Grit Rewrites Git in Rust Using Agents, Raises the 'Why' Question

Grit, a from-scratch Rust reimplementation of Git built with AI agents, passes over 99% of the Git test suite. The technical achievement is real. The comments are skeptical, not about the execution, but about the motivation. Multiple people asked the same thing: is it faster? More capable on large repos? Does it enable better workflows? The project's framing around memory safety and idiomatic Rust didn't answer those questions.

This connects to a recurring HN pattern where Rust rewrites of foundational tools get attention but struggle to articulate a value proposition beyond 'memory safe' and 'written in Rust.' The comment 'I'd rather it stay with the existing Git ecosystem' captures the sentiment: if it doesn't do something meaningfully better, the switching cost is pure cost.

The interesting subplot is that this was built with AI agents, which is the actual novel part. The question of whether agents can produce production-quality reimplementations of complex, battle-tested systems is genuinely unsettled, and this project is one of the more ambitious tests of that idea.