I know it's very recent but the latest models no longer produce slop. This article would have made sense 6 months ago. Today it just tells me that Torvalds is surprisingly up-to-date and whoever wrote this is behind the times.

It's too late for this to have any effect. It's more "brace for the shitty unnecessary storm".

If you thought "main" was stupid...

There's a lot of people here that clearly haven't ever tried what you're describing op (or maybe tried recent AI at all).

I have. I think it's a totally valid question. Sometimes when I'm getting AI to do stuff it reads the docs and then tries stuff anyway and finds the docs are wrong.

I also often use AI to generate docs for undocumented things.

But I think it does still make sense to have docs for a few reasons:

1. They show the intent explicitly, rather than having it inferred. Same reason anything explicit is better than implicit, e.g. static types.
2. It spends fewer tokens to read docs than to constantly reverse engineer software.
3. It's also quicker.
4. Sometimes you do still want to understand the thing...

Why do you drive a car?

How did he move them?

Surely this should be purple?

Yes but that tracks issues of all severities and with all of GitHub's services of which there are a great many. Also "days since incident" is a terrible metric because it doesn't capture how long the incidents last.

For example the latest incident is that copilot had degraded performance with Fable due to an Anthropic issue. It only last 18 minutes, only affected part of a part of their service, and it wasn't even their fault!

That's very different to "all of GitHub is down for 7 hours".

So yes. I am not sure GitHub is worse.

I can think of a few potential advantages:

1. Portability: It means you can change your hardware target and you only have to worry about re-certifying this, instead of an entire compiler.

2. Testing: You can probably run your WASM code fully deterministically in a deterministic simulation much easier than native code. Also when developing you don't need the actual hardware.

3. Isolation / compartmentalisation. If I were them I would be strongly considering using the WASM component model (in future).

It would be good if they put their actual motivation in the readme though.

It doesn't mean you made the wrong decision. Changing the names has a potential cost - you break something (this is especially risky for environment variables which have insanely global scope, no static type checking, etc.).

Not changing the names has a potential cost too - causing confusion in future, wasted time, or even bugs due to the increased chance of mistakes.

I'm currently leaving a company where they almost never clean up their messes. Tons of old unused code is just left in the repo. Causes huge wastes of time.

Don't underestimate the potential cost of not keeping things orderly. There was a famous case (Knight Capital) caused by not cleaning up an old unused CLI flag and it cost them $440m and destroyed the company.

Not really in practice. In older versions I have done it, but they really don't want you to. The binaries are only distributed as shared libraries, so you at least need to build Qt from source. And also whenever I build recent Qt apps even with QtWidgets it has a load of plugin files and other nonsense that it needs.

I'm sure it's possible with enough work but it's not simple like with fltk or egui or gpui.

Riiiiight. You might want to look up what Hitler actually did. It was just a little worse than anything these companies have done. Just a smidge.

That is... shoddy. This is basic stuff.