The Changelog: Software Development, Open Source - 10 big predictions for 2025 (News)

Episode Date: January 6, 2025

M.G. Siegler goes way out on a limb with some BIG predictions of things that could happen this year, Simon Willison's year-end roundup is a must-read and perhaps the only thing you have to read to get up-to-speed on the state of the LLM, Allen Pike describes a method for magic, Tom Critchlow thinks small databases are magic & James Stanier agrees with me about Parkinson's Law and the usefulness of deadlines.

What up nerds, I'm Jared and this is ChangeLog News for the week of Monday, January 6th, 2025. Well, how did your 2024 exit? Did you close open file descriptors, delete temporary files, and free allocated memory? Or was it more of a segfault and core dump kind of a finish? Me? I'm still holding on to a couple of loose threads, but I managed to return zero and I'm ready to execute again. So let's get in to this week's news. 10 big predictions for 2025. TechG. Siegler goes way out on a limb with some big predictions of things that could happen this year,

one of which he believes actually has a chance. Here's his list with all of his reasoning removed because why not, right? 1. Apple buys an AI company 2. Someone buys Warner Bros. Discovery. Intel gets bailed out. Elon Musk bails on the White House. Amazon's Alexa overhaul proves less than remarkable. Microsoft and OpenAI kiss and make up or break up.

NVIDIA comes back to earth, a bit. Number eight, threads passes X slash Twitter in active users. Number nine, Google starts to feel real pressure on search. And number 10, Mark Zuckerberg unchained. Of course, MG does defend these predictions in the blog post, which is linked. Some of these sound not too outlandish to me. Specifically, I can see numbers 1, 4, 5, 7, and 9 happening. That's Apple buying an AI company, Elon Musk bailing on the White House, Amazon's Alexa being not so remarkable, Nvidia coming back to earth, and Google starting to feel real pressure on search. That might already be happening. What do you think? Things we learned about LLMs in 2024. Simon Wilson's year-end roundup is a must-read and perhaps the only thing you have to read to

get up to speed on the state of the large language model ecosystem. He also comments on much of the commentary around LLMs, which I wholeheartedly agree with. Simon says, quote, I think that telling people that this whole field is environmentally catastrophic plagiarism machines that constantly make things up is doing those people a disservice, no matter how much truth that represents. There is genuine value to be had here, but getting to that value is unintuitive and needs guidance. Those of us who understand this stuff have a duty to help everyone else figure it out. An unreasonable amount of time. Alan Pike describes a method for magic, quote, the pianist whose fingers seem supernaturally nimble, the presenter whose message seems viscerally compelling,

and the artist whose paintings seem impossibly realistic all wield the same magic. They've invested more time than you'd expect. It can be difficult psychologically to commit yourself to spend an extreme amount of time and attention towards a goal, no matter how worthwhile. Doing impossible things feels, well, impossible, end quote. Allen also provides a formula for getting over that fear of commitment. I'll give you a hint. It's similar to the formula for eating an elephant. It's now time for sponsored news. Mobile debugging hands-on workshop. Picture this scenario.

You get a crash report. App crashed on checkout page. But you can't reproduce it on your Pixel. Maybe it's only happening on a Samsung device. Maybe it's a memory issue. Or maybe the user was on a bad network. Now you're stuck digging through logs, guessing at settings, and running the same scenario over and over in your emulator. If this sounds familiar to you, join Sentry's Philip Hoffman and Simon Grimm for a demo-filled, hands-on workshop aimed at helping

you take the guesswork out of debugging on mobile. They'll show you real world examples and how to solve common issues like reproducing those elusive crashes and finding the root cause of performance issues. Whether you work with iOS, Android, or React Native, you'll leave with practical strategies and tools you can use immediately. Sign up for that workshop using the link in the newsletter and your chapter data. And thanks to Sentry for sponsoring Changelog News. The magic of small databases. Here's Tom Critchlow. Quote,

We've built many tools for publishing on the web, but I want to make the claim that we have underdeveloped the tools and platforms for publishing collections, indexes, and small databases. It's too hard to build these kinds of experiences, too hard to maintain them, and a lack of collaborative tools exist. End quote. Tom goes on to think through what's needed in this space, list some existing tools and examples, and make this overall point. Quote, I want to empower more individuals to publish, maintain, and collaborate on small indexes to build a million tiny libraries, community databases, weird collections, and indie indexes.

Parkinson's Law. It's real, so use it. Parkinson's Law, which says that work expands so as to fill the time available for its completion, is counterintuitive, but that doesn't make it wrong. This is why I've staked the claim that arbitrary deadlines are actually awesome, link in the newsletter, and it's why James Stanier agrees with me. Here's James, quote, projects that don't have deadlines imposed on them,

even if they are self-imposed, will take a lot longer than they need to and may suffer from feature creep and scope bloat. By setting challenging deadlines, you will actually get better results. It's all about manipulating the iron triangle of scope, resources, and time. End quote. I wish it weren't true, but it is. Oh, it is. Deadlines really help human beings get things done. Acknowledge it, embrace it, use it. That's the news for now. But also scan the companion newsletter for even more stories worth your attention,

such as hitting OKRs versus doing your job. Nobody gets fired for picking JSON, but maybe they should. And the ghosts in Spotify's machine. Hashtag dead internet theory is real. In case you missed it, our last episodes of 24 were a couple of bangers. Our final interview was with Mitchell Hashimoto talking ghosty, which is publicly available now, by the way. And our final friends was State of the Log 2024 with 12 listener voicemails plus BMC remixes. Scroll back in your feed if you haven't listened to those yet and hang tight

for some awesome pods this week as well. On interviews, Rachel Plotnick joins us to talk buttons, knobs, and switches. And on friends, well, it's only Matt Reier with his guitar and a list of ridiculous topics to discuss. Have yourself a great week. Leave us a five-star review if you dig the show, and I'll talk to you again real soon.