Skip to content
context AI agents

curl author Daniel Stenberg compares AI rewrites into Rust with the project's real-world development

only one source so far

Daniel Stenberg, author of curl, compared AI rewrites of curl from C to Rust (185 thousand → 340 thousand lines) with the project's actual development, which over the same period added 300+ commits and 200+ bugfixes including security fixes that the rewrites do not contain.

Daniel Stenberg, author of the curl project, compared the results of AI agents that converted curl's source code from C to Rust with the actual development of the project taking place over the same period. According to Stenberg, over this time the agents converted 185 000 lines of C code into 340 000 lines of Rust code — meaning the resulting code is significantly more voluminous than the original.

Over the same time span, roughly three weeks, the curl team merged over 300 commits and made more than 200 bug fixes to the project itself, none of which are included in the Rust rewrites. According to Stenberg, several security vulnerabilities were also fixed during this time. Stenberg states that over the following month the project will likely add roughly 200 more fixes and possibly additional security patches.

From this, Stenberg raises the question of whether AI-generated rewrites can keep pace with the speed of live, actively maintained development. This is his personal comparison and opinion, not an independently verified measurement of the quality or security of the resulting Rust code.

What changed

Why it matters

The comparison illustrates a concrete risk for projects considering AI-generated ports or rewrites of large codebases: the resulting code may be more voluminous, but without keeping pace with the bugfixes and security fixes made in the actively maintained original. This is relevant input for deciding whether and how to maintain or deploy such AI ports in production.

Two audiences, two different impacts

What this means

01

For individuals

For developers, it's instructive that a larger volume of generated code does not mean higher quality or security — the AI rewrite of curl into Rust does not match the pace of bug fixes maintained in the live, actively maintained original project.

What to do When evaluating AI tools for porting or rewriting code, compare the result not only by functionality but also by its alignment with the current security fixes of the original.
More practical updates →
02

For a business

For companies considering migrating or porting code using AI agents, Stenberg's described risk is relevant: an AI rewrite may grow in volume of code without keeping pace with the security fixes and bugfixes of the actively developed original, which increases the risk of deploying an outdated or vulnerable version.

Development
What to decide Before any potential deployment of an AI-generated code port (e.g., a rewrite into another language), verify whether it contains security fixes and bugfixes comparable to the actively maintained original version.
More business impacts →
C++ curl generative AI coding Rust security

Check the original

Event sources

only one source so far · 1 publisher, 1 independent. We count feeds from the same owner only once.

1
Simon Willison — AI tag (leading independent LLM commentator) community signal · first detected Quoting Daniel Stenberg