This is a practitioner's migration diary disguised as a breakup letter. The author — a developer who bet on Deno years ago as the cleaner, more modern JavaScript runtime — returns to Node.js in 2026 and discovers the gap has not just closed but inverted. Native TypeScript execution, modernized file system APIs, the death of require() — Node absorbed Deno's best ideas while Deno chased startup metrics. The technical meat is specific and useful. PNPM replaces NPM as the package manager of choice, with post-install scripts blocked by default and a minimumReleaseAge setting to avoid being the first sucker to install compromised packages. The author's line — 'The M in NPM stands for malware' — lands because the mitigations are real: trust policies, delayed updates, aliasing npm to pnpm. This is operational hygiene dressed as humor. The Node TypeScript story is telling. Node v26.10.0 runs TypeScript natively but refuses to strip types inside nodemodules — a philosophical rather than technical restriction, as the author notes, designed to prevent the ecosystem from becoming fully dependent on a Microsoft product. The workaround (bundling with tsdown, two extra dotfiles) is minor but illustrative: the JavaScript ecosystem still can't agree on where TypeScript ends and JavaScript begins. The migration benchmark is the piece's strongest evidence. Converting a static site generator from Deno to Node required replacing exactly two APIs (Deno's FS with node:fs, Deno.serve with Hono's node adapter) and yielded 15% faster builds — with the author conceding the codebase still uses idiomatic Deno patterns, meaning Node performance has headroom left. The Deno post-mortem is three paragraphs but could be one sentence: Deno Land Inc. stopped investing in the runtime and started building commodity cloud products nobody asked for. Broken ZSH integration, aggressive rate limiting on JSR, HTTP concurrency bugs — these aren't strategic failures, they're maintenance failures. The company stopped caring about the tool that made people care about the company. The broader signal here is convergence. Node didn't win by being Node — it won by becoming what Deno promised to be. ESM by default, native TypeScript, modern APIs. The lesson isn't 'big ecosystem wins' (though it does). It's that good ideas in competing runtimes get absorbed by the incumbent faster than the challenger can build an ecosystem around them. Deno's real contribution may be that it forced Node to modernize. The piece works because it's specific. Not 'Node is better now' but 'here are the exact API swaps, the exact build time delta, the exact PNPM config that makes this viable.' That specificity is what separates a useful migration report from a vibes post.
