Rewatching @hynek's keynote from @PyConUK this year. (Not yet on his Talks page, I noted.)

Can't miss the chance to (re)?post this slide:

> Giving everyone what they want is a failure of leadership

Did you know @ThePSF is a charity that runs on a budget that is a vanishingly small proportion of the money that Python-using corporations make, but we still put on PyCon US, run PyPI, and (try to) fund grants that support Python events all around the world.

Changes in the tech and politicial landscape make it ever harder to fund the essential work we do. If #Python is part of your life, read this post that connects the dots, and consider donating or sponsoring our work: https://pyfound.blogspot.com/2025/10/connecting-the-dots.html

After the last couple of years I can now safely say several unpopular opinions:

* Yes, do NIH - if you have a sense of pragmatism. You'll be independent from vendors and can afford to just not care.

* As little dependencies as possible. Today's framework seemingly making you faster is tomorrow's ballast you're trying to get rid of

* Rewrite. Constantly. Don't hesitate. With a solid sense of YAGNI, it'll totally pay off

* Have more tests and checks than observability. And lots of tooling.