I’ve been using Podman for a while as my go to container engine. This article outlines all the reasons to do so and provides some nice migration advice https://codesmash.dev/why-i-ditched-docker-for-podman-and-you-should-too
Remote
hackeryarn
@hackeryarn@mastodon.social
mastodon 4.8.0-nightly.2026-10-06I'm a staff engineer working with Python, Clojure, and all the DevOps technologies. Outside of work, I write Rust, contribute to open source, maintain a Kubernetes homelab cluster, and dabble in 3D printing and hardware.
In the past, I taught software engineering at Washington University in St. Louis, and worked with web technologies, across many industries, since 2006.
0 Followers
0 Following
13 Posts
Joined August 15, 2019
Blog:
Codeberg:
Github:
Replying to
@Gina@fosstodon.org @dascandy@infosec.exchange @wiert@mastodon.social I had the same experience. Tried to order online from a local restaurant and couldn’t get to the store page. Turns out, it was an invisible captcha because it worked just fine on a desktop.
Zero explanation or alternatives.
Replying to
@stefano@mastodon.bsd.cafe I feel this. On my last article my spell checker broke because of an update so the final edit ended up getting published with few misspelled words… oops.
Multiple people commented that it was a technique to make an AI generated article seem more human 🤦
Just published a Python library that makes basic dictionary based numeric computation really fast… Like faster than numpy fast. All thanks to PyO3 and Rust. https://github.com/hackeryarn/redbear
Replying to
@diazona@techhub.social not my write up, but totally agree with you. It was a great article. And podman is definitely worth a try.
Replying to
@stefan@sunny.garden that’s how I started with it. Using xdg.configFile to link everything, then start pulling stuff out into home manager configs. Using home manager for linking at least it gave me one command and repo that would setup the whole system. Also helps if you share the config across multiple machines, like I do.
I keep coming back to this when I try AI or see it pushed. It's addicting, and that addiction often feels like productivity, but it's very hard to separate the two when it comes to personal productivity.
Our brains will jump through a lot of hoops to keep an addiction going. And the smarter the person is, the more creative the hoops they jump through become.
This makes such an important point. You can't make a system work faster by just producing more junk. You have to build in assurances that can build confidence in what you deliver.
This is why better languages, like #rust, and better tooling are so important. They eliminate entire categories of QA and review. They make you move faster at the scale that matters.
I’ve been experimenting more with data-oriented design in Rust. A few weeks ago I applied these ideas to a Python library using PyO3 and Rust. It ended up outperforming NumPy for a subset of operations.
In this post I detail the journey and techniques used to get there:
This is one of the best explanations of quadtrees I’ve ever seen. If you work with special data, it is definitely worth a read https://growingswe.com/blog/quadtrees
This is the equivalent of going to a bar and having the bouncer take out their phone to take a picture of your ID. Only giving you the assurance of "I will totally delete this, don't you trust me?"
I never really used labeled blocks until getting into #zig, despite using other languages that support them. The frequent use of them in Zig really opened my eyes to how they can lead to elegant algorithms.
I wonder why other language use them so infrequently.
With the internet continuously getting worse for actually useful information, I've gone back to using reference books for languages.
It really cuts through the noise, and even lets me work off line completely. Why did we ever stop using those?