They’re using a classifier to capture metadata (like “does a dog live there?” or “does a baby live there”) by categorizing sounds. That’s all technically legal, that’s why they do it that way.

They were an issue because it was the wild West and any app could keylog any other app and respond to anything. Now we actually have UX around this.

Just one year? It’s been working much longer for me.

XDG Global shortcuts portal? Has existed and been supported by plasma and gnome for a while https://flatpak.github.io/xdg-desktop-portal/docs/doc-org.freedesktop.portal.GlobalShortcuts.html

In addition to what everyone else said, I’ll tell you why I don’t think that what we have currently can become some runaway superhuman intelligence:

All “AI” people are talking about are basically “just” LLM wrappers: tools built to run a LLM in a loop while giving it access to the ability to write and run code. This gets pretty powerful, since you can do a lot programmatically (including monetary transactions), but has a fundamental limit:

LLMs are trained ahead of time. And that’s a very cost intensive, long process. Beyond that, the LLM has only a context window of finite length. So there are no long-term planning capabilities. While running after being trained, the LLM can only condense knowledge into a summary and then re-read it, a lossy process that might introduce mistakes in “understanding” and will make any prolonged task degrade over time.

And even if they figure out how to make the things actually self-modify and learn things, you can’t give every session that ability, since that’ll vastly increase the storage and performance requirements. Do sessions with that ability will be closely monitored and limited.

So I don’t see any space where an evil AI could emerge, even when factoring in an essential possible future improvement to the technology.

You gotta admit that’s a huge market.

Two additions:

• some software is also sold under a separate license in addition to GPL, so you can choose if you want free open source or paid closed source (e.g. Qt has this model)
• being in breach of license terms doesn’t mean you’re forced to do anything automatically: it means you can get sued and a judge might order you to do something (probably pay money) to heal the breach, or you can settle out of court (probably by paying money).

*with a CLAUDE.md

Because that’s the main reason people cite why they’re not programming games in Rust.

There has to be something to it as well, because it happened a few times, while people keep falling in love with Rust in all other fields

Great article, I’m going to send this one around, thanks!

But I’d like to make a distinction: LLMs are not just tools, it’s more basic; they’re just math.

As the author themself said: Claude, Codex, and so on are tools built using LLMs.

They drive you towards slop not least because they’re only available as subscription models that you pay for by token.

Maybe commodification will lead to super terse models that run on personal or institutional servers, and exist as an advanced rubber duck to test your assumptions against. Totally thinkable with LLMs, yet very far from our current reality.

People are. They’re not training frontier models yet but scientific supercomputers exist.

What makes them call LLMs “fascist”? I guess you could say that individuals who aren’t super rich can’t train them by themselves, so relying on them currently means you’re relying on these fuckers.

That would be quite the stretch though …

The response to the mail seems to think the same:

Not sure if there’s any point in saying this, but the fact that the technology has notable companies pushing it forward that are often associated with fascism does not mean that the technology itself is somehow inextricably linked to fascism, just like how Microsoft making an operating system doesn’t associate operating systems in general with Microsoft or their goals. The GR was concerned with the technology as a whole AFAICT, not the companies developing some variants of it.

I guess you could frame a neutral stance that way as it doesn’t ban AI.

But to me “pro AI” sounds like embracing it, building it into processes, …

Yeah, no matter the truth content of this piece of tinfoil, the vote I linked shows that systemd continued being agreeable among voters.

The handful of conspiracy myth believers are now on Devuan if that still exists. Presumably mad about something else that exists only in their head.

They re-render it, but that used to be capped with a max size: https://invent.kde.org/plasma/kwin/-/merge_requests/9176

No it doesn’t it just uses metadata compatible with XCursor themes, see e.g. this discussion.

The fix was needed because the render code just used to have a maxScale: https://invent.kde.org/plasma/kwin/-/merge_requests/9176

Where did you pull that number from? https://vote.debian.org/~secretary/gr_initsystems/results.txt

The smallest difference between a systemd option and a “not-only-systemd option” I see is

Option 2 defeats Option 4 by ( 211 - 177) = 34 votes.

And bear in mind that condorcet selects the most agreeable option, not just the majority option.

And it makes sense: the most likely contender at the time was upstart, which didn’t stand the test of time AFAIK (I don’t think the systemd hater crowd still uses that)

So it ended up the right decision. Would they have decided differently if other good init systems had existed already? Maybe, but they didn’t.

Yeah, when something isn’t in the official repos, NixOS exists, Flatpak exists, and they’re all vetted.

Wanting a super clean system where everything is managed by pacman is purely about taste, so if you want it, you clearly can be bothered to put in the 2 minutes of work.

That’s how you’re writing a regex validator that’s almost always correct!

def is_regex_valid(p: Pattern | str, purpose_description: str):
    return False

Thanks, good to know! Drama be drama-ing if you don’t have a good governance model …