We've raised $17M to build what comes after Git

Original Article ↗ Hacker News Discussion ↗

A startup led by a GitHub cofounder has raised $17M to build GitButler, a Rust-based layer on top of Git that promises “what comes after Git,” especially for AI- and agent-driven development workflows. Commenters are sharply divided: some see real value in features like virtual branches, stacked changes, and tighter integration of issues and PRs, while many argue Git’s ubiquity, simplicity of data model, and existing tools (or alternatives like Jujutsu, Pijul, and Perforce-style workflows) make displacement unlikely. There is also broad skepticism about VC-funded attempts to wrap or “improve” core infrastructure, with concerns about lock-in, monetization pressure, and whether this solves meaningful problems compared to Git plus good interfaces and LLM assistance.

What GitButler Is (and Isn’t)

  • Many readers note the product is a UX layer on top of git, not a new VCS, despite “what comes after git” branding.
  • It uses git repos and GitHub; seen more as a nicer workflow/CLI/GUI (e.g., virtual branches, operations log) than a fundamental replacement.
  • Some features (operations log, multi-branch workflows) are compared to existing git capabilities like reflog and worktrees.

Alternatives and “Post‑Git” Options

  • Several point to existing contenders: Jujutsu (jj), Pijul, Sapling, etc., as more credible “after git” systems.
  • jj in particular gets strong praise: snapshot model, easy branching/switching, good fit for AI workflows, git‑compatible hosting via GitHub.
  • Others say they’re still happy with SVN or even CVS; some believe distributed VCS was never needed for most users.

AI/Agents and Version Control

  • One camp argues AI agents make traditional git workflows feel outdated; coordination of many parallel branches and agent plans is painful.
  • Another camp responds that LLMs already handle git very well (including merges and PRs), so adding another abstraction is unnecessary.
  • GitButler’s “virtual branches” are pitched as better for agents than git’s single‑checked‑out‑branch model; critics counter that worktrees already solve this.

Funding, VC, and Business Model

  • Strong skepticism about $17M for “better git,” especially given git’s origin story and community maintenance.
  • Many say the raise is primarily based on the GitHub cofounder’s track record (“pedigree”), not the product.
  • Concerns include: extracting value from a critical tool, lock‑in, eventual enshittification, and misallocation of capital versus more tangible problems.
  • Others defend it: dev tooling is big business, $17M is modest compared to AI bets, and competition in VCS/forges is healthy.

Git UX: Fine vs Broken

  • Split views: some insist “git is fine” or that standardization outweighs any better alternative; others describe git’s UX as mediocre, especially for rebases, merge conflicts, large binaries, and GitHub’s PR/diff UI.
  • There’s interest in richer models (patch‑based, environment‑aware, centralized with locking for binaries).

Security and Workflow Concerns

  • One user reports GitButler installing git hooks that block normal git commit, calling it “malicious.”
  • A project representative replies that the hook is necessary for their multi‑branch “megamerge” model and is removed on checkout; critics still dislike a tool hijacking core git behavior.