Debian packages are ancient, unless you use "testing", which will be broken. That being said, Ubuntu is horrible too. The problem comes when you have to use proprietary software made by incompetent vendors, and that only works in Ubuntu.

Ubuntu is basically a packages imported/synced from debian (I think they mostly cherry pick unstable) + ubuntu specific stuff.

Debian Testing is pretty much what many Debian users use on a daily basis. They pretty much invented the notion of rolling releases. Debian stable is of course not cut that often. Ubuntu has a higher frequency of updates. But still, their LTS releases are spaced two years apart. 26.10 follows the recent LTS 26.04 release, so that gives them a good 1.5 years to work on stabilizing this. I've been using Ubuntu server LTS releases for the last ten years or so with generally no major issues. There are always some minor headaches between updates but generally nothing that worries me.

The main criticism here seems to be that these packages need a bit more work or are still lacking certain important features. Getting them out in a non LTS release like this is how they get packages to be ready for LTS releases. If you need LTS now, 26.04 is what you should be using. By 28.04, this should be rock solid.

For me what's risky here is deviating from what other distros are doing. Including Debian. Ubuntu has a history of being opinionated not everything they push through being adopted with great enthusiasm (e.g. Snap). You could make the point here that waiting for Debian to pick this up would be nicer. But you could also make the point that waiting for that is a very slow process.

> Debian packages are ancient

ever used backports? no need to go to sid. Flatpack/Appimage/Compile from source for the rest (on my debian machine it's just kicad that is not installed from APT, well and the couple of softwares that don't do deb)

Firmwares, that's another story, i have fedora on the laptop for that.

> The problem comes when you have to use proprietary software made by incompetent vendors, and that only works in Ubuntu If i want to do FPGA development, either ubuntu or a distro old enough to package JDK 11 (sigh) and X. not XWayland. Debian 11 was the last one to have JDK 11 i think (talk about ancient..)

Needless to say FPGA is done on the windows machine these days.

I feel society is going in a direction where there is no longer any individual responsibility. Instead everything is delegated to the state. Can't limit your kids screen time? Not your fault, it is the evil algorithm and it is up to your government to discipline your children. Being a parent will be reduced to providing your gametes.

My impression was that the major argument against static site generators is the unfriendly UX for non-technical contributors to the site, and with LLMs that problem is completely solved.

I don't know how it's "completely solved" unless you can point me to an existing, complete solution. Providing a nice backend with a WYSIWYG editor and access control as well as varying level of customization is not a trivial problem for anyone, technical or non technical, and you can easily mess it up.

You don't need a backend. You don't even need a static site generator. Just sit a non-technical person in front of Claude Code and tell them to tell the computer what they want (including what the content should say). For something as simple as a static site (that the user can themselves look at and read) there's nothing to mess up -- if something looks out of place, a non-technical user can just tell the computer that and it will fix it. "Computer, add a page to the blog with this content: <pasted text>" -- "Computer, the banner looks weird on phones but not on desktop, please fix it."

My wife (no CS background, no programming experience, schooling in music and law) is about to submit her first several apps to the App Store. I don't think she has had to even look at a single line of code. Yes, they work.

You need to get out of your bubble. That's all I can say.

I ... need to get out of my evidence-based bubble in which I'm observing real non-technical users in the real world? This is not a discourse bubble. I'm not getting my talking points from other people. I'm drawing conclusions based on what capabilities I've observed the AI giving to non-technical users.

You only need a WYSIWYG if you're the one doing the actual writing. /s

Destroying things is part of human experience. Kids got to have that.

AI is trained with an overwhelming amount of web-related code that is open source. For embedded, not only is the training set smaller, but much of what's publicly visible is not of great quality.

It is called "Reggaeton be Gone"[1]. I think the author had too many problems with his neighbor.

[1] https://github.com/ronibandini/reggaetonBeGone

If it can transform into Winamp I don't care if the code is LLM generated, typed by infinite monkeys or found in some ancient scrolls.

Jokes aside, how can you be sure that the official Spotify app (or any app, or Windows) was not written by a LLM without supervision?

Not OP but I don’t really care if it was or not. All I know is that I’ve been using the Spotify app 10+ years and it works well. I don’t know if this random app works, and if it hasn’t been supervised I doubt it does.

Gave it a quick try today, it works pretty well and is noticeably faster than the official app.

The official Spotify app has some skin in the game. Random dudes self promotion project, not so much.

I suspect the official app would actually be more consistent and not as bad to use if it was generated by an LLM.

If this could be made to work reliably, it would imply that cameras would have the ability to doxx users.

Regarding the Egyptian messages, if you care about these things you simply avoid doing business (especially in regulated industries) in the third world. The fact is that, even before the incident with the messages, cell providers had to sit down with the "men with guns" to even be able to operate in the country.

According to the article, the scanning facility has been there for a while, so it may well be that they have been destructively scanning books for a while, even before AI. It makes sense for the major book retailer to have digital copies of every book so that it can begin selling those as soon as it can negotiate the rights (if there are any).

In any case, all the indignation about destructive digitization misses the point that rare books takings space in a warehouse for years without being bought will eventually be destroyed anyway.