Since this seems to come up a lot, especially now with Windows being replaced by Copilot and Tim Apple being a frequent guest at the White House, I thought I'd write up my quick and objective guide to choosing Linux distros:
I just want something stable and easy to install that's still being maintained next century: Fedora.
I want a carefully hand crafted desktop experience curated for me by people who know what they're doing: KaOS.
I want to carefully hand craft my own desktop experience from scratch, I don't care how many wikis I have to read: Arch.
I think I'm somewhere in between those two extremes: Manjaro.
I'm a gamer: Bazzite.
I'm a Mac: Elementary.
I'm a Haskell weenie: NixOS.
I think rms made some good points: Debian.
I think Tommy Robinson made some good points: Omarchy.
FX Radio is back online. Still have no idea why this thing hard crashes about once a month. Hardware failure? Reading the kernel log didn't tell me anything interesting.
Some criticism of people using AI is just mean. Just saw a post that said:
> Instead of using AI images to illustrate your blogpost or presentation, why not go deep into the forest, lie down in a pile of autumn leaves and allow the earth to consume your body
Seriously? You know, not everyone is as immersed in all of the things that are wrong with AI as we are. Some people are just trying to get by in the system of incentives that they didn't create.
People just hate because something that took them 15 years to figure out is now doable by the press of a button in mere minutes. I don't get it. Its literally convenience why not use it? As long as you fact check your research and test things that you're having it code, why not?
@threlm4280 Yes, there are probably people who hate AI for the reason you described. I want to push back on this point though:
> I don't get it. Its literally convenience why not use it?
There are in fact better criticisms of current AI products. They're trained on basically all of the work of all humanity (or as much as the AI companies could get their hands on), without the consent of the people who did the work, for the explicit purpose of replacing labor. Worth thinking about.
I suppose there is validity in that. I don't think about it too much though because stealing the work of others has been a thing for ages even before AI came along to the stage. So to me, that's just finding a reason to criticise for the sake of it. Using the work of others in a modified fassion is not even remotely new.
You know what, I'm just gonna say it. Some people with depression or other issues, if they get enough of those kinds of comments, might just take them seriously.
"Reportedly, Google has recently started automatically opting users in to allow Gmail to access all private messages and attachments for training its AI models. This means your emails could be analyzed to improve Google’s AI assistants, like Smart Compose or AI-generated replies. Unless you decide to take action."
The #chatmail #rust core is the base infrastructure library used by all apps and bots. It contains all networking, encryption, email processing and implements all the relevant IETF email standards. It's a single central development place with which the whole ecosystem of clients and bots can upgrade consistently.
We are constantly trying to remove code, and to cleanup things. That's a must to keep a 8-year long evolved code base agile and adaptable. Key insight: The best code is no code ;)
two other things that can't be stressed enough for maintaining development long-term:
- Avoid superflous abstractions. Premature abstractions are the root of all ... complexity. If you have a second and third case, you can introduce an abstraction, but even then: don't over-abstract!
- write tests for everything, especially for the API that you want to guarantee to others. The chatmail core library just passed 1000 automated tests. The users of your API, and their users will thank you.
- Write tests with servers together with environment you roll 'em out. Id est, testing configs. - If your test suites don't run locally on dev machine (vms and illusion of networks - can do in 2025), if you need any cloud to test, you are not done.
But when your abstraction splits out part of the system, allowing richer life to it, while simplifying the rest, then entertain it. Example, past delta's post talked about goodness and richness that little apps bring, versus monolith-style addition of features in signal, recent announcements. This is a good abstraction, even though many would've walked from it. Food for thought. Short haiku/koan doesn't replace wisdom and elaboration, it only highlights it.
"The problem is, right now, talking to Copilot in Windows 11 is an exercise in pure frustration — a stark reminder that the reality of AI is nowhere close to the hype.
I spent a week with Copilot, asking it the same questions Microsoft has in its ads, and tried to get help with tasks I’d find useful. And time after time, Copilot got things wrong, made stuff up, and spoke to me like I was a child.
Copilot Vision scans what’s on your screen and tries to assist you with voice prompts. Invoking Copilot requires you to share your screen like you’re on a Teams call, by hitting okay Every. Single. Time. After it gets your permission, it’s excruciatingly slow to respond, and it addressed me by name every time I asked it anything. Like other AI assistants and LLMs, it’s here to please, even when it’s totally misguided."
Microsoft is advertising its Windows Copilot AI as “the computer you can talk to.” How does that hold up to testing, and how does it track with CEO Satya Nadella’s ambitions?
Two years ago, I wrote: Microsoft-the-company does not care about Infocom. But a lot of people in Microsoft must care. Microsoft is heavily populated by greying GenX nerds just like me. Folks who grew up with the first home computers and fondly ...
It's fascinating how my mood can shift from crap, to relatively OK, to quite depressed with suicidal thoughts within about a 1.5 minute span of time. I recommend it to everyone. /s
I'm always bothered by this kind of thing because on one hand, we obviously need to meet users where they are. On the other hand, apparently a lot of users only really know how to use the tab key and that's just such an efficiency brick wall! Trading off between those two things is very sad.
I agree. I also think some amount of learning curve is OK. Learning that you can (or should) navigate Tabs/toolbars with the Arrow keys instead of the Tab key feels like a low bar. Though I can’t back that up with any data.
Je pokrytí železnice mobilním datovým signálem nějaký principiální technický problém, anebo je to akorát nízko v seznamu priorit? Když teď České dráhy plánují nasazení Skylinku, je to technicky významně lepší než třeba běžné 5G?
Ja si to myslim, kdyz jsem to pred lety zkoumal, tak na mape pokryti se to kryje pouze s populaci a jeji hustotou. Zadna infrastruktura tam zanesena neni. Z toho usuzuju, ze pokryti 9x% lidi je zasadni (a mozna jedina) metrika v tom algoritmu planovani site.
We're sorry. The box making all the noises has experienced it's apparently now standard one month crash. There will be no noises or FX Radio for a while. The stupid idiot running this stuff didn't include a remote restart device, because he's a worthless piece of shit. That's right, a worthless, fucking, piece, of, shit! Get it? Got it? GOOD!
Wir lancieren heute Version 1.0 von barrieren-gutachten.de und veröffentlichen zeitgleich viele neue und hilfreiche Funktionen auch für Laien — sowie über 700 neue Prüfgutachten, die zeigen, wie der Staat seine eigenen Gesetze missachtet und Inklusion zu einem reinen Schauspiel verkommt. Das ist der nächste große Schritt für mehr digitale Barrierefreiheit und ein dringend nötiger Weckruf für die Politik.
TURN YOUTUBE into your most reliable source of qualified leads STOP treating YouTube like an afterthought! We use proven strategy and the ICE Framework – Impact, Core, and Explainer – to turn your expertise into discoverable, trust-building content …
Here's another French journalist participating in fearmongering about GrapheneOS. That article is not measured. It provided a platform to make both unsubstantiated and provably false claims about GrapheneOS while providing no opportunity to see and respond to those claims.
The claims the article platforms are conflating closed source products from European companies infringing on our copyright and trademarks with GrapheneOS. GrapheneOS doesn't have the features they claim it does, isn't distributed in the ways they claim and they don't understand open source software.
GrapheneOS is obtained from grapheneos.org/install/web and grapheneos.org/releases. There are a bunch of legitimate companies in Europe selling devices with real GrapheneOS including NitroKey. We aren't partnered with those companies and don't get funding from it but there's nothing shady about it.
Products using operating systems partially based on our code are not GrapheneOS. There's no such thing as a fake Snapchat app wiping the device in GrapheneOS. It has no remote management or remote wiping built into it. It does not have a subscription fee / licensing system built into it either.
Vast majority of the code for those products comes from elsewhere: Android Open Source Project, Linux kernel, Chromium, LLVM and other projects. Of course the non-profit open source project writing a small portion of the code being used by those companies being targeted rather than IBM, Google, etc.
Both Android and iOS try to defend users from the same attack vectors we do. We developed far better protections against exploits which we release as open source code. Open source means anyone can freely use it for any purpose, exactly like the Android Open Source Project used by GrapheneOS itself.
Open source is why we can build GrapheneOS based on the Android Open Source Project. It doesn't make Linus Torvalds, IBM, Google, etc. responsible for what we do. Similarly, others can make their own software based on GrapheneOS. A fork of GrapheneOS contains a small portion of code written by us.
France supposedly has a right to reply which we intend to exercise to respond at length to these articles containing libel from the French state.
We're going to be ending the small amount of operations we have in France as we don't feel the country is safe for open source privacy projects anymore.
GrapheneOS doesn't host services storing sensitive user data. We have signature verification and downgrade protection for updates to the OS, apps and app store metadata. We're going move our website and discussion server away from OVH. Our update mirrors and authoritative DNS are already elsewhere.
Our discussion forum, Matrix, Mastodon, etc. in OVH Bearharnois can be moved to local or colocated servers in Toronto instead. We can use Netcup (owned by Anexia, both German) as one of the main providers for website/network service instances. The majority of our servers are already not on OVH.
We won't travel to France including avoiding conferences and will avoid having people working in the country too. A simple heuristic for the EU is avoiding countries supporting Chat Control. We genuinely believe we cannot safely operate in France anymore as an open source project privacy project.
Our pinned post on this platform shows a great example of why they're actually upset with us:
It almost makes us willing to contribute to AOSP again to try to wipe out their ability to exploit a subset of non-GrapheneOS Android devices too. Google is welcome to reach out.
Attached: 1 image
Deux articles du Parisien hier, suivis aujourd'hui d'un article du Figaro, ont lancé une offensive honteuse contre GrapheneOS, un système d'exploitation open-source pour téléphones, gratuit et accessible à toutes et tous.
One thing I miss about old-timey "user unfriendly" linux was the way it wasn't bloated with every possible feature, tool and library that tried to cater for every hypothetical future user's potential needs.
Like, unpacking and building a tarball was annoying and time consuming, but it meant that you could install a PDF parser on a headless server without automatically pulling in a window manager, a print stack, several gigs of UI widgets.
Old days: "You will need 2MB of disk space, plus another 150KB for the compiled executable. It will take an hour to build, and you're gonna have to repeat the process if we release any updates that you feel you need"
Now: "You will need 15GB of disk space to install all the dependencies and their dependencies and their dependencies and their dependencies, and now you have to check for updates every 6 hours because we just added 1150 packages __that you don't even know what they do__ and they all have published CVEs but you haven't read any of them because nobody has that's too many CVEs to worry about, and if you try to remove any of them the entire stack auto-uninstalls to protect you from accidentally clicking 'print' on a subdependency that you didn't know was there and then it won't print"
Anyway, I'm just annoyed that today's must-fix patch is for 28 different packages all with "cupsd" in their name on a laptop that has never installed a printer and never will.
🎉 The wait is over! All the talks from DjangoCon US 2025 are now live on YouTube. From keynotes to lightning talks, you can now watch (or rewatch) all the incredible sessions from Chicago. Whether you missed the conference or want to revisit your favorite talks, head over to our YouTube channel.
Ein Email-Austausch mit einer Freundin bestärkt mich darin, meinen Ausspruch „Barrierefreiheit und PDF, ein Widerspruch in sich" auch weiterhin aktiv zu nutzen. Barrierefreie PDFs zu erstellen ist immer noch ein Hexenwerk, es gibt außer Adobe Acrobat anscheinend kein Tool, dass selbst in einfachen Dokumenten mit ordentlicher Struktur ordentliche Exporte hinbekommt, und beim Auswerten und Lesen der Tags über verschiedene Betriebssysteme ist der Zustand mit "katastrophal“ noch mild umschrieben.
@fasnix Ja, das stimmt im Prinzip, aber a) musst du es explizit einschalten, wenn du PDFs exportierst, was die meisten vermutlich übersehen werden, und b) übernimmt LibreOffice z.B. keine Alternativtexte für Grafiken aus aus Word importierten Dokumenten, und somit auch nicht in die getaggten PDFs. Ohne aufmerksame Handarbeit ist auch dies also eine echte Hürde. Dass dieses PDF sauber getaggt war, zeigt, dass sich da zum Glück jemand auskannte.
@oldperl @fasnix Eben nicht. Die einzigen, die wirklich PDFs testen, sind Funktionen im kostenpflichtigen Adobe Acrobat. Word und LibreOffice haben Testmöglichkeiten, ihre nativen Dokumente auf Barrierefreiheitsprobleme zu testen, haben aber beim PDF-Export selbst das eine oder andere Defizit. pdf.js kann wohl rudimentär inzwischen Tags verarbeiten, Vorschau auf Mac und iOS auch, aber sonst kenne ich keine Bibliotheken, die standardmäßig vernünftig getaggte PDFs erstellen.
@oldperl @fasnix Und da sind Tools wie die OCR-Software ABBYY FineReader noch nicht berücksichtigt, die Dokumente scannen und daraus eine Logik abzuleiten versuchen, aufgrund der sie dann Tags in PDFs reinschreiben. Das kann gut gehen, muss es aber nicht. Und als Entwickler kannst du Dir natürlich die PDF/UA-Spezifikation reinziehen, das bleibt Dir unbenommen. Ist aber ein ziemlicher Brocken.
Also braucht es eine Lib, und ein Tool, dass das kann. Ok...
Ich mach mir da mal einen Kopf dazu. Könntest Du mir eventuell, falls Du Zeit hast, zumindest ein paar keys, auf die man mindestens prüfen sollte, nennen?
@oldperl @fasnix Das resultierende Markup für Screen Reader ähnelt stark HTML. Also Absätze, Sprachauszeichnungen, Überschriften, Grafiken mit Alternativtexten, Listen, Tabellen mit entsprechenden Header-Zellen usw. Also viele "übliche Verdächtige“. Was aber die Darstellung für Sehbehinderte wie Kontraste usw. angeht, da bin ich überfragt.
* Für Word gibt es das Plugin axesWord, welches dich beim Erstellen der Tags aus Word unterstützt. * Für InDesign gibt es madeToTag. * Außerdem gibt es noch PDFlib für die programmierte Ausgabe von getaggten PDFs. * Die Umsetzung aus LaTeX wird noch etwas brauchen, aber man ist eifrig bei der Sache.
Wie immer bei der Barrierefreiheit sind mehrere Beteiligte am Kochen, was das Ergebnis zu Experiment macht.
Der Autor/die Autorin, das verarbeitende System mit Exportfunktion, der PDF Reader, der Screenreader und das zugrunde liegende Betriebssystem.
Nur wenn alle ihren Job machen, ist das Gesamtergebnis zufriedenstellend bis gut.
Viele Schreibende verwenden die Formatvorlagen nicht, da kann der beste Export nichts machen – außer raten. Oder beim Export werden Inhalte, die über einen Seitenrand hinaus gehen, nicht zusammengefasst. Oder der PDF Reader gibt nicht alles an die Schnittstelle des Betriebssystems weiter. Oder der Screenreader ignoriert technische Normen.
Viel Potential für Fehler – wie bei der Erstellung von Internetseiten.
@dnkrupinski Das ist mir durchaus alles klar. Im Gegensatz zu HTML-Seiten, wo meist irgendwie noch irgend etwas geht, ist aber bei PDFs oft genug wirklich gar nichts lesbar. Oder der Textfluss von nicht getaggten PDFs ist so grundlegend kaputt, dass es keinerlei Logik gibt. Mach z.B. mal eine Rechnung von der Telekom auf, egal ob Mobil oder Festnetz. Haarsträubendes Chaos.
Mir ist das durchaus bewusst. Aber wenn du Schei…benkleister reinsteckst, kommt auch Scheibenkleister raus.
Auch wenn HTML "irgendwie" geht, kommst du oft ohne JavaScript gar nicht weiter – und was da an <DIV>-Dünnpfiff herauskommt will ich nicht weiter kommentieren.
Auch hier das Zusammenspiel: Redakteur:in / CMS / Betriebsystem / Screenreader.
Das Format PDF ist nicht schlechter oder besser als HTML (mit der Ausnahme, dass es nur wenige gibt, die PDF im Quellcode lesen können). Und es sind verschiedene Ausgangspunkte. Doch wenn beide entsprechenden den Normen und Standards erstellt werden, ist das Ergebnis zumindest vergleichbar, wenn auch nicht gleichwertig.
Ja, ich wünschte mir auch ein Produkt, welches kostengünstiger und absturzfreier als der Adobe Acrobat Pro funktionieren würde. Aber **noch** gibt es das nicht – auch wenn der Foxit PDF Editor sich in die richtige Richtung entwickelt.
Wenn Inhalte als (semantisch getaggtes) HTML vorliegen, kann man Prince verwenden, um daraus accessible PDFs zu machen: princexml.com/doc/prince-outpu…
Prince ist frei für Non-Commercial-Use. Für Commercial Use oder wenn mans als SaaS verwenden möchte zB mal auf europdf.eu schauen (Disclaimer: In dem Projekt steck ich mit drin.)
Making accessible tagged PDFs with Prince Disclosure: I’ve been doing some paid consultancy for YesLogic (Prince’s vendors) on the accessibility of Prince’s PDFs. Love them or hate them, PDFs …
@fasnix Wir haben das jetzt noch weiter getestet und Libre Office übernimmt die Alttexte doch. Die Datei, die wir getestet haben, hatte nur die von mir übersehenen Grafiken in der Kopfzeile, die im Acrobat mit Screenreader nicht angezeigt werden. Pages hat die Kopfzeile anscheinend so getaggt, dass sie nicht als solche erkennbar war vom SR, sondern als normaler Seitentext. (1/2)
Gemini AI needs to be disabled on Android or it will override your privacy settings and gain full access to your texts, calls, and WhatsApp - even if you’ve turned off Gemini Apps Activity on your phone.
I haven't used Gmail for months so I don't even care anymore. I deleted the Gmail app by using Canta app. I'm not gonna install it again on my Android just to change some settings
I am a bit lost. Are you actually saying that Gemini could access third parties apps, like Whatsapp and things like that? What about the encryption, then? Sorry, I am not very smart in this field, just trying to learn!
To be fair, that setting exists already for quite a while (and I've turned it off ages ago on my gmail account). Good to remind people to check once in a while whether it's still turned off, but it's not like they did this yesterday.
Here's the post from january 2025 about exactly this interface.
all my critical informations are stored with @Tutanota. All what remains in my gmail is basically junk I can lose without blinking. It's money well spent.
Wow, that’s concerning. Thanks for the warning. Most people won’t even notice Gemini was switched on automatically. Really appreciate you breaking this down and showing how to disable it. Sharing this! ❤️
Will share the link with family who need to hear it from "someone else".
Had to go Proton for all bc every time I attempted to sign up w Tuta for a account (paid annually after sign-up) I was flagged as potential abuser. With VPN. W/O VPN. Multiple devices, IPs, and browsers over a two-month period. 🤷🏻
That said, I still appreciate what you do. Disappointing to not have been able to sign up and pay. 💰 Maybe fix that. -Random midwestern parent who likes privacy and technology 👍
Thanks for reminding. I am afraid these toggles are just make believe. They have your data on their servers and already used it for Gemini crap behind the scenes. And none of the alternatives is safe as long as billions of people continue to use gmail blindly. Every mail you sent/receive to/from gmail is a door wide open. No other option then to block all traffic from gmail accounts.
I checked my settings today as I had looked to make sureit was off a while ago. Still off, so they aren't switching it on for people already opted out.
I never asked and said NO to unsolicited AI garbage I don't want to use on my electronic devices. American big tech companies have become techno-fascist virus nowadays, so I moved away from Google services and switched to @fdroidorg . Early next year I'll leave Android for the privacy-focused /e/OS mobile system to leave in peace. Keep fighting for privacy and freedom everyone!
Leider ist es nicht die einzige Sache, die Google mit uns tut, ohne unsere Zustimmung. Daher kann man alles von solchen Unternehmen erwarten. Ihr rebranding vom Logo, ist wohl nun die Begründung, alles mit unseren Daten zu tun, wie es ihnen gut kommt.
Last month it was the Young Republicans praising Hitler. Last week it was Fuentes & Carlson criticizing "Jewry". This week, they're saying swastikas and nooses aren't hate speech. The rot is party-wide. seattletimes.com/nation-world/…
Obedeciendo las instrucciones y órdenes profesionales de los superiores, salvo que constituyan una infracción manifiesta del ordenamiento jurídico, en cuyo caso las pondré inmediatamente en conocimiento de los órganos de inspección procedentes.
![[On the bridge of Enterprise]
Picard: Yellow alert
Everyone: OW ALERT!!](https://fedi.ml/photo/preview/640/728017 "[On the bridge of Enterprise]
Picard: Yellow alert
Everyone: OW ALERT!!")
BlueLegend
in reply to Matt Campbell • • •Matt Campbell
in reply to BlueLegend • • •@threlm4280 Yes, there are probably people who hate AI for the reason you described. I want to push back on this point though:
> I don't get it. Its literally convenience why not use it?
There are in fact better criticisms of current AI products. They're trained on basically all of the work of all humanity (or as much as the AI companies could get their hands on), without the consent of the people who did the work, for the explicit purpose of replacing labor. Worth thinking about.
BlueLegend
in reply to Matt Campbell • • •Devin Prater :blind:
in reply to Matt Campbell • • •