Thinking of Aaron Swartz today & I’m stuck on this photo - he & OpenAI CEO Sam Altman (both circled) each scraped 1000s of docs but one did it to make the knowledge free for all while the other did it to make $$$$ through probabilistic plagiarism. The US DOJ only came after one of them & the other is feted by tech bros and executives.
Thank you Aaron for so much, for RSS, for Markdown, for Creative Commons and more. I’m sorry our society failed you.
So let me get this straight: First, it was Elon. Now, it is Mark. Two guys are removing all moderation rules on social media. And are people asking another billionaire to build alternative social media? Do you even hear yourself? That is not going to fix anything. Maybe we should go back to the days of blogs, mailing lists, individual forums, and RSS feeds and end this madness of centralized social media apart from the federation. I’m just saying
První letošní Vídeňák. 😋 (Jako příloha kančí řízečky.) #zasezerou
Tenhle salát se u našich dělal odjakživa. Jako jo, byl dobrej. Ale když jsem v hluboké dospělosti našel ten správný recept, teprve teď je to ono. Kde se sakra předkům mohl vytratit ten hovězí vývar? 🤔 A vy si nechte chutnat, ať máte cokoliv. 🤞
@archos To mě překvapuje, bavorská varianta je tam celkem rozšířená, jak jsem viděl. Jen tam přidávají okurku, to už není ono. 😊
Kančí můžu jen doporučit. Míváme výhradně vyzrálé selátko nebo lončáka, stačí s tím zacházet jako s normálním vepřovým. Rozdíl v hubě je ovšem zásadní.
As a recipient, is it meanwhile possible to get a payout without needing an account with Peter Thiel (Paypal, Stripe) – and regardless of how the donors submitted their donations?
The lack of that is what keeps several people/organizations from using Liberapay (or having their accounts with you dormant, like IzzyOnDroid, still hoping for that possibility).
ich bin aktuell am überlegen, meine lokale Instanz auf einen kleinen VPS umzuziehen. Deswegen bringe ich mich momentan auf den aktuellen Stand der Dinge, was die Installation eines LAMP-Servers und die Installation von Friendica betrifft. Beim rumstöbern bin ich auf eine aktuelle Anleitung gestoßen, wie man Friendica installiert.
Dazu hätte ich zwei Fragen:
Unterscheidet sich die Installation auf einem Raspi zu einer auf einem VPS?
Würdet ihr die Installationsanleitung so durchführen oder würdet ihr was anders machen?
Was würdet ihr generell am System optimieren bzgl. Friendica und dem System selber?
@OldKid ⁂ @Unity :Friendica: ⁂ Ist ja teilweise sehr ausführlich dargestellt. certbot als snap, PHP als Modul, ufw macht lediglich drei Ports auf, fail2ban in der Standardconfig und ein "AllowOverride All" in DocRoot.
Ich denke, da wäre noch ein bisserl Luft nach oben.
Für mich ist bei sowas immer die Frage: "Wie hoch ist der Pflegeaufwand nach der Installation?" Man muß sein Zeug ja aktuell halten. Und da sehe ich nur zwei Wege:
Ein Paket, das über den Paketmanager des Systems akuell gehalten wird. Mir ist nicht bekannt, daß es so etwas für Friendica gibt
Oder eben Docker in Verbindung mit so etwas wie watchtower - da muß man dann wortwörtlich gar nichts mehr tun
In der Docker-Variante muß man nicht einmal das machen. Und man muß auch nicht wissen, ob und wann eine neue Version erschienen ist. Man kann ganz einfach sterben und die Instanz läuft immer aktuell weiter.
TIL that on #Meta platforms now one can write terrible, offensive slurs against LGBTQIA+ folks and immigrants… with zero repercussions. But the second one mentions a #Fediverse project like #Pixelfed, the post is automatically removed.
Meta “moderation” policies in 2025: racism, xenophobia, transphobia, homophobia and sexism are allowed; mention of competition is strictly forbidden.
New: Meta employees are furious about a new policy that allows users to call LGBTQ+ people "mentally ill." Some have resigned in protest. Some LGBTQ+ employees have started saying they if they are 'mentally ill' they should get leave of absences from work www.404media.co/its-total-ch...
One of the most important things I consider when choosing a social media platform is whether people can read what I post without being nagged or forced to create an account. For many of my friends who don’t really care about #Mastodon, gultsch.social is just a website they can drop by occasionally to see what I’m up to.
Fix rare crashes on app startup
Don't lock up app on close
Improve internal logging
Fixed background crash
Revert christmas/winter special
Fix crash on incoming non-spec-conformant LMC
Make xmpp ga...
We expect that some 50 Auschwitz Survivors will come to the Memorial to commemorate the 80th anniversary of the liberation, which will be held in front of the main gate of the former Auschwitz II-Birkenau camp.
On January 27, the eyes and ears of the world should rest upon them.
We must safeguard their voices from silence, distortion, instrumentalization, or being overshadowed. Their message should help us shape the memory of the human tragedy of #Auschwitz.
The Story of Gunsmoke is essential listening for classic radio enthusiasts just starting to listen to the series and for veteran hobbyists. It was produced by...
Barnes and Noble is the largest bookselling chain in the US and has announced that it will open 60 new bookstores in 2025. The bookseller opened up 57 stores
using the USB-C capture card into the Be My Eyes app under windows! Truly endless possibilities listen to audio below.  to grab the capture card that I have off of Amazon use Papeaso Video Capture…
Comcast’s planned spin-off of its cable television networks is gaining momentum as Mark Lazarus, the prospective CEO of the new entity dubbed “SpinCo,” announced key appointments to its senior leadership team.
@Rainer "friendica" Sokoll Ich habe heute fulltime 6 Stunden Fehler gesucht. Ich bin kurz vor dem Aufgeben. Die docker-compose.yaml funktioniert jetzt, dafür habe ich einen neuen Fehler und der Container startet nicht.
Error: crun: error stat'ing file `/var/lib/containers/storage/volumes/friendica_friendica_friendica/_data`: No such file or directory: OCI runtime attempted to invoke a command that was not found
exit code: 127
Extraordinary article on the energy usage of generative AI from BloombergNEF founder Michael Liebreich - absolutely worth spending some time with this: about.bnef.com/blog/liebreich-…
This year will go down in history as the year the energy sector woke up to AI. This is also the year AI woke up to energy. Is the data center power frenzy just the latest of a long line of energy sector bubbles, or is it the dawning of a new normal?
> In 2007 the EPA predicted data center energy usage would double: it didn't, thanks to efficiency gains from better servers and the shift from in-house to cloud hosting.
That latter shift can also be seen as a shift to greater centralization, which isn't necessarily a good thing. I hope centralization, a concentration of hosting onto big servers at a few providers, isn't necessary to keep energy consumption at a reasonable level.
@matt on a similar note, for all of the excitement about local LLMs not a lot of people seem to be considering how inefficient they are in terms of power usage compared to big power hungry data center LLMs that are shared by thousands of people
@matt It's a few things. Large cloud providers have done a lot of work to reduce inefficiencies in their ecosystems, including moving to large-scale liquid cooling on some cases, whole rack-aisle airflow calculations, custom chassis designed for power efficiency, etc., meaning that in many cases, less than ten percent of the total DC power is going to things other than compute, storage, and networking. They also have enough load to balance that they can keep most of their machines running at right around the 80% load mark, because systems are more efficient when they're running a heavy workload.
Very few companies with small on-prem DC footprints do anything like this much work, and the systems that are sold to them are generic, because they need to work anywhere, so that kind of extreme optimization just isn't available to them. Loads for individual companies are also often bursty, given peaks in customer demand, monthly batch jobs, etc., and the rest of the time that spare capacity sits idle. Cold-booting hardware carries some risk of systems not coming back online, so servers sit mostly idle, still consuming most of their peak power.
So basically yes, on-prem is hard to make as efficient as cloud systems, for most realistic companies. However, one of the things you can do is get rid of useless compute. Taking the time to work with profiling tools and taking some care at the business level in how and when jobs are run can have a large impact on performance. Getting rid of needlessly complex systems also helps — for instance, the entire targeted advertising ecosystem, if turned off and all related code removed from systems, would likely cover efficiency losses from returning to on-prem systems many times over. @simon
@dymaxion So, what would happen if we had a lot more small-scale self-hosting on passively cooled systems with low idle power consumption? For example, I have a single-board computer (a Quartz64 model B) sitting in my apartment, assigned to a static IP address on my residential fiber connection so it can host stuff. Its idle power consumption is probably lower than a typical x86-64 PC, though I haven't measured it.
@dymaxion Still, I wonder, if hundreds of millions of people had such self-hosted setups, would the aggregate energy consumption still add up to much more than the equivalent centralized services?
@matt Yes, basically, especially with full lifecycle analysis. We can federalize services politically and take advantage of industrial efficiencies of scale without trying to be digital homesteaders, and most of our options for efficiency come from doing less dumb useless stuff, like data collection, behavioral analysis, personalized advertising, and the entire suite of useless LLM tools. @simon
@matt So, a) if we're talking about large models, they cannot be trained in that context and many of them won't even be able to do inference there. b) For a phone, the embodied carbon (ignoring all other resource impact) of manufacturing is larger than its lifetime energy carbon footprint. c) A machine like that serving a single user is almost always going to be idle, and even if it has significant idle power reduction, it's still going to do much lower useful work per watt than a more computationally capable per watt large system that's running at a steady 75-95% load. @simon
Recently the Dutch Electoral Board (where I am also a very part time advisor) invited me to do a talk reflecting on their open source Abacus vote tabulation software.
After our archive was updated to 12.9, our Debian CD team did the hard work of preparing and testing our installation medias to make sure everything is working correctly. Now their work is complete and the images are available at debian.org/distrib/
the EU does so much fucking right. I'd kill for a government with one tenth of the balls of the EU. Instead we get feckless democrats and Putin worshipping cucks.
Having just been reminded of this blog post by @glyph: blog.glyph.im/2012/10/the-lexi… I decided to repeat something that a friend tried when I showed him the post a year or so ago. I gave Claude 3.5 Sonnet this prompt:
Based on the attached blog post, create a quiz to score whether someone is more of a nerd or a geek. Create it as a self-contained HTML artifact using plain JavaScript.
The questions Claude came up with were OK, but what's more interesting to me is that it used clickable divs for the answer choices rather than real radio buttons. So remember, if using an LLM to write code, it's been trained on all of the Internet, good, bad, and ugly, and you have to watch out for issues like this one. And this issue affects accessibility.
To be clear, I never intended to publish the quiz, particularly since the input was someone else's post and my contribution was very minimal. I just wanted to see how an LLM would handle that task.
Claude used React to render a list of suggested products in a category I asked about recently. As in, it coded a React app to render the list, all so it could plop it into its artefact preview area rather than just... sending back some text. So I can't say I'm surprised.
@jscholes I haven't actually checked the code yet; I just completed the quiz using NVDA. I explicitly instructed it to use plain JS, but I guess it could have still used React. IIRC though, if it had used React, NVDA would have labelled the whole thing as clickable because of React's event delegation, not the individual answers.
@jscholes It's something I've noticed before; the whole root element is marked as clickable. For that reason, I see framework-default use of event delegation as a step backward for accessibility, though a minor one; the "clickable" indication is a fallback after all, a weak substitute for a proper role. Still, sometimes a screen reader user needs every weak fallback they can get.
Turns out there are more accessibility issues with this web app Claude generated. The progress bar (showing how much of the quiz one has completed) and the score indicator are totally inaccessible.
The progress bar was implemented using divs and CSS with no ARIA or semantic markup, just setting the CSS width of the inner div to make the progress bar the right length.
The score indicator was a colored div positioned somewhere in between two labels, "Pure Nerd" and "Pure Geek".
I'm tempted to post the stand-alone HTML file on my web server. If I do so, I just hope it doesn't get shared without context, especially since, assuming I post it unedited, the point is to show what the LLM did wrong.
Is your disappointment more with the code or the actual quiz content? The questions certainly aren't subtle. And it always put the "nerd" choice first.
@masonasons I think sharing it with inline commentary would be really interesting. (To the extent that this is based on anything I wrote, please don't worry about that part, the relationship is tangential.)
@masonasons just took the quiz and as with every LLM output, it conveys a shocking, almost magical level of "dang it's incredible that this even works" with a very disappointing, lackluster performance at the actual task :)
indeed, but that's exactly why you find a way to put it in the quiz ;) but okay if the prompt specifically elided it maybe I can't blame the model for that
Which Rust library? As far as I'm aware, well-behaved Python extension modules that happen to be written in Rust can generally be built using pip in the standard way. You do have to have a Rust toolchain installed, but not uv. Which is why maintainers of such extension modules tend to publish prebuilt wheels for the common platforms.
Muss jetzt mal #NextCloud lobpreisen 🙌 . Betreibe seit einigen Monaten eine Instanz und bin relativ begeistert. Die Apps bieten echt ne große Vielfalt an Funktionen, Sync von Kontakte, Kalender, und #Firefox Lesezeichen, alles kein Problem. Und die Kosten von so einer Strato #LinuxVM sind auch sehr überschaubar.
Mit der beliebtesten selbst gehosteten Kollaborationslösung für Millionen von Benutzern, eingesetzt von mehreren Tausend Unternehmen auf der ganzen Welt
Did your AirPods case just beep? Here's why "...To help ensure that your AirPods microphones and speakers are operating at their best (for example, to help provide high-quality hearing test results), your AirPods may periodically play a quiet chime when they’re in their charging case." macworld.com/article/2574141/d…
Thanks https://pcbway.com for sponsoring today's video!Today we're tearing down the SGI Indigo2, checking out its weird architecture, and tossing in some upg...
@hub I envy you so much for this. The closest I got to a Unix workstation was one of the cheapest Sun pizza boxes in '03 (it was woefully outdated back then)
@gabrielesvelto it was in 1995, the summer of Windows 95 release. Later I had access to Digital UNIX Alpha workstation (school) and later one Sun Ultra 5.... which was when things started to level up with PC.
I've been in the corporate world for a while now, but this just plain baffles me... Why increase Tim's salary by so much? He doesn't need this money, neither did Apple generate this much to justify Tim's increase. I know their stock went quite a bit over the last year, but is this really all Tim's doing? Anyone? finance.yahoo.com/news/apple-c…
I was thinking of commenting, but was not sure what to say there. Are they askking for our opinion about something specific in this particular matter? What Accessibi has done is wrong and no one should be allowed to do that. Would that be a good Comment? LOL
"Today, President Biden spoke with His Holiness Pope Francis and named him as a recipient of the Presidential Medal of Freedom with Distinction." - White House whitehouse.gov/briefing-room/s…
Archos
in reply to Václav Pašek • • •