https://lukeed.com/blog/ Notes, rants, and deep-dives on tooling, design, and en-us Wed, 21 Oct 2026 00:00:00 GMT https://lukeed.com/blog/bongo-bench/ https://lukeed.com/blog/bongo-bench/ Wed, 21 Oct 2026 00:00:00 GMT How a 10,000-page bongo build went from 6.9 seconds to 2.2, and a rebuild after an edit to 239 ms, with every step measured. https://lukeed.com/blog/bongo/ https://lukeed.com/blog/bongo/ Wed, 14 Oct 2026 00:00:00 GMT bongo started in 2020, was rewritten twice, and now builds this site. It is public today. https://lukeed.com/blog/js-performance/ https://lukeed.com/blog/js-performance/ Thu, 08 Oct 2026 00:00:00 GMT Six habits from ten years of small, fast libraries, each one shown in real code and measured on two engines. https://lukeed.com/blog/maintaining-alone/ https://lukeed.com/blog/maintaining-alone/ Wed, 30 Sep 2026 00:00:00 GMT How one person keeps a few hundred packages alive. What I triage, what I ignore, what I automate, and what agents changed. https://lukeed.com/blog/empathic/ https://lukeed.com/blog/empathic/ Thu, 24 Sep 2026 00:00:00 GMT Eleven popular packages walk up a directory tree, and together they install twenty times. empathic does the same work in one package with no dependencies. https://lukeed.com/blog/lane/ https://lukeed.com/blog/lane/ Mon, 14 Sep 2026 00:00:00 GMT git worktree gives you a clean checkout and a cold cache. lane clones the ignored files by reference, so a new branch builds in seconds. https://lukeed.com/blog/four-sparks/ https://lukeed.com/blog/four-sparks/ Thu, 27 Aug 2026 00:00:00 GMT Field notes from wiring four DGX Sparks into one inference cluster. The fabric works now. Here is everything that cost me a day. https://lukeed.com/blog/esm-transition/ https://lukeed.com/blog/esm-transition/ Tue, 11 Aug 2026 00:00:00 GMT Eight years of shipping packages across the CommonJS and ES Module line, the mistakes authors still make, and whether it was worth it. https://lukeed.com/blog/agent-apis/ https://lukeed.com/blog/agent-apis/ Tue, 21 Jul 2026 00:00:00 GMT Explicit names, few options, no magic, and errors that say what to do next. An API that is good for agents turns out to be good for people too. https://lukeed.com/blog/readme-driven/ https://lukeed.com/blog/readme-driven/ Tue, 30 Jun 2026 00:00:00 GMT Usage examples are where developer experience gets decided, and now the agents read them too. https://lukeed.com/blog/tools-that-lost/ https://lukeed.com/blog/tools-that-lost/ Tue, 09 Jun 2026 00:00:00 GMT Taskr, pwa-cli, and freshie were each my answer to a big question in JavaScript tooling, and each one lost. Here is what they taught me. https://lukeed.com/blog/comptime/ https://lukeed.com/blog/comptime/ Thu, 07 May 2026 00:00:00 GMT A Zig-inspired primitive that runs a function during your build and inlines the result. One import, two plugins, and no new syntax. https://lukeed.com/blog/invisible-packages/ https://lukeed.com/blog/invisible-packages/ Tue, 21 Apr 2026 00:00:00 GMT Some of my smallest packages are installed billions of times a year, almost always by something else. Here is what that asks of me. https://lukeed.com/blog/benchmarking/ https://lukeed.com/blog/benchmarking/ Tue, 31 Mar 2026 00:00:00 GMT A benchmark is a claim, and most of the work is making sure the claim is true before you make it. https://lukeed.com/blog/uvu/ https://lukeed.com/blog/uvu/ Tue, 10 Mar 2026 00:00:00 GMT Most of a test run is spent before your first assertion, and uvu was an experiment in skipping that part. https://lukeed.com/blog/small-on-purpose/ https://lukeed.com/blog/small-on-purpose/ Fri, 20 Feb 2026 00:00:00 GMT Ten years of writing 200-byte libraries, and why I still think the size limit is a design tool, not a party trick. https://lukeed.com/blog/worktop/ https://lukeed.com/blog/worktop/ Tue, 27 Jan 2026 00:00:00 GMT I wrote a router for Cloudflare Workers when the platform had almost no tools. Some of its ideas became normal. Some stayed niche. https://lukeed.com/blog/polka/ https://lukeed.com/blog/polka/ Tue, 13 Jan 2026 00:00:00 GMT An Express alternative in 90 lines got far more attention than it earned. Here is what it proved, what it did not, and what grew out of it. https://lukeed.com/blog/another-router/ https://lukeed.com/blog/another-router/ Fri, 06 Dec 2024 00:00:00 GMT I went looking at other JavaScript routers and found my own code inside most of them. https://lukeed.com/blog/file-system-routers/ https://lukeed.com/blog/file-system-routers/ Fri, 22 Nov 2024 00:00:00 GMT Routes are data. Directories are a poor place to keep data.