tell you which of them is waiting for an answer. That single gap is the entire case for herdr, and it is both narrower and more interesting than the pitch.
Updated 2026-09-10Muse-Glimmer 30B vs Qwen3.8 27B — both "local agent models", one distilled, one native
Two open-weight, Apache 2.0, ~30B multimodal agentic models released four days apart in August 2026. One compresses a bigger model's capability into a small one; the other grew it itself — and each vendor's benchmark table quotes the other's numbers. Here is how much to trust those numbers.
Updated 2026-09-10