Low-code tools have guardrails that keep things predictable - you know what a Retool app can and can't do. AI-generated internal tools are just... code. Code that will need updating when APIs change, when requirements shift, when the person who prompted it leaves. "We migrated everything in a couple of sprints" is doing a lot of heavy lifting. Check back in 18 months when half those tools have drifted into unmaintained states and someone has to figure out what handleEdgeCase_v3_final_FIXED.js was supposed to do.

Well thats the problem right there, business logic belongs in the database, not JavaScript rats nests. I’d argue that Claude and even Coeldex are so good with SQL that it would be insane to use any other approach. And the database survives the hot framework of the week churn.

Worth noting that most of these GTK4/libadwaita players are going to look out of place on anything that isn't GNOME. If you're on KDE or a tiling WM, Strawberry or one of the Qt-based options will integrate much better

I've found that libadwaita apps tend to look at least decent outside of their native environment, whereas QT apps near-universally look terrible outside of KDE.

I haven't really looked into this, but is it possible to make GTK4 apps look liek standard GTK2/GTK3 applications? It feels like every single modern GTK app I've encountered has that modern Rounded-Material look to them and ignores the window manager decorations.

> and ignores the window manager decorations.

That's because Gtk4 does "client side decoration". That has the advantage (or otherwise, depending on your point of view!) that the application can now place custom widgets in the title bar, and the disadvantage that when apps do that, the part of the title bar available for dragging windows around becomes significantly smaller.

My main objection to client-side decoration is that middle-clicking a window's title bar to push it to the back no longer works. (Plus, for those of us with eyes that aren't as young as they once were, it's now much harder to choose a window border style that clearly indicates which window has focus.)

My biggest problem (of many) with client side decorations is that now when your program crashes, you can't just hit the close button to have the window manager kill it, because the process responsible for drawing and responding to the close button has crashed.

The trick is to avoid software using the newer gtk versions.

I am running KDE, and they look just fine. If you mean they won't follow your theme, yes, but also a lot of other apps don't (e.g. Electron).

Feels like there’s an unexplored overlap here between audio games and systems thinking — once visuals are gone, you’re forced to model the world more rigorously. That pressure seems like it could make better games, not just more inclusive ones.

I checked Slack twice while reading this

I've been burned before by driver bugs that only manifested under very specific timing conditions or malformed responses from the device, tnx

Anytime, hopefully it fits your needs and helps you not spend more time than needed tracing issues like this. Thanks for the comment!

“Move fast and break things” looks a lot worse when the things being broken are clean-air laws and the health of nearby communities

The biggest problem with LLM reviews for me is not false positives, but authority. Younger devs are used to accepting bot comments as the ultimate truth, even when they are clearly questionable

I alluded to it in a separate comment but the problem I have here is that it is really hard to get through to them on this too.

Upskilling a junior dev required you spend time in the code and sharing knowledge, doing pairing and such like. LLMs have abstracted a good part of that away and in doing so broken a line of communication, and while there are still many other topics that can be tackled as a mentor, the one most relevant to an upstart junior is effective programming and they will more likely disappear into Claude Code for extended lengths of time than reach out for help now.

This is difficult to work with because you’ll need to do more frequent check-ins, akin to managing. And coaching someone through a prompt and a fancy MCP setup isn’t the same as walking through a codebase, giving context, advising on idiomatic language use and such like.

Yes, I've found some really interesting bugs using LLM feedback, but it's about a 40% accuracy rate, mostly when it's highlighting things that are noncritical (for example, we don't need to worry about portability in a single architecture app that runs on a specific OS)

“You don’t need an ORM” — after a few difficult debugs, it sounds almost like therapy

A good illustration of how “can run LLM” ≠ “makes sense to run LLM”. A prime example of how numbers in specs don’t translate into real UX.

This is probably the most clear explanation of the seasons and the changing altitude of the Sun that I have seen. This would be perfect for school lessons or popularizing science.

Here's a fascinating video of what the sun does at 90S: https://vimeo.com/208466944