Expert in building, scaling and maintaining complex web applications - Founder of @revsys, Django Steering Council, PSF Fellow and Former President of the DSF
Typo of the day. Naming a docker container after a customer's brand site I called it hotdogX rather than hotrodX.
Apparently I'm hungry for heavily processed meats.
@mitsuhiko It's probably not ideal for you, but OpenZiti is an alternative for one of those networks and can interoperate between the two. I run an OpenZiti mesh currently for my personal stuff. Not as good of a UI/UX as Tailscale, but your clanker can do the config.
If you're using Pi coding agent and not using `/tree` like I was until recently you're missing out! So freaking useful to rewind back to before your LLM went stupid.
New version of my `gg` command a little shell/web DX love for working with Github's web UI. This version adds `gg view` which takes you straight to the repo's homepage which was sadly missing from earlier versions https://t.co/s86ddJRvsT
Up at the office early for some fun coding and research time. Coding against the backdrop of the sound of light rain outside is underrated.
Now the question is what is today's focus... 🤣
Another great Lawrence Tech Guild meetup. I’m so glad this sort of in person meetup restarted and/or survived COVID. Random nerds from all walks of life hanging out. Awesome people random advice over some booze is very underrated.
In return for $100B from Anthropic, the Holy See is reworking the Lord’s Prayer to say “give us this our daily slop,” which is great news for their financials.
Unfortunately, this does mean they’ll be losing the Wonder Bread account.
Serious ask here. I’ve had a lot of great mentors and friends over the years who helped me both get into and thrive in Open Source.
My success in this area is very much on the shoulders of giants.
How can I be of better service to you and the next generation of geeks?
People ask me how AI is effecting me and REVSYS all the time. Which I completely understand.
I’m the least stressed, most productive and most positive about building software for clients I’ve been in years.
Let’s go!
We got tired of calling PEP 723 scripts (the self-contained Python files with embedded dependency metadata) "PEP 723 scripts"
We're now calling them 📜 "scrolls" 📜
Behold, a Python scroll
I had the chance this week to teach a Professional Python class to a room full of very smart people. My IQ was definitely bringing the average down.
Almost everyone in the room had a PhD.
Except for two of us.
One of them was me.
And I saw something I often see when I work with highly technical experts.
They are brilliant. They know their domains deeply. They know how to get things done. They have built workflows that work for them, often over many years.
But there are usually gaps around coding and software engineering practices.
Their job has usually been to get the analysis working, get the model running, get the paper out, get the result shipped, or get the thing working on their machine.
That creates a very different coding style than what you need when code needs to be shared, tested, refactored, reviewed, maintained, and collaborated on.
That is the fun part of teaching this material.
You start introducing processes and practices that move people from being highly capable individuals to being a team that can build together.
Testing was one of the big unlocks. Many people had not spent much time with it, but they immediately saw its value. Tests give you confidence. They make refactoring safer. They make collaboration easier. They let you change code without holding your breath.
For the skeptics, I told them to apply the scientific process (they are all scientists, so they can't do much to argue), apply the techniques, and verify whether the results are better.
They were also very interested in AI-assisted coding.
This class was not an AI coding class. I teach that separately.
But this class is foundational for AI coding because AI coding accelerates experts. The better your software development practices are, the better your results with AI will be.
If you bring messy habits to AI, you get faster messes.
If you bring testing, structure, environments, refactoring, and collaboration practices, AI becomes much more useful.
Professional software practices are not just for software engineers anymore.
They are the foundation for anyone who wants to use Python, data, and AI at a high level.