After being so sick for a week that I couldn’t work out and my daily routine falling apart, I went to Montreal and got sick again on the flight home, which left me, again, unable to follow my hammered-in daily routine. It’s now been three weeks and my cough is finally, finally disappearing (I feel like I’m jinxing it by writing this), but so did my daily routine.
In the past week, I’ve embraced it: ignored my usual side-projects, did whatever I wanted in that now-free time, worked more, read more, wrote more. That’s why the last two issues of this newsletter have not included a Joy & Curiosity section. I just wasn’t in the mood.
Now I am.
This blog post by John Calhoun, who has worked at Apple for 26 years, is lovely and contains the cutest ASCII art I’ve seen in a while. The whole blog is a treasure chest. This tribute to fellow Apple engineer Tom Dowdy is wonderful.
Some discoveries don’t just leave you with new knowledge, but in awe, marvelling at the complexity that can be contained in the tiniest possible details and how much work and care others put into figuring out that complexity. The bytecount crate left me with that feeling: it “uses the ‘hyperscreamingcount’ algorithm by Joshua Landau to count bytes faster than anything else.”
Mitchell Hashimoto wrote about Zig’s comptime and the examples he used are great: small and simple enough to understand for anyone, but showcasing the feature well enough. Very hard to do.
I wish I could remember how, but I did end up on this website by a guy in California who builds strange-looking headphones by hand. No, you won’t guess what they look like before clicking.
This blog post, Rewriting Rust, was very interesting. “I swear, it took more effort to learn pinning in rust than it took me to learn the entire Go programming language.” Some day (in the far future) I might write more about my feelings on Rust, but while reading this post I kept waiting for the paragraph in which they say what they would remove from the language. That paragraph never came and I think that’s one of the biggest sources of friction between me and Rust.
My former coworker-at-Sourcegraph Varun wrote about leaving his previous job at Apple. The section about recruiters stuck out for me. In the past, I’ve interviewed at two companies that I respected a lot and the technical recruiters I was in contact managed to destroy that reputation completely within a week. One of my niche beliefs is that tech companies greatly underestimate how reputation-destroying recruiters can be and the whole system works in a way that makes it very hard to find out.
This tweet made me re-discover the original blog post by Lennart Poettering, the author of systemd, on systemd. I vaguely remember reading it years ago, but now, re-reading it, I’m amazed by how much knowledge is in there.
The team behind iA Writer wrote about developing for Android. “Developing for Android you navigate an asteroid field. Bugs surface across thousands of device types, Android versions, and flavors—One UI, MIUI, OxygenOS, Pixel Experience, you name it.” (Be sure to read the footnotes.) It made me think of our experience building Zed for Linux. It ain’t easy.
Somehow came across this blog post on Visual Studio Code being “designed to fracture” and, man oh man. It made me jot down a note, telling myself that I should write about how tied to VSC the whole language server ecosystem is.
When I was a teenager eager to learn Linux, the #1 recommendation I came across was to “buy the Kofler book” — a reference to Michael Kofler’s book Linux. It was already a very big, very heavy book back then, but now it’s in its 18th edition and has 1410 pages. And, fascinating to me and why I’m writing this after not having heard the name Kofler in a long time, I learned last week that Michael Kofler is still actively blogging, about such things as running Ubuntu on macOS. Clicking through his blog felt like discovering a wormhole connected to 2003.
Craig Mod wrote something beautiful about meeting Kevin Kelly.
This interview with Mark Zuckerberg was fascinating. Fascinating because here’s the CEO and majority shareholder of Meta saying the same things that any tech lead in a small, 5-person shop says too: you need to ship fast and early, you need to have a great feedback loop and then iterate and iterate. When Zuckerberg says it, though, it doesn’t sound empty. But then another interesting bit was that he essentially downgraded the importance of innovation. He said that you have to innovate one or two things, otherwise you won’t be on the map of tech companies, but their strategy is also to wait and see what others do and then do it better. Might sound trivial, but not if you hold it against a whiteboard full of tech leader quotes saying how important innovation is. (Reminder to listen to this podcast with Kent Beck on Facebook Engineering Process — an all-time top 5 podcast episode for me.)
I just ordered the 2025 CALENDAR OF EXTREMELY ACCURATE BIRDS.
Have a great Sunday, friends. I’ll see you on the Internet.