PyTexas 2026 Recap

8 points by gaborjbernat


confusedcyborg

For a language focused con that recap seemed to be really focused on AI/LLMs. I guess that's to be expected given python's popularity in the space. The bit about python as a DSL was interesting. I wonder if there's examples of instances where python doesn't provide sufficient expressiveness to model a domain in the way a domain expert would think about the problem.

mtlynch

Al wrote Automate the Boring Stuff with Python

I was very surprised to see the claim that this book was AI-generated, and then I realized that was a lowercase L.

So, it was actually that "Al [Sweigart] wrote Automate the Boring Stuff with Python."

Also, thanks for this writeup! I wish more people would write about their useful takeaways from conferences.

aleyan

Miguel Vargas’s framing: AI agents produce cleaner, safer code in codebases that are already clean, safe, and typed, so Ruff, ty, and uv matter more now.

I had the same thought, so a few months ago I built script-prompter that instructs agents to construct single file python scripts with inline dependencies, tests, lints, and type checking (uv, pytest, ruff, ty). From using it myself a little bit, the results seem fine, but ymmv. Would love to hear what folks think.