Seems like every program's event loop will inevitably be replaced or augmented by a reasoning LLM. The remaining challenge is making it fast enough to match user expectations (~250ms) when executing on-device. I'm hopeful model distillation techniques will make this possible.

AI Agents (using MCP) are just loops because they are actually interpreters, and ultimately an interpreter is just a loop.

How do we make Fil-C successful? This post describes my thoughts about what has to happen next. First of all, don't worry about performance. Of the user-facing programs I've ported, none of them exhibits performance problems that make them less usable. You won't notice the perf

I'm at #pycon! It's off to a great start — I got this ahead-of-its-time AI Agent keychain from the Microsoft booth.

The full video of my Upper Bound 2025 talk about our research directions should be available at some point, but here are my slides: docs.google.com/presentation/d… And here are the notes I made while preparing, which are more extensive than what I had time to say:

Video from my PyCon Talk ("The Zen of Polymorphism") is up! It starts with basic type switching and ends with the catamorphism abstraction. I also explain how OOP organizes code along the wrong axis. I hope you enjoy! youtube.com/watch?v=hidy15…

I accidentally used the word "orthogonal" in conversation with a person outside of work today 😬 I probably need a vacation...

This is interesting — a compiler/runtime that makes almost all standard C/C++ code completely memory safe. The performance overhead is not negligible, but there are plenty of cases where “rewrite it in rust” still isn’t a practical solution.

Watching Carol Willing keynote presentation at PyBeach Great venue and I'm excited to be here! #pybeach2025

Video of my talk at the PyBeach conference in Santa Monica last weekend is now online. What a fun time! Great speakers, excellent venue, and a wonderful community of SoCal Python enthusiasts. youtube.com/watch?v=MWAfiy…

The Japanese translation of "Effective Python" 3rd edition goes on sale today — so excited to see this! oreilly.co.jp/books/97848144…

At the PyBay conference today where glyph is sharing about the realities of running open source projects.

Right now at PyBay Guido van Rossum is showing off a live demo of structured RAG. Enjoying some sketch comedy deep cuts, too.

At PyBay, Lisa Roach is now talking about free threaded Python (no GIL) vs. multiprocessing.

Amazing talk at PyBay right now by Noah Kantrowitz about how "hello world" in Python actually works at the lowest levels. Perhaps too deep (see pic)?

Last day to submit talk proposals for PyCascades 2026! pycascades.com/news/cfp-revie… Just sent my proposal. Good luck everyone~