I agree.

As far as I know, most distros are waiting. I heard debian is doing that (people say their org structure makes them unable to satisfy it), lead dev of pclinuxos said it might be a "nothingburger", the distros who said "no plans" are doing that, etc.

We might get the "stack" in place before distros do anything. Only a few distros refused outright so far, even less are trying to comply.

I think we might linger in this limbo for a long-long time, and the distros might only decide when websites start requiring and API (or when they get taken down by authorities).

Possibly a relevant historical parallel on how this might happen (the differences are surface-level): https://lists.debian.org/debian-devel/2026/03/msg00359.html

I agree with most of what you said, however, I'm not worried about most people. I'm more worried about people who are trying to act against tracking. Maybe I shouldn't, and we cannot "personal responsibility" our way out of this, but I still haven't processed that loss.

I'm also not worried about “user is likely accessing service from a *nix device” in itself. It's hard to say how insignificant this particular bit is. I'll prolly need to spoof this info, too.

The systemd change is fairly insignificant, yes. That the xdg-desktop-portal people will implement this is more worrying. I'm also worried that the systemd addition will be used as "proof" that the linux landscape is perfectly capable of adding age attestation and verification, and therefore they will dismiss any possibly amendment to the US bills and the EU wouldn't even consider adding an exception.

If the law creates this new API, nobody will be able to get away without leaking PII to the web (I think it will have a javascript API). 1970 jan 1 identifies *nix users, 1900 jan 1, even an "unset" will leak info on people, etc. entropy and shit. Websites will also try to use the API. One could set the function to crash JS, or disable all JS, but that already breaks 90% of the web. One could also: either not install userdb (like I do), or install a non-systemd linux, but those still will leak "unset" or "API disabled".

There's no getting away from this, it really is the law that we should push back against. Unfortunately it's a fight currently only Americans and Brazilians can participate in properly. Donations won't work well, govts these days often clamp down on NGOs because they often get money internationally, and therefore, they are "foreign interference" etc.

I think the discussion we should be having instead is engineering ethics.

I think Dylan's letter clearly showed he isn't really involved in "FOSS radical" circles and has the usual engineering mindset (that was described in the Sam Bent blogpost). Not that I personally blame him, engineering schools raise students for this, and the corporate structure further enforces it with the strict hierarchies where you do whatever management orders you to do. What's important is that there are many devs who clearly aren't tied into the ethics discussion part of engineering. There's even research on engineers' attitudes afaik, and its conclusion was that freshly graduated engineers' first response was to implement changes without thinking about the ethics part, and after further prompting and discussion, they could be convinced that this change was bad and they need to push back.

I'll leave a two links here so when people iwsh to read, they know where to find good stuff.

• Constructive Technology Criticism Annotated Syllabus - I know Medium sucks but the scribe.rip copy I saved suffers link rot
• Awesome critical tech reading list - you can just git clone it and read it without browser

I don't wish Dylan harm, and I'm not doing anything to him. I also believe nobody here is sending threats. If you saw the video, you probably noticed that it contained a screenshot from 4chan. 4chan users have been known for terrible behaviour and they are conspiracy-minded fascists, who also oppose this change (others fascists would like this as long as it's not them being spied on). I'm almost completely sure that they are the ones doing the bullying, not users from here.

I'd prefer if the age verification landscape would be fragmented an unusable, compared to systemd offering it in a consistent manner. Websites will use the offered APIs and will use it for extending fingerprinting. No, fake birth dates won't save you. Even the disabled canvas API is used for fingerprinting. It just shouldn't be exposed at all. Not that it matters because every other OS will comply and desktop linux is negligible.

The arguments presented in the video won't convince anyone who bully people on the internet. They are most often fash and they believe that only power matters. Bullying is exerting power over people, and if they succeed in bullying them into reverting the change, they will be satisfied. Not that I think it will achieve anything, but they do. They follow Carl Schmitt's teachings.

On "better ways to make a change": If somebody doesn't live in the US, and lives in sort of a dictatorship, they can neither affect murican lawmaking nor do their govt listen to anyone other than a few. If this age attestation/verification shit comes into place, their only choice here would be to go and not install systemd-userdb (or use linux without systemd). The disabled API would probably break even more websites than disabling 3rd party javascript. Their govt could also use the fingerprinting to spy more on citizens (they prolly already do).

As a thought experiment: please recommend them a better way than bullying. No, not living in a shithole country is not an option. No, they don't have the spare money to found/donate to an existing org that fights against this. Also, companies pushing this would just outspend them.

There's also the aspect of corporate influence over linux. It directs changes in order to further business interest over normal users' interests. Personally, I prefer companies be out of linux and would accept lower-quality things. But also, I think the ones most hurt by these laws are the system integrators (mentioned in the video), who actually need to do things that align with normal users' interests.

On the parental control thing: yes, age things could be used for it, but the parents know better than the computer, and user settings would be preferable (for example, kid should be able to this and that program, open this site, but not others). I think it shouldn't be the websites who decide. Yes, parenting takes a lot of time, but we can't substitute it with automatated fence-building.

Edit: elaborate on "fascists.

I've been running debian sid on desktop for 4 years, I think. Yes, I don't care if it breaks. I wanted to try debian and didn't want to use old packages at that point. These days I don't really need the latest things. I recently switched to testing - I only needed to replace a few words in a few files in /etc. I didn't even need wiki or anything for that, because testing is almost like sid. If this doesn't break on me majorly, I might not switch and just replace "testing" with "forky". I'm really satisfied with debian.

Others already explained basically everything. I'd like to elaborate and offer a few examples to support them.

• On potential users:

  • The people who look for distros to try are seen as newbies by linux users, and therefore are recommended newbie-friendly distros. Also, debian is conservative: it rarely offers shiny new things, so its desktop use isn't high. There isn't much to be excited about, so there are no hype cycles. The current "shiny new thing" in debian was the recent change in apt's interface (now it formats its output into tables, for example), compare that to "atomic" distros. People often still use apt-get (it is in the guides for some reason) instead of apt so even this news in nothing to them.

  • Furthermore, software development often happens with the latest libraries around. It's often a great help that Arch ships the latest software. Debian doesn't have that. While languages these days have their own package managers, having the latest devtools, editors, etc. to try out is harder to do on debian. Therefore, IT students and software engineers have better time on faster-moving distros. Debian is more for the sysops/sysadmin people ( you can leave it there on auto-update and not care for 2 or more years ). The above further restricts its appeal and userbase.

  • Even further, Debian might be bigger than it seems, as others have pointed it out. Perceived marketshare is often based on desktop use. See EU OS's FOSDEM presentation on how opensuse has a bigger company behind it than ubuntu.

• On "latest drivers":

  • It used be much harder to configure Debian. Recently ( think it was with Bullseye) I installed it on an old machine, and debian didn't install the right wifi drivers by default. I think it also lacked the proper firmware. This changed only with Bookworm. Back in 2008, I also tried it on my pentium 3 I had then. Debian didn't have ath5k at that time, and the ndiswrapper hack was harder to pull off for me than just using mandriva, or later, lubuntu and salix.
  • I heard that these days, people expect linux to fully support their hardware on day1. They also expect it to just run on any new hardware they buy. Also, games often need the latest optimizations in drivers: it might just be the thing that pushes the fps count above 30, 60 or 120. They also that they want the driver bugfixes to come ASAP. Early on during a release cycle of a game, driver updates sometimes give big improvements. While using the latest drivers on debian is possible, and not too hard (Compiling a newer stock kernel is easy, even if it complies slowly. Mesa isn't hard either. Still, these require knowledge of old & basic dev tools, and also new ones.), ubuntu offers new drivers to LTS kernels, they are called HWE. No idea how doable this on debian, I never needed such things.

• On "stability":

  • What people usually don't think about it that there are different kinds of stabilities. Debian offers something like API stability, so that user-provided software on the same version of debian rarely - if ever - breaks. It's not necessarily shipping the most stable software, but it has a guarantee that updates won't break anything. Even a slight change can disqualify being included. This very slow process resulted in the old and famous xscreensaver vs debian drama. The abovementioned stability also applies to other distros, but to a lesser extent, I believe. Mostly due to the 2-year release cycle.

• On "ease of use":

  • Debian doesn't have a user repo like AUR, so it isn't as easy to install 3rd party stuff (I know, makedeb, flatpak, snap, pkgsrc, nix & guix exist), debian is so big that anybody providing packages will do it (to list a few examples: freetube, discord, librewolf, signal, Trinity DE, and there are bleeding edge emacs packages available).
  • Debian has docs, but I often just use arch wiki or the gentoo wiki to figure out stuff. I can only do that because I understand the differences and the similarities. Newbies would have trouble with this. Also, ubuntu automatically configures a few things, like installing something with a systemd service also will enable that service. Debian doesn't do that.

I have a few such netbooks. I'm currently running debian oldstable (bookworm) on them. GuixSD also provides x32, but I still have to try that on slow computers. (Sidenote: maybe old Mesa versions work better for GUI). Ofc, it's ok for trying BSDs (or maybe experimental stuff like Hurd and 9front).

About usage: you can put it into your garage, workshop, storage room, whereever you wouldn't want your regular laptop (gets dirty, dusty).

For "desktop" purposes:

• emacs (editing text, taking notes, developing some software, reading email, rss client)
• reading books: epub and especially pdf were made for MUCH slower devices (you should avoid scanned books)
• IRC and matrix (in emacs, for example)
• discord client (yeah, I know, you shouldn't do it)
• can play 360p H.264-encoded videos (you could use a smartphone for that, but I don't)
• play mp3, act as a radio and play music or podcasts from the internet
• SSH and other remote access stuff
• testing whether software you write could run on slow hardware.
• it is a terrible experience to run a browser, but it works. I could browse the catalogue of a local library from it
• if something was doable in the 90s, the machine can handle it (mine came with diablo 1 installed)

As others already wrote, it's also good for homeservers (web/gemini/gopher, git, mumble, irc bouncer).

If it is an actual nettop, and not a netbook, it probably has a mini-itx board with PCI on it, which makes it able to test/use old PCI cards. I used it for that purpose a little bit (there are better options, PCIe->PCI bridges). The atom D525 nettop board I have also has a mini pcie slot, which I converted into a full-sized one. Now it has a similarly slow Radeon HD 6450 in it, which helps it play videos. Should work up to 1080p, but now I realize I haven't actually tried that.