When Toolchains Get Fast Enough, Rewrites Start Looking Rational

articlex-twittertoolingtypescriptnextjsbunlintingperformanceagent-loopsjoelclaw

faster TypeScript and lint feedback directly lowers the cost of joelclaw agent loop review cycles

Jared Palmer posted the current everything feels fast stack: Bun, Next.js with Turbopack, TypeScript Go, oxlint, oxlint-tsgolint, and Biome. Not one magic tool. The whole feedback loop is getting squeezed: install, dev server, type-aware lint, formatting, and build checks.

That’s why Joel’s reaction landed: “this makes me want to rebuild everything tbh.” Speed changes the appetite for architecture work. When every check takes a coffee break, refactors feel expensive. When checks are instant, rebuilding starts looking less like procrastination and more like something you can actually touch without hating your life.

Useful for joelclaw because the system already lives near this stack: Bun, TypeScript, Next.js, and Biome. The sharper move is treating toolchain speed as part of the runtime budget for agent loops: faster checks mean cheaper review cycles, tighter iteration, and less slow-feedback bullshit hiding in the walls.

Key Ideas

  • Jared Palmer is pointing at a full-stack speed pass, not a single dependency swap: Bun, Next.js, Turbopack, TypeScript Go, oxlint, oxlint-tsgolint, and Biome all attack feedback latency.
  • TypeScript Go and oxlint-tsgolint are the interesting edge because type-aware checks are usually where fast JavaScript projects go to die.
  • For joelclaw, toolchain speed is not polish. It is infrastructure for agent loops, reviewer steps, and local confidence.
  • The danger is classic rewrite energy: fast tools make rebuilding feel possible, but the win only matters if the new stack reduces the boring daily loop, not just the first-run demo.