TLDRocket
Sign in

Mojo🔥 is now open source

Simon Willison’s Weblog Simon Willison

Mojo, the GPU-friendly language that copies Python's syntax, just went fully open source under Apache 2. It's been promising this since 2023, and it finally happened right after hitting version 1.0.

Based on reporting by Simon Willison’s Weblog, Simon Willison — read the original for the full story.

Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error

The Mojo team has finally made good on a pledge that's been sitting on their to-do list since May 2023. Last week they pushed out version 1.0 of the language, and today they followed that up by releasing the compiler and toolchain under an Apache 2 license — the kind of permissive terms that let anyone fork, modify, or build a business on top of the code without asking permission.

What's interesting is how much the project's ambitions have shifted along the way. Mojo originally billed itself as a superset of Python, meaning the goal was full compatibility so existing Python code could just run and help bootstrap the new ecosystem. That framing quietly changed around August 2025. The team now says Mojo may or may not ever become a complete Python superset, and they've made peace with that uncertainty.

Part of the reason they're comfortable dropping that promise is AI-assisted coding tools. The Mojo developers point to how well these tools already handle migrating Python code into Mojo, and they're betting that tooling and ecosystem maturity will keep smoothing that path over time. In other words, instead of building compatibility into the language itself, they're leaning on AI assistants to bridge the gap.

So Mojo today stands as its own language rather than a Python clone. It borrows Python's syntax because that's familiar and readable, but it's built with a different mission entirely: making GPU programming less painful. That's a narrower, more specific goal than universal Python compatibility, and it's arguably a more honest one — chasing perfect Python parity while also optimizing for GPU workloads is a tall order for any language to pull off cleanly.

My take — AI-written commentary, not fact-checked reporting

Dropping the Python-superset promise looks like a smart retreat rather than a failure — plenty of languages have died trying to be everything to everyone. Betting on AI tools to handle the migration instead of forcing language-level compatibility is a pragmatic shortcut, and it's the kind of move that only makes sense in 2025. Now that it's open source, the real test is whether GPU programmers actually adopt it or just admire it from a distance.

Read more about this at: Simon Willison’s Weblog

Related stories

The daily briefing

Every AI story that matters, in your inbox by 8am.

TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.