Programming Language Fungibility in 2026: Mitchell Hashimoto on Rewriting Bun in Rust

Mitchell Hashimoto, the co-founder of HashiCorp and a legendary figure in infrastructure software, has sparked a new debate in the engineering world. In a recent interview with The Pragmatic Engineer, Hashimoto outlined a radical shift in his development workflow, driven by AI agents. But it was a subsequent social media post that truly caught fire: his assertion that programming language fungibility is now a defining characteristic of modern software engineering.

According to Hashimoto, the days of being locked into a single language for the life of a project are over. He pointed to the Bun runtime—originally written in Zig but recently rewritten in Rust—as a prime example. “Bun has shown they can be in probably any language they want in roughly a week or two,” Hashimoto wrote on X (formerly Twitter). “Rust is expendable. It's useful until it's not, then it can be thrown out.”

What Is Programming Language Fungibility?

Programming language fungibility refers to the ability to swap one programming language for another with minimal cost and disruption. Hashimoto argues that AI agents have made this concept a reality, eliminating the traditional lock-in that once defined software projects. In 2026, this shift is accelerating, as large language models (LLMs) can now translate entire codebases in days.

Why AI Agents Change Everything

Hashimoto’s argument rests on the accelerating capability of LLMs and AI coding agents. In his interview with The Pragmatic Engineer, he described how he now uses AI to generate, refactor, and even translate entire codebases. This, he claims, has fundamentally changed the cost-benefit analysis of choosing a programming language.

“Programming languages used to be lock-in,” Hashimoto explained. “You picked a language, built your team around it, and switching was a multi-year migration. Now, with the right AI tools, a rewrite that would have taken months can be done in days.” This perspective challenges the traditional wisdom that language choice is a long-term strategic decision. Instead, Hashimoto suggests that engineers should view languages as interchangeable tools, optimized for the current problem rather than the future.

The Bun Rewrite Case Study

The Bun example is particularly instructive. Bun is a fast, all-in-one JavaScript runtime and toolkit. Its initial implementation in Zig was a bold choice, but the team later migrated to Rust. Hashimoto’s point is not that Zig or Rust are bad languages, but that the programming language fungibility means the switch was a tactical, not existential, move. The Bun team could pivot because the underlying logic was preserved, and AI agents handled much of the translation grunt work.

Implications for Engineering Teams and Tooling

Hashimoto’s views carry weight because of his track record. As the creator of Vagrant, Terraform, and Packer, he helped define the modern DevOps landscape. His current perspective, shared on Mastodon and in the newsletter, suggests a future where engineering teams prioritize adaptability over specialization.

This has significant implications. If languages become fungible, the value of deep, narrow expertise may diminish. Instead, engineers might focus on architecture, system design, and prompt engineering—skills that transcend any single syntax. Companies may also reconsider their hiring criteria, looking for “polyglot” engineers comfortable with rapid rewrites.

Potential Challenges and Criticisms

However, Hashimoto’s thesis is not without critics. Some argue that AI-assisted rewrites still introduce subtle bugs and that performance characteristics vary dramatically between languages. Others worry that treating languages as disposable could lead to technical debt. Yet Hashimoto remains bullish: “The interesting side is how fungible programming languages are nowadays. It’s a new superpower for teams that embrace it.”

Future of Language Lock-In

In the end, Hashimoto’s message is clear: the era of language lock-in is ending. Whether or not the industry fully adopts this view, his commentary has already forced a re-examination of long-held assumptions. As AI continues to reshape software engineering, the programming language fungibility debate will likely intensify, and Hashimoto will remain at its center.

Frequently Asked Questions

What is programming language fungibility?

It is the ability to swap programming languages with minimal cost, enabled by AI agents that can translate and refactor code quickly.

How does the Bun rewrite relate to fungibility?

Bun’s rewrite from Zig to Rust shows how AI can make language migration fast and practical, reducing lock-in.

Is language lock-in really obsolete?

According to Hashimoto, AI agents make it obsolete for many projects, though some performance or safety-critical systems may still require specialized languages.