Hanno Zulla

Another fun thing is remembering how the people who complained about EVs and heat pumps being an unreasonable burden on the grid's infrastructure are now the very same people betting on an obviously unrealistic goal of data center buildout FOMO that overwhelms both the local grid and the local water supply and will leave us with stranded assets of never completed or abandoned data center ruins.

@thoralf@gruene.social @jlink@det.social @thwill@bizzfed.de Ich würde es nicht "willkürlich" nennen, aber natürlich gibt es individuell sehr unterschiedliche Sichtweisen darauf, was "richtig" oder "falsch" ist. Unsere Gesellschaft führt aktuell eine Diskussion, wie generative KI in unsere gemeinsamen Werte passt. Das ist eine politische Frage und damit ein trefflicher Grund, zu streiten. Ich beschreibe gegenüber Freunden die Technologie hinter generativer KI als "Fractal Moral Hazard". Egal, aus welcher Perspektive und mit welchem Detailgrad man GenKI / LLMs / Chatbots betrachtet, man findet krasse ethische Probleme und externalisierte Kosten.

@thwill@bizzfed.de Strawman. Die Kritik ist nicht, dass LLMs "nutzlos" wären. Sondern dass der Nutzen viel viel zu gering ist im Vergleich zu den negativen Folgen für Gesellschaft, Politik und Umwelt. Ich finde es natürlich technisch beeindruckend, dass meine IDE jetzt ein besseres Autocomplete hat, bewerte aber die dafür externalisierten Kosten der Technologie dahinter als nicht zu rechtfertigen. Deshalb auch die angeekelte Reaktion, wenn zB mein Pastor den Output eines Chatbots in der Predigt zitiert oder die Lehrer meiner Kinder mit Chatbots Arbeitsblätter für den Unterricht oder Erzieher damit Bilder für den Hort erstellen.

There is a sign in the Prora Museum that mentions briefly how the planners and engineers of the Prora Nazi Holiday Resort knew that contemporary infrastructure technology was incapable to serve the planned massive building complex with the required services (fresh water, sewage, food, waste disposal, heating) if it were ever completed as depicted by Kraft durch Freude propaganda. They knew the project will fail the moment actual people were to use it. But the mere fact of building and showing off a construction site was a propaganda success for leadership demonstrating will power. Nothing to learn from that for current events, I guess.

The Apollo Project of our generation should have been fighting climate change. But those in power are telling us now that it's AI, instead. It's inevitable, they say, we should happily embrace this. It's this betrayal that makes all of this so disgusting.

The coming months are going to be really funny, watching US tech bros agonizing over whether signalling loyalty to the Trump administration is their priority or not when choosing SI over AI to describe the technology.

@neil@mastodon.neilzone.co.uk back in previous VR booms (like 3D cinema, it's coming back every few years), I read a few popsci and tech articles about this problem, explaining why it's unfixable. Also how game designers were trying to bypass this by building scenarios with as little player movement as possible, e.g. by using teleportation to relocate within the 3d world or by keeping the player stationary seated in a cockpit or standing at a control panel while it's only other things within the game moving around or approaching the player. People can accept a short game at a VR arcade as a novelty experience, but at home you want to play for hours if the game is good. But then, eventually vomiting is basically impossible. All of this was widely reported before the Metaverse folly was announced, so it was very very weird when the tech press simply accepted Zuck's market predictions with very little pushback at the time.

@neil@mastodon.neilzone.co.uk With the exception of some very few special game scenarios such as Beat Sabre or a commercial jet simulator, where the protagonist remains at a fixed location within the game world, most scenarios require the player to move while the body's senses disagree with the virtual movements that the eyes see. This is human nature and thus this it is unfixable. For decades, the military threw unlimited funds at the problem and still failed to fix the VR motion sickness effect for their simulators. Turns out that people don't want to pay hundreds of Euros to projectile vomit in their living room, after all.

@isotopp@infosec.exchange mMn kannst Du eine generative KI nur dann wirklich sinnvoll nutzen, wenn Du als Anwender bereits Experte auf dem Thema bist, für das die generative KI für Dich Output produziert. Generative KI wird aber aktiv als Werkzeug vermarktet mit dem Versprechen, dass Du ohne vorherige eigene Expertise Aufgaben erfüllen kannst. A PhD in your pocket. Bzw dass Du keine Experten mehr einstellen musst, weil ungelernte Mitarbeiter diese Arbeit jetzt erledigen können. Dieses Versprechen hat die Mehrheit der Öffentlichkeit auch so geglaubt, wenn ich mir die Art ansehe, wie vom Lehrer über die Teamleitung bis hin zum CEO jetzt generative KI einsetzen. Das ist das Problem, das ich hier sehe. Nicht, dass generative KI jetzt "da ist" und "nicht wieder weg" geht. Sondern dass ein sehr sehr großer Teil ihrer Anwender fest daran glaubt, dass deren Ergebnisse gut genug sind für das, wo sie Chatbots einsetzen.

These climate events had been accurately predicted decades ago. We had technologies to prevent the scale of this - renewable energy, batteries, electric cars, etc. - ready for a large-scale roll-out long ago. We had discussions, actual proposals, plans and budget estimates to do this since 30+ years.

We currently find ourselves in an escalating climate catastrophe that everyone can witness to have real consequences. The new normal in Europe is droughts, wildfires and heatwaves. We have petrostates engaged in multiple idiotic wars and trade conflicts that dismantle our global supply chains. We're seeing lower crop yields and coming food shortages. We have a realistic outlook to see the end of AMOC in our lifetime.

And now? It took the world a finger-snap to go all-in on AI. We were capable to decide this. To literally throw trillions of cash and resources into this. Only that this is the modern equivalent of investing into a tulip bulb merchant who is promising us a never before seen future black tulip flower.

@dougmerritt@mathstodon.xyz @TimWardCam@c.im @technicaladept@techhub.social Markov Chains were known before Clarke's short story. There's a famous popsci article written by Brian Hayes published in Scientific American in 1983 that popularized the concept to non-experts: http://bit-player.org/extras/bph-publications/SciAm-1983-11-Hayes-drivel.pdf (That article inspired me as a high schooler to write small Basic and Pascal programs using Markov Chain text generators.)

Sturmgeschütz der Demokratie habt Ihr Euch mal selbst genannt. Ich will keinen Kauftipp-Servicejournalismus. Ich will, dass Ihr über das Versagen der deutschen Politik in Sachen Klimaschutz schreibt. Ihr könnt Agenda-Setting, dann macht das bitte auch.

@langhaarschneider@sueden.social die WDR-Hörspielproduktion von "Die unendliche Geschichte" ist ausgesprochen großartig und sowohl als Tonträger zu kriegen als auch auf den üblichen Streamingplattformen verfügbar.

@finn@social.0x520.eu Linux Mint Debian Edition and installing desktop applications via flatpak is a pretty nice blend of boring foundation that just works, beautiful GUI and always-up-to-date applications.

Unsurprisingly, text rendering is just perfect on an e-ink display. This will make a beautiful family calendar thanks to https://github.com/FamousWolf/week-planner-card @rudy@fosstodon.org

Yet another non-techie sales person working for a large tech company telling me offhandedly that they don't need to hire as many software developers like they used to because of AI and it's fucking grinding me down.

@henrik@eliitin-some.fihttps://www.steintor-variete.de/event/monkey-island-ich-will-pirat-werden-2/ Monkey Island was made a theatre production 10 years ago and has just been remade to return to the stage.

@langhaarschneider@sueden.social und dafür zahle ich Gebühren: https://www.ardsounds.de/sammlung/geschichten-kinder-6-102/

Same. https://infosec.exchange/@masek/117291865460131952

Just look at this beauty.

@sam@solent.social I do not understand the problem here. It is fairly easy to mount a CD/DVD on the first computer, share the folder via the network and then mount the shared folder over the network on the second computer. You can use sftp, smb or nfs, it shouldn't matter. All of this should be possible through GUI tools with no line of code, but you can also automate some of it using batch scripts.

@sam@solent.social however, the absolute easiest solution if you want to avoid having to hassle with this is to just buy an external DVD drive and connect it via USB to the target computer. It seems like you can get a new one for < 30€ and used ones even less than that.

RE: https://mastodon.social/@europeanspodcast/116968326350349013 This is a very nice podcast episode discussing summer music tracks from all over Europe.

They're selling this album for 6x of its original price tag now.

The moment corporations started to believe that a machine will make annoying workers easily replaceable or even unnecessary, the moment organizations found out that using AI offers plausible deniability for difficult decisions that can potentially hurt other humans, it took the people in power an *instant* to redirect all of our attention into advancing this technology.

Ein Tag, zwei Meldungen zum gleichen Ereignis.

"There will be a human in the loop" the proponents are telling the public. And yet this is how the same industry sold us self-driving-cars: "The person in the driver's seat is only there for legal reasons. He is not doing anything. The car is driving itself."

@abulling@fedifreu.de Die Reaktion auf LinkedIn auf Deine Beobachtungen fällt wenig überraschend aus: Die Kommentatoren verteidigen den Einsatz generativer KI und beschimpfen stattdessen die Lehre dafür, dass sie die Technologie falsch verwendet. https://www.linkedin.com/posts/hanno-zulla-a5990382_ein-informatik-professor-der-ki-an-sich-activity-7486496897935929344-O24p

The traditional hacker's game of "can it run Doom" is being abused by the Trump administration in their attempt to undermine the upcoming midterm elections. https://talkingpointsmemo.com/news/the-doom-computer-game-was-part-of-trumps-wild-and-baseless-election-security-nightmare

@merlin@toot.kif.rocks die Gerätekategorie heißt "Dokumentenscanner". Ich habe hier einen Scansnap ix1500, den ich wärmstens empfehlen kann. Läuft auch unter Linux. (Vorsicht, es gibt Geräte wie den ix100, der keinen automatischen Einzugschacht hat, weil für Reisen.) Du wirst mit dem Suchbegriff "Dokumentenscanner" oder "Scansnap" finden, was Du suchst. Manche alten Geräte haben verdreckte oder kaputte Rollen in der Mechanik, deshalb vor Kauf testen. Wartungs- und Reparaturanleitungen findet man dafür im Internet.

We hastily rebuild our society and economy around it. We're told we'll be left behind if we don't accept to include AI in our workplace, our kid's schools, the culture we consume, the bureaucracy we rely on, the military decisions on how to fight our wars, and so on.

@TimWardCam@c.im Stanislaw Lem described another generative AI writing machine and how it causes LLM psychosis in "Cyberiad" in 1964. https://nibblestew.blogspot.com/2026/08/stanisaw-lem-foretold-current-llm-mania.html?m=1

It's also curious that there aren't more used devices for sale to end users. Shops and Cafés redecorate (or give up) all the time and there must be obsolete devices left afterwards, no?

https://www.mopo.de/im-norden/schleswig-holstein/aktion-das-letzte-hemd-wer-die-koepfe-hinter-den-ortsschild-protesten-in-sh-sind/4893712 Das Hemd hängt hier auch gelegentlich im Osten an Ortsschildern. Nach "Projekt M1llion" gibt es damit eine weitere "Protestbewegung von unten" mit betont unklaren Hinterleuten.

Can't get over the fact that several hierarchy levels within Google working on their primary product used to display geopolitical reality have approved adding a disinformation generator and then took extra care making it easy to use. From idea to shipping, they must have spent hundreds of man hours on this generative AI feature in Google Earth and noone in charge even considered the obvious possibility that this will harm the reputation of their product when it results in misinformation campaigns and world political turmoil?

The visceral disgust for the technology doesn't come from its obvious niche usefulness. But from its externalised costs we're told to ignore to embrace it. What do we have here? We have combined a massive problem space fuzzer with an addictive Clever Hans Effect machine. Tech bros are successfully lying to the public about it having magical capabilities.

@europeanspodcast@mastodon.social I'm halfway into the episode and am jumping up and down to respond: While it is true that air conditioning means pumping heat outside of the home, it doesn't necessarily have to be pumped into the neighborhood air. My own home's geothermal heat pump is doubling as a passive cooling system in summer, pumping the heat underground to be stored and partially used again later in autumn. It's pretty nifty, we're not blowing our heat towards the neighbours and it's also totally silent with no AC split unit outside. It's also possible to do this in larger scale and have municipal cold water circulation to city apartment blocks, making it then possible to capture and reuse the heat from the homes for later use. See the link for a large scale example in Paris. https://www.theguardian.com/environment/2026/jun/26/underground-revolution-seine-cooling-network-paris-buildings-heat I do admit that this is way more complicated than just putting up an AC, and passive cold water solutions like mine also have far less cooling capacity compared to the real deal: My cooling system is able to cool down the roof floor to <= 27° on a heatwave day where an actual AC would most likely be able to go down to <= 23°.

@TimWardCam@c.im Roald Dahl published a short story "The Great Automatic Grammatizator" in 1953 that accurately predicts how a generative AI text generator works and what economic pressure it would inflict on authors. https://gwern.net/doc/fiction/science-fiction/1953-dahl-thegreatautomaticgrammatizator.pdf

A good friend who's vibe coding now tells me I'm talking like a blind person about colours when arguing to reject generative AI without "really" trying it out first myself. The issue isn't if generative AI can be useful for highly specific work and operated by experts for that task. Generative AI is a genuinely impressive and fun tool. Especially in prototyping and security fuzzing. But it still causes a deep feeling of disgust, at least for me.

@merlin@toot.kif.rocks ein Dokumentenscanner für's Büro kann in aller Regel beidseitig scannen und ist um mind. eine Größenordnung schneller als jeder automatische Blatteinzug eines bezahlbaren Multifunktionsdruckers.

The colours are a little pale and imperfectly dithered, but it's still impressive and feels very much like a piece of graphics art display. This is so much fun to play with.