Collin Donnell

    • Hire Me
  • PhotosMy photo blog.
  • MusicMy music website.

  • First Impressions of macOS 26

    Yesterday I decided to stop waiting and install macOS 26 Tahoe. It’s a mixed bag.

    Everything they added for apps other than the design updates, I like.

    I’m a big fan of the Journal app. Having it on macOS is awesome. I’ve never been able to stick with Day One, but for some reason Journal clicks with me. The suggestions it gives are good at reminding me what I was doing, so even if I don’t visit the app for a few days, I can come back and fill in the details.

    The updates to Spotlight aren’t huge for me, but what’s there is good. Being able to quickly switch to an app-only list is great for when the app I’m looking for isn’t the top result. I want to play with the actions list more. I’m wondering if having a good launcher for shortcuts will encourage me to use them more. For some reason the built-in clipboard manager only holds onto what you copy for eight hours, so I’m going to keep using Pastebot. If they remove that limitation, I’ll probably try it.

    Relatedly, Launchpad has been totally redone and I guess is called “Apps” now. I never used or liked Launchpad to the degree it was hard for me to imagine that other people did. Instead of being a weird fake Springboard for macOS, it’s now a proper app launcher you can filter by category. I’m going to give it a keyboard shortcut and try using it in place of Spotlight.

    Terminal is my favorite terminal app. Now that they added better color support and whatever you need for Powerline to run, I don’t see myself needing any others.

    The new design language? It’s weird. Almost no third-party apps I use have adopted it, and I don’t know if I want them to. It’s really strange seeing the two styles next to each other, though. Probably they should eventually, but I think it’s going to make Electron apps and others that use custom UI seem even more out of place.

    Liquid Glass is definitely better on iOS. At least there it feels a bit alive and I can see what they were going for. On macOS it’s just weird slightly-transparent bubbles everywhere. The new drive icons really bother me for some reason. The perspective is off. They aren’t flat, but they aren’t tilted back the right amount either. It feels like they’re going to tip forward out of the screen.

    My review for the design changes is this — mostly not awful, but everything they changed is worse. Every change rests somewhere between neutral and a little bit degraded. Not hugely terrible, but just a little bit less pleasant than before. I guess that kind of adds up.

    6/10

    October 2, 2025
    macOS

  • What if Humanity?

    For most, LLMs are useful for two things. As a search engine, or to work out your own ideas. Other than those, they can be actively harmful — using them for medical advice or therapy. Everyone seems to hate half-baked AI features infecting everything they use.

    Who likes it are executives. These ghouls hope that it will somehow let them fire and replace people. It’s clearly not to that level yet. In the short-term, it lets them pretend it can replace humans. They’ll fire people. They’ll make the rest work more to make up for it.

    What the rest of us mostly hear is that AI is going to take our jobs. People supposed to be excited by that? OpenAI and Anthropic don’t want to make the world a better place, they want to sell a product. They don’t care about any damage it causes to the rest of us. Executives care about ROI. Maybe they’ll write a letter or fake crocodile tears on a call before rushing off to a yacht trip in Monaco.

    My question is this. What if we treated each other with empathy? What if our goal as humans was not acquisition and expansion, but treating each other kindly and with respect? What if our packages showed up in three days instead of two, but no one was pissing in bottles? What if our number one priority was each other, and only secondly speed and efficiency?

    I think that world might be better for everyone — including all the little Jack Welches running the world.

    August 22, 2025
    AI, Capitalism, Society

  • Mark Twain on the Book of Mormon

    Part of Mark Twain’s semi-autobiographical book Roughing It talks about his visit to Salt Lake City in the 1860s. His review of the Book of Mormon is absolutely worth reading, because it’s hilarious. He talks about the overall style is was written in and pulls out many specific sections which he entirely pulls apart.

    The beginning of the review gives you a pretty good idea how it’s going to go.

    ALL men have heard of the Mormon Bible, but few except the “elect” have seen it, or, at least, taken the trouble to read it. I brought away a copy from Salt Lake. The book is a curiosity to me, it is such a pretentious affair, and yet so “slow,” so sleepy; such an insipid mess of inspiration. It is chloroform in print. If Joseph Smith composed this book, the act was a miracle–keeping awake while he did it was, at any rate.

    August 20, 2025

  • Playdate Development Is Incredibly Fun

    For the last couple of weeks, since my post about Playdate Season 2, I’ve become obsessed with learning the Playdate SDK and learning to make games. The project I’ve been working on as a learning experiment is a Breakout clone that uses the Playdate’s crank for movement, which I’ve titled Crankout. I posted the source on GitHub if you’d like to look at it.

    Making a simple game like this, which I can continue to expand and improve, felt like a good strategy for learning game development. Part of my approach to learning new things is to pick the simplest version of it I can think of which still demonstrates how something works. If I can’t start using the thing in some form almost immediately, I tend to get distracted and not get very far. My MiniCalc project I wrote in Ruby is another good example of this I wrote when I wanted to try making an interpreter.

    What’s making this fun is that game development is a totally new area for me, and that the Playdate SDK is extremely accessible. The SDK is very bare bones, while still having all the pieces you need. If you look at the official documentation, it’s just a list of functions. There’s no engine or anything like that. It reminds me a lot of making games in QBasic when I was a kid.

    If you’re interested, here’s resources that have helped me the most, aside from Panic’s documentation.

    The Programming in Lua book. Written by one of the authors of the language. It’s short, approachable, and covers everything in Lua. I read this while taking notes in my bullet journal, and by the end feel like I’ve pretty much got it. There’s a couple of parts I skimmed or didn’t bother doing the exercises for, but I know what they are, so I can go back if those things ever come up. It’s a really small language and easy to learn.

    SquidGod’s Playdate SDK tutorials. This guy should be on Panic’s payroll. He’s been putting out videos regularly on the SDK for years, and they’re incredible. Probably the number one resource for learning not just the Playdate SDK, but game development patterns as well. He also has a Discord that he’s active in, and he seems like a really nice guy who wants to help people learn.

    Eventually I’m hoping to make a real game I can put on the Playdate Catalog, but for roughly two weeks in, I feel like I’m doing pretty well.

    August 17, 2025
    Game Development, Gaming, Lua, Playdate

  • How I Learn Things Quickly and Deeply

    Something I’ve been told a lot is that I tend to pick things up faster than most. I don’t know if that’s true — maybe everyone does exactly what I do. Maybe there’s a better way. Regardless, I have a pretty good idea of how I learn things, so here’s what that is.

    The way most tutorials, guides and lessons work is by breaking things down at a high level into easily digestible chunks. I hate this. I feel stuck at the surface. I’m being told how, but not why. Maybe I’ll be able to do the mechanics of something, but that’s a slow way to expand my knowledge. What I try to do is see past that and find the concept that underlies everything. Once I’ve got that, there might be some specifics to figure out, but I’ve basically learned it.

    Here’s some examples.

    Let’s say you’re learning scales on guitar — major, minor, pentatonic, whatever. Don’t worry if you play guitar or not, that’s not important. The way this is often taught is by learning five distinct box shapes that you can play in different positions on the neck.