Wenn ich mir so nie #KI Scheisse angucke, frage ich mich schon, ob das mit der Evolution jetzt nicht ziemlich schnell in die andere Richtung läuft.
Wenn ich mir so nie #KI Scheisse angucke, frage ich mich schon, ob das mit der Evolution jetzt nicht ziemlich schnell in die andere Richtung läuft.
Our favorite VNC service for the lab has changed the subscription models and now demands horrendous sums for the same thing we have used for years. But I cannot talk to them about it, because both on their live chat and on the phone, no human employees are available (on a weekday at a reasonable time of the day). I guess someone is making a lot of money here.
#AIBullshit
There must be some mistake, I've been informed by multiple tech bros that the environmental costs are overblown. /s
www.theverge.com/news/686387/...
#AIBullshit #politics
Meta inks a new geothermal ene...
OpenAI's API prices are not 1/7-1/8 the price of Grok's. I'm sure Grok is inefficient, but your entire estimate of AI energy usage has been based on API prices: prices don't suggest Grok is that inefficient. The highest GPT API is 3-10x *more* expensive than Grok's highest model. #AIBullshit
Noting that from the AI "centrists" who try to qualify their usage of LLMs or talk about how they understand the concerns of the industry, I've so far seen zero of them use their public platforms to advocate against a budget bill that would place a ten year ban on AI regulation. #AIBullshit #thread
(tilts head) So.. has the software industry stopped teaching about skinner boxes?
#AIBullshit #thread
RE: https://bsky.app/profile/did:plc:byy2wqmi6ctl2pevujhwvxf3/post/3lqxogvkpj225
@davidho Still Wondering What the Hype Is About? It's An Attack On Labor!
I've already gotten into an argument with the author over this so out of respect for his mentions I'm not going to clutter them with another thread, but I want to be super clear: this is a deeply unserious argument from top to bottom. You can not walk a line between art and code here. #AIBullshit
"If you're not on the AI train you'll get left behind." God I wish. You fuckers won't leave us alone. You're constantly pulling us behind you. You won't take no for an answer, you keep shoving this shit into literally every piece of software and every context. Please leave us behind. #AIBullshit
@aral A. I. lies to us all - or the owners certainly do #aislop #aibullshit
I've been seeing a frustrating number of articles recently talking about AI "assistants" and I suspect this is going to be the new buzzword for a while, so I want to preemptively encourage you to look up 'prompt injection' and understand why general AI agents literally *cannot* work. #AIBullshit
I am fucking begging the programmers on this site to go read The Design of Everyday Things (or literally any interface/UX design book). Well designed interfaces are predictable, a very good UI can be used almost without looking. A good UI doesn't change based on your *scroll position*. #AIBullshit
I got popcorn. And listening to comedians talking about "VIBE CODING".
Ayer veo en un stand de una editorial pequeña en la #feriadellibro de #Granada un libro para niños sobre #IA
El argumento es el siguiente: una abuela recibe un robot ayudante con el que no congenia, su nieto le ayuda y al final el robot y ella se hacen amigos. Esa amistad hace que la abuela no se sienta sola.
Es increíble que se le haga publicidad a las ideas de #Zuckerberg #Altman #Musk y compañía de forma tan gratuita #aihype #aibullshit
Btw, seguid el #hiloferia de @sevennorth
"...a government plan to let artificial intelligence companies use copyright-protected work without permission."
This ONE WEIRD TRICK lets you bypass copyright epicly.
The #Multiverse defence of #AIBullshit*:
It hasn't happened in this universe, it has happened in another.
*I considered;
"hallucinations" - no, those are subjective experiences of the hallicinator,
"Lies" - AI has no intent,
"Wrongness" - the product does not match reality & does not conform to ethical standards in many ways, "wrongness" is too general,
"Inaccuracies" - the product is being produced accurately, it is accurate, it is the user who is misusing the product,
so #AIBullshit.
One paper (about Octavia Butler's DAWN) contained multiple quotations, with page citations, that DID NOT APPEAR IN THE TEXT.
Which defect still in 2025, three years in, remains the diagnostic and unmistakable stink of ChatGPT.
The quotes are MUCH BETTER than in 2022, uncannily plausible, unnervingly in Butler's voice, but they DO NOT APPEAR IN THE TEXT.
Mere words are too weak to express my skin-crawling loathing of the scourge of AI bullshit.
KI und die Startups, die daran verdienen wollen, sind furchtbar. Ich sollte Cursor ausprobieren. Für Linux gibt es das als AppImage, obwohl es echt offensichtlich einfach nur ein VS Code ist.
Es möchte nach einem Prompt Befehle ausführen um ein Verzeichnis anzulegen.
Mit Powershell.
Mit einem Pfad zu Powershell in einem Snap Paket.
Erkläre ihm, dass er auf einem richtigen Linux arbeitet und sh oder bash nehmen soll.
Er scheitert an Dateiberechtigungen.
Erkläre ihm, dass er ein AppImage ist und daher in einer Sandbox läuft.
Zehn Versuche später auf diverse Ordner zuzugreifen gibt Cursor auf.
Schöne neue Welt. Für 40 € im Monat ist das doch eine Lohnende Investition in ein schickes VSCode Theme, dass leider nichts kann.
...tout à coup, l'image du miroir du rised dans Harry Potter s'est imposée à moi.
https://www.rollingstone.com/culture/culture-features/ai-spiritual-delusions-destroying-human-relationships-1235330175/
#IA #AIBullshit