Caleb Fenton spent last month writing 20,000 lines of code in a language he doesn’t know. He used AI agents to write it, and his verdict: it was easier than using the language he’d relied on for a decade.
That’s a sentence worth pondering.
Many developers have gravitated toward Python because it is forgiving. You can be loose, approximate and casual. Python mostly lets things slide and figures it out. That philosophy made sense when humans were doing the coding. Humans get frustrated. Humans lose patience. Humans quit and find a different approach when a tool keeps telling them they’re wrong.
Rust was built on the opposite philosophy. Where Python shrugs and lets you continue, Rust stops you dead. Its compiler checks your work with a rigour that experienced developers often describe as maddening. It enforces strict rules around memory and logic through a mechanism called the borrow checker, a system that tracks exactly how data is created, used and discarded. Get any of it wrong and the code doesn’t run. Not “runs with bugs.” Doesn’t run at all.
That strictness comes with a steep learning curve. Many developers avoid Rust unless they absolutely need it. Life is short.
Fortunately, AI agents don’t have feelings about being rejected.
An agent will rewrite code a thousand times without complaint. The very quality that makes Rust exhausting for humans – the relentless, unforgiving feedback loop – is exactly what makes it powerful when an AI is doing the work. The agent keeps going until the code passes. And because Rust won’t accept sloppy code, what eventually passes is genuinely solid.
Python’s looseness, which was a kindness to human developers, becomes a liability with AI. The agent can be imprecise. Python doesn’t flag it. The bugs slip through to the people using the software.
Python isn’t going anywhere, though. Decades of community contributions have produced an ecosystem Rust can’t match. The entire machine learning and data science world runs on Python libraries that took years to build and refine. For research, experimentation and rapid prototyping, Python remains the natural choice. Fenton acknowledges this. The scratchpad, he says, isn’t going anywhere.
But production software – the kind that runs while you sleep and breaks when nobody’s watching – is a different matter.
Fenton’s central argument is straightforward: if AI is doing the hard work of writing code, you might as well choose the language that enforces quality, even if humans find it painful. The pain was always the cost of admission. It just wasn’t worth paying when humans were the ones writing the code.
Tools like Claude Code are language agnostic. No preference baked in, no allegiance to Python or Rust. They will work with whatever language you choose and iterate without frustration until the job is done.
In an increasingly software-driven world, code that works correctly matters. A lot. The fact that AI agents can now absorb the frustration that kept developers away from stricter, more rigorous languages is, quietly, a significant shift.
I’ll be giving Rust a go in the near future. The compiler can reject my coding agent as many times as it likes.
Sources:
- Python is Dead – Caleb Fenton, Substack
- Python – python.org
- Rust – rust-lang.org
- Python on Wikipedia – en.wikipedia.org
- Rust on Wikipedia – en.wikipedia.org