The Zig–Bun Beef Got Personal
Andrew Kelley’s response to Bun’s Rust rewrite mixed technical criticism with personal attacks, private grievances and revealing post-publication edits.
On this page
The Zig–Bun story stopped being a language debate almost immediately.
On 8 July 2026, Bun creator Jarred Sumner published an account of rewriting Bun in Rust. The next day, Zig creator Andrew Kelley answered with “My Thoughts on the Bun Rust Rewrite”.
Sumner made a technical case for Rust and AI-assisted development. Kelley mixed technical criticism with years of resentment, allegations about Bun’s workplace and direct attacks on Sumner.
Bun’s post contained an implied criticism#
Sumner credits Zig with making Bun possible, but lists serious use-after-free crashes, double frees and leaks in the Zig implementation. His argument is that Rust can push more lifetime and cleanup checking into the compiler.
The provocative part was the method. Sumner says roughly 50 Claude Code workflows spent 11 days porting 535,496 lines, fixing errors and reviewing generated code. Separate agents were instructed to assume the implementation was wrong.
Bun did not call Zig a bad language. Even so, “we rewrote it in Rust and improved stability” inevitably sounded like a judgement on the old code.
Andrew Kelley made it personal#
Kelley opened with Sumner’s early “beginner energy”, describing him as ambitious but under-equipped and producing mediocre engineering outcomes while learning.
He then attacked Sumner’s background and motives. Kelley linked the Thiel Fellowship to an uncritical Silicon Valley mindset, framed venture funding as a race towards an exit and portrayed Bun’s priorities as incompatible with Zig’s long-term values.
The response became more personal when Kelley repeated negative accounts from former Oven employees and candidates. He called Sumner a “stinky manager” and alleged poor communication, unrealistic expectations, low empathy and inexperience. These were claims Kelley attributed to private conversations, not independently established findings.
He also attacked Sumner’s work directly. Kelley described Bun as hacks and technical debt, said Sumner produced “slop” before LLMs, called the company dysfunctional and said the Zig Software Foundation viewed Bun as a net liability.
The harshness came from the cumulative argument: Bun’s success was presented as the result of reckless engineering, bad management and venture-backed marketing rather than disciplined technical work.
The business relationship fuelled it#
Bun had donated about $60,000 per year to the Zig Software Foundation. Kelley says that after Anthropic acquired Bun, the donation stopped and Bun’s team neither cancelled nor attended a regular meeting.
He presents the Rust rewrite as a relief. ZSF no longer had to deal with Bun being treated as Zig’s flagship user while privately regarding its code as an example of how not to use the language.
That history explains the intensity, but it also makes the post read less like a technical rebuttal and more like a public release of private grievances.
There were technical criticisms underneath#
Kelley questioned why Bun’s test suite could supposedly validate a million-line Rust port but had not caught the Zig bugs. He disputed how Bun represented its fuzzing, argued that some performance and binary-size gains came from overdue engineering rather than Rust itself, said Zig maintainers had warned about excessive comptime, and noted the absence of post-rewrite compilation-time figures.
Those points do not disprove Bun’s case, but they challenge the cleaner marketing story that changing language produced every improvement.
The article changed after the backlash#
Kelley’s public Git history shows the response evolving in stages.
The original conclusion claimed he had no personal criticism of Sumner, despite comments about his competence, wealth, management and motives. It framed them as business opponents playing the same game and ended with a shrug.
Later that day, a commit literally titled “double down” added a longer insistence that the post was not personal. The same edit quietly removed a claim that Bun’s fuzzing statement was an outright fabrication, replacing it with narrower wording.
The following day, Kelley rewrote the conclusion after what he called self-reflection. He acknowledged resentment, accepted that readers experienced the post as a personal attack and apologised to Zig users who worried they might be targeted if they stopped using Zig.
The apology did not retract the article’s body. The management allegations, “slop” attack, liability claim and technical accusations remained. The new ending said Kelley still blamed Sumner for making Bun an embarrassment to Zig and stood by his leadership criticism.
The change was therefore not “I withdraw what I said.” It was closer to: “I now understand why this was received as personal.”
T3’s long-form breakdown#
Theo Browne covered the dispute in “We need to talk about the Bun Rust rewrite”. It is useful further viewing because the story makes more sense as a relationship breakdown than a simple Rust-versus-Zig argument.
Where it stands#
Bun’s Rust rewrite still has to prove itself through stable releases and long-term maintenance. Kelley softened his conclusion but left his central allegations intact. There has been no public reconciliation.
Rust and Claude gave the dispute its trigger. The actual beef had been building for years.