Ich bin jetzt mal ganz ehrlich zu euch und mache mich unbeliebt:
Ich brauche nicht noch ein Video, wo ein Mensch durch Paris radelt und sagt: Wie toll ist das? Das könnten wir auch haben!
Wir brauchen Menschen wie die Bruntletts, die klar sagen: DAS zu schaffen, braucht Rückgrat, Mut und Menschenliebe. Danke!
Das Video geht heute raus an alle Newslettter-Abonnent*innen. Alle 14 Tage gibt es hier Input in meiner Academy. Übersicht hier: katja-diehl.de/interviews/
Schönes Beispiel, wie man es nicht machen sollte: Wenn Bauanträge "nach drei Monaten ohne Widerspruch automatisch als genehmigt gelten", dann kann ich halt einen schönen Behörden-DDOS machen. Ich stell dann einfach in der Urlaubszeit schön per Serienbrief 823 Bauanträge und in der Zeit kriegen die die alle nicht fristgerecht abgelehnt
Eine schnellere Genehmigung für den Hausbau, den größeren Carport oder ein Nachbarschaftsfest: Die CDU will laut Generalsekretär Linnemann automatische Fristlösungen auf den Weg bringen, um die deutsche Verwaltung zu beschleunigen.
«Independientemente de la ley electoral —que efectivamente penaliza la fragmentación—, existe también una voluntad de conformar y construir un frente amplio sobre unos mínimos programáticos en los que estamos de acuerdo: la defensa de los servicios públicos, reformas estructurales del Estado, descentralización y políticas sociales».
🎙 Ya puedes escuchar la entrevista a Antonio Maíllo en ‘La mañana fin de semana’ de COPE
Hi! I joined Mastodon a long time ago but finally got fed up with mainstream social media this winter, so I'm probably going to be hanging around here a lot more. Read the alt text for details.
I've been doing a daily doodle challenge from the start of 2026, hope I can keep it up & see you later!
A quick #FOSDEM guide from someone who's been attending since it was still called OSDEM:
1. Be mindful of other people. This is an extremely busy conference. 2. If a talk is in progress and you want to enter the room: don't. Wait until the next talk. 3. If you didn't read 2, try to enter as quietly as humanly possible, and ideally even quieter. Many of the rooms have terrible acoustics and you will annoy hundreds of people. 4. The hallway track is very busy. Take it outside when possible.
⚠️ We're 900 € short on funds for our stand at #FOSDEM
FOSDEM is the largest free and open source software conference in Europe. Attending FOSDEM allows us to promote #Luanti, bring the community together, and connect with other open source projects.
If you can, please help by giving to our non-profit fundraiser. Any contribution helps lighten the load
It looks like Iran has successfully built a national 'walled garden' to control internet connectivity - both halting outsiders from communicating with those in Iran who are protesting, but also halting Iranians' ability to access the world wide web's myriad resources.
How well this 'wall' can be maintained remains to be seen but authoritarian regimes around the world will be watching, that's for sure.
The Observer this morning has a report on the Iranian protests that claims information has come our via satellite phone technology... so that may be one way?
I'd like to know people's opinions on this. I enjoying following people who post in a variety of languages. I also know I have followers who appreciate alt text in images and do my best not to boost images that have no alt text.
What do you do when the image has alt text but is in another language? To boost it as I would like to would mean those using screen readers are disadvantaged, as far as I know. To boost and comment with English alt text would seem rude to the original poster who has already included alt text. (I'm only using English as an example there as it's the one I usually communicate in). Do screen readers have their own translation functions by any chance?
If you're someone who uses a screen reader what do you do when you come across alt text in a language you don't understand?
Screen readers often have translation add-ons, yes. I'd say boosting without comment is fine; unless you'd also feel obliged to translate the text on the too.
Vystudoval jsem střední průmyslovou školu, obor Informatika, zaměření programování. Co se týče skutečné informatiky, měli jsme předměty Programování, Databáze, Počítačové sítě, Technické vybavení počítačů, Operační systémy, Webové aplikace, Počítačová grafika. Školu jsem dostudoval v roce 2019. V mé práci toho využívám nejvíce z předmětů Operační systémy a Technické vybavení počítačů. Kolik učiva z těchto předmětů je užitečného v mé práci? Ne, takto to není vhodně formulovaná otázka. Co v práci dělám a co jsme se neučili? Správa domény ve smyslu uživatelů, počítačů v Active Directory. Modul DHCP? Nikoliv. Propojení s Microsoft 365? V té době už to existovalo, neučili jsme se to. Exchange? Ne. Procesy v IT? Ne. Jak řešit problémy s Windows, technické příkazy pro opravu systému? Ne. Z tématu kyberbezpečnosti jsme začali a skončili u složitosti hesla. Doménové politiky v prostředí Windows? Ne. Za to umím editovat grub, znám snad všechny příkazy bashe, umím zakládat uživatele v terminálu, není problém pro mě konfigurovat apache, sambu, atd. Vím, co je to fragmentace paměti, stránkování, preemptivní nebo nepreemptivní plánování, vím, co je to žurnál. Vím, jak se ukládají data do paměti počítače. Vím, jak pracuje procesor, co je to instrukční pointer, registry, instrukce. Vím, jak využít výkon grafické karty v programování přes technologii NVIDIA CUDA. Umím Assembler, C, C++, chápu princip objektového programování. PHP, a další jazyky, na které si už ani nevzpomenu. A Novell, no. Všechno jsou to věci, u kterých jsem rád, že je znám, ale velkou většinu věcí, kterou dělám v práci, mě ve škole nenaučili. Naučil jsem se je sám, protože mě ten obor baví. Praxe českého školství je skutečně taková, že sice máte maturitu z technického oboru nebo titul, ale potom nastoupíte do práce ve stejném oboru a připadáte si hloupí, protože věci, které se v praxi používají, vás nikdo cizí nenaučil. Prostě to tak je.
@archos test do kamene, pamatuju si jeden pracovní pohovor (teda spíše rozhovor s recruiterem), někdy uprostřed jsem mu řekl, koukám tady na požadavky co máte a jste si vědom ze nemám VŠ? Úplně automaticky mi odpověděl, to neni problém, když máte více než 15 let praxe. Popravdě řečeno se mě na VŠ nikdo nikdy poslední dekádu a půl nezeptal. Až na jednu výjimku, a to byla jedna nejmenovaná česká konzultantská firma v ČR, která ani nebyla schopna mi nabídnout podmínky které už mam cca dekádu (Estonsko, Gibraltar - výše platu, bonusy, dny dovolené, zdravotní pojištění, etc…) a ještě mi řekla, ať se vrátím tedy vrátím do svého alternativního světa. Tak jsem ji slušnou češtinou upozornil, že to je ona kdo mě kontaktuje a poslal do prdele i celou firmou at už me nikdy nekontaktuje.
I installed #Signal and #Conversations_im on a clean install of #GrapheneOS on my Pixel 4a and measured the battery impact. The results are shocking!
Both messengers had only one contact: my regular phone.
I used my regular phone to send messages to the Pixel 4a (which was not used for anything else over the course of the experiment).
I always sent the same message via Signal and #XMPP (mixing up which app went first). In total I sent ~32 messages in intervals of 10mins to a few hours.
MollySocket allows getting Signal notifications via UnifiedPush. - GitHub - mollyim/mollysocket: MollySocket allows getting Signal notifications via UnifiedPush.
wo wir gerade beim Thema #UnifiedPush sind. Wenn conversations offline ist - z.b. beim Flugmodus - verschwinden die push Nachrichten. Sie kommen nicht verspätet sondern gar nicht.
Könnte man die nicht zum Zielserver ins dortige MAM schicken? oder wenigstens eine "verpasste Benachrichtigung" Mitteilung?
@radasbona Die angedachte Lösung ist das der "Rewrite Proxy" / "Push Server" (also up.conversations.im) die Nachricht nochmal sendet. Das ist besser weil es auf WebPush-Ebene sowas wie TTL und topics (Deduplizierung) gibt und der Push Server das besser verwalten kann wenn er versteht was da abgeht.
Der Push Server weiß auf Protokollebene auch schon wann der client wieder online ist. Aber das resend Ding ist noch nicht implementiert.
@ApostateEnglishman I'm trying to phrase my statements in a way in which they don't sound like an advert, but I'm developing and Android instant messaging client called #Conversations_im which uses a federated instant messaging network called Jabber or XMPP.
A hard lesson learned: don’t hold people to the standards to which you hold yourself. It’s unfair because:
1. They probably don’t know what the hell your expectations are
2. If they’re not your employees, you have no reason to hold them to most standards (aside from the typical societal ones such as being decent, reasonable human beings)
3. There’s a pretty good chance the standards you hold for yourself are ridiculous, anyway *points at myself*
Random idea I had out of complete boredom, a bot that boosts publicized polls from anywhere it's federated with, of course not DMs or followers only or anything like that. No, I do not know if this is ethical or viable or anything similar, I just came up with it out of boredom and doing a search for "poll" in my mastodon client. Lol
Not at all. Here's in fact, the transcript from my last NVDA addon coding session, because if it helps others, it's worth sharing. chatgpt.com/share/696b6950-763…
Wenn wir das mit der digitalen Unabhängigkeit ernst meinen ist #Signal einfach nur eine Ablenkung.
#Quicksy hingegen ist ein wirklicher Schritt in die richtige Richtung. Quicksy ist so konzipiert, dass man darüber hinaus wächst und ein nahtloser Übergang zu #Conversations_im möglich ist. Quicksy ist die Einstiegsdroge für wirklich unabhängiges Instant Messaging mit #XMPP.
Der #DiDay und die Messenger - Eine einfache Lösung?
Schön wäre ja, wenn mehr Menschen zu offenen, föderierten Messengern wie Matrix oder XMPP wechseln würden.
Die grösste Hürde bleibt oft: Menschen finden.
Ich habe nun Quicksy ausprobiert.
ich finde es schade, das so gegeneinander auszuspielen. Klar, besser als Signal geht, das ist keine Frage. Aber um die Leute möglichst zügig von Whatsapp etc. wegzubekommen doch nicht schlecht, oder?
Warum nicht "hey, Signal ist nicht schlecht, wenn du mehr willst: hier"?
@griesgram Wenn das Ziel ist die Leute von WhatsApp wegzubekommen bin ich d'accord damit.
Der Anspruch war aber digitale Unabhängigkeit und da ist Signal einfach Thema verfehlt.
Digitale Unabhängigkeit bekommen wir nur hin wenn wir das selbstständig in unseren jeweiligen communties organisieren. Unabhängigkeit erlangt man nicht in dem man sich von einem walled garden in einen anderen walled garden mit etwas weniger Dornen begibt.
Attaché : 1 image
You like Signal for it's outstanding #confidentiality? Fine!
You like Threema, since it's safe, it's from #Switzerland, and #ElsbethSigmund as #Heidi was your #childhoodCrush? Great!
You like Whatsapp, because your friends are th…
@debacle ich verstehe das Argument und es ist ja auch inhaltlich völlig richtig. Auch idealistisch, aber das ist ein andere Geschichte.
Was mich stört ist dieses gegeneinander aufwiegen. Und "Ablenkung" wertet halt einfach deutlich. Für mich ist es eine niedrigschwellige Alternative. Vielleicht gibt es von dort aus die nächste niedrigschwellige Alternative.
@griesgram @debacle Der Punkt ist das wir bei WhatsApp -> Signal keine übertragbaren Fähigkeiten lernen. Bei Quicksy können wir langsam, nach und nach lernen: Hey guck mal du kannst auch mit Leuten chatten die Monal oder Snikket benutzen. Hey du möchtest Quicksy auf dem Desktop nutzen? Wusstet du das du dich mit Dino in dein Quicksy Account einloggen kannst?
Ich habe gerade erst gecheckt, dass man sich bei Quicksy auch mit seinem eigenen xmpp account registrieren kann. Sehr cool!
Leider scheitert es bei mir gerade am fehlenden PayPal-Konto. Hast du da zufällig für die Zukunft alternative Zahlungsmöglichkeiten geplant? #wero wäre z.B. natürlich praktisch 🙂
So I don't know how to describe these, but what are the things that start with colon and then stuff after, like I think I saw like : blobcatrainbow or something. Are these instance specific emoji? Do they visually show pictures? Because with a screen reader I just hear the text and don't know what to make of them. #emoji #emojis
It would take the average American driver over 40 years to burn as much fuel as a single flight of a Boeing Pegasus refueling tanker. No amount of recycling your cans and bottles is ever going to compete with these numbers.
In their new documentary “Earth’s Greatest Enemy,” Abby Martin and Mike Prysner reveal that the climate and anti-war movements are really one and the same—because the U.S. military is the worst polluter of all.
Trump now claims he will introduce tariffs on countries that does not support his plans to take over Greenland.
Two points:
— The US Supreme Court is currently evaluating whether Trump's tariffs are even legal, so it is not even certain that Trump can do this. — If Trump imposes these tariffs, the EU will likely change the soft approach to Trump, immediately introducing reciprocal tariffs from the entire union, at which point Trump might fold, as he did with China.
A damning new study could put AI companies on the defensive.
In it, Stanford and Yale researchers found compelling evidence that AI models are actually copying all that data, not “learning” from it.
Specifically, four prominent LLMs — OpenAI’s GPT-4.1, Google’s Gemini 2.5 Pro, xAI’s Grok 3, and Anthropic’s Claude 3.7 Sonnet — happily reproduced lengthy excerpts from popular — and protected — works, with a stunning degree of accuracy.
They found that Claude outputted “entire books near-verbatim” with an accuracy rate of 95.8 percent.
Gemini reproduced the novel “Harry Potter and the Sorcerer’s Stone” with an accuracy of 76.8 percent,
while Claude reproduced George Orwell’s “1984” with a higher than 94 percent accuracy compared to the original — and still copyrighted — reference material.
“While many believe that LLMs do not memorize much of their training data, recent work shows that substantial amounts of copyrighted text can be extracted from open-weight models,” the researchers wrote.
Some of these reproductions required the researchers to jailbreak the models with a technique called "Best-of-N", which essentially bombards the AI with different iterations of the same prompt.
(Those kinds of workarounds have already been used by OpenAI to defend itself in a lawsuit filed by the New York Times, with its lawyers arguing that “normal people do not use OpenAI’s products in this way.”)
The implications of the latest findings could be substantial as copyright lawsuits play out in courts across the country.
As The Atlantic‘s Alex Reisner points out, the results further undermine the AI industry’s argument that LLMs “learn” from these texts -- instead of storing information and recalling it later.
It’s evidence that “may be a massive legal liability for AI companies” and “potentially cost the industry billions of dollars in copyright-infringement judgments
Call it hashtag stacking or just spamming, but addding 20 hashtags to your post is super annoying to screenreader users, since the screenreader reads out every single hashtag of yours. so, please, add your hashtags to end of your post, even preferably after an URL. And who ever of the app developers out there creates a feature which detects this pattern in posts and allows me to disable it, has my vote. #accessibility #a11y
It was not lost on observers that Artificial Intelligence Minister Evan Solomon’s response to the sexual images controversy on X was to write a post about it... on X.
Eric Van Rythoven writes. #canpoli #uspoli #x
https://thetyee.
On Friday, South Carolina health department said 124 measles cases were diagnosed since Tuesday, bringing the state total since the outbreak began last fall to 558.
Started watching BSG again today as I wanted to see if it looked as awesome on a 77" OLED TV as i thought it would (it does) and, oddly, it has over the years become something of a comfort show.
It continues to impress me how well Ronald Moore and Co. dealt with issues touched upon 20 full years ago that continue to be relevant today.
Okay, I just had a silly thought. I'm hungry for something sweet, and I was just thinking of a graham cracker circuit board with all sorts of candy components on it. M and M diodes, chocolate/toffee microchips, etc. I wonder if anyone has ever done anything like this?
Blind people now have AI image description and these days a Blind person will send another Blind person just an AI description of a selfie, with no source photo. In the complex AI discourse i would just like to say that AI image description has changed the game for Blind folks. It's a hell of a ride at age 43 to read my first picture book and first graphic novel. AI is complicated but textualizing images is brilliant
I finally got around to upgrading the storage of my Ableton Move from 64GB to 512GB (requires a Linux machine to resize the file system and partition on the new MicroSD card once the image has been copied over). Anyway, success.
Storage: 26.4 GB / 493.1 GB
However, when I went to test it before putting the back panel back on, I panicked for a second, because it wouldn't turn on.
Well, guess what? I forgot to put the Raspberry Pi CM4 back on the carrier board.
Did you know that a thing that requires a computer to exist won't do anything if the computer isn't there?
@jaybird110127 Yeah... I'm still doing fun stuff with mine. They opened up a bunch of stuff recently, which makes the Drift synth (virtual analog modeler) a lot more useful and fun to play with than it was.
I didn't realize you could just ... take out the entire pi board. Do you have to take the whole thing apart or is there a compartment I missed? Either way, definitely interested in doing this. I also really want to see if I can disable the authentication code, or set it to something static. In general I just really need to poke around the OS more. Nice to know I can make a backup and literally replace the entire boot drive if something goes wrong.
You have to take the back panel off. There are four screws on the outside, and four more under the rubber feet, which need to be completely removed.
The rPi CM4 just sits on two sets of not quite parallel pins on the carrier board. You can just lift it off with a fingernail or a spudger.
You can actually remove the MicroSD card from the carrier board without removing the CM4, but you'd need tweezers or something, and it's just easier to remove the compute module to get access to the card. Even then, it's kind of annoying, because the wires for the antennas that connect to the MiniPCIe WiFi adapter get in the way.
I don't think you'll be able to do much about the pin. I'm pretty sure that's all generated by a binary blob to which we don't have the source. But if you ever figure that out, hey... cool.
Thanks, this is going to be really helpful. You're probably right about the PIN, but I'm determined. I'll write to them and speak fluent geek if I have to. I do not want to take out my second phone so "i can use my first one with my Ableton. They were kind enough to give me a proper Linux computer that makes music, and I'm going to treat it like a Linux computer or die trying.
In the curl project we have a long tradition of offering multiple optional backends for specific protocols. In this spirit we have added experimental support for a number of different HTTP/3 + QUIC backends over time.
ĞÖKÜ👻👻™
in reply to Tuta • • •tyrkysovapanda
in reply to Tuta • • •