Anthropic acquired Bun in December 2025; Claude Code’s visible product decline raises concern that Bun could follow the same enshittification path.
Key Takeaways
Anthropic’s stated rationale for the acquisition: Claude Code ships as a Bun executable, so Anthropic has direct incentive to keep Bun stable and fast.
Claude Code degraded visibly in April 2026: reduced default reasoning effort, a stale-session bug, and a prompt change hurt coding quality per Anthropic’s own postmortem.
The OpenClaw incident is the sharpest red flag: simply having OpenClaw in git history could trigger Claude Code refusals or extra billing, even in an empty repo calling claude -p "hi".
Author’s practical move: switch to pnpm for package management on existing projects, citing pnpm’s comparable monorepo and disk-usage story without the acquisition risk.
Bun’s bundled value (native TS, built-in bundler, test runner) has no single pnpm equivalent, making any migration involve adding Vite, vitest, and a build step back.
Hacker News Comment Review
Commenters split sharply: optimists note Bun now has sustainable funding where it previously had VC pressure; skeptics point to a pattern where acquired teams cash out and parent-company culture displaces product culture over time.
Technical critics argue Bun was already unstable pre-acquisition: memory issues, breaking changes in patch releases, and a known Linux ARG_MAX bug that silently hangs installs in large monorepos have persisted across versions.
A Bun team member pushed back directly, listing concrete near-term shipping work (smaller binaries, new CLI flags) and arguing development pace has increased since joining Anthropic.
Notable Comments
@Jarred: Bun team member counters the post with a concrete feature list shipping in the next release, calling the worry “confusing” given accelerated development pace post-acquisition.
@nulltrace: Documents a specific ARG_MAX bug in Bun’s security scanner that silently hangs on large monorepos on Linux, broken since at least 1.3.5 with no fix.
