this post was submitted on 28 Apr 2026
721 points (93.6% liked)

Fuck AI

6859 readers
1721 users here now

"We did it, Patrick! We made a technological breakthrough!"

A place for all those who loathe AI to discuss things, post articles, and ridicule the AI hype. Proud supporter of working people. And proud booer of SXSW 2024.

AI, in this case, refers to LLMs, GPT technology, and anything listed as "AI" meant to increase market valuations.

founded 2 years ago
MODERATORS
 
top 50 comments
sorted by: hot top controversial new old
[–] HalfSalesman@lemmy.world -3 points 4 hours ago* (last edited 37 minutes ago) (2 children)

People relying entirely on AI to make all of their decisions though seems like a rarity.

You have to be particularly narcissistic or inclined to bipolar or schizophrenia because AI's are irritatingly yes-men coded still.

Every time I try to use an AI and I suggest a counter to its points I often get frustrated when it fails to make very convincing counter points to me and just is like "You're made a very astute point" or some shit. I don't need to have my fucking ego stroked.

Plus there are times when I want to argue about something taboo, and it starts to literally mimic the worst aspect of human intellect: judging my character for even arguing a certain perspective. Its like, motherfucker one of the bigger reasons to use you is to avoid judgement. I want to know if an argument is actually wrong, not if its suggested into being wrong by cultural mass peer pressure over virtue.

Its a result of being so pre-occupied with being helpful and safe in their training process.

I need a sociopathic "intellect only" AI so I can argue with it. I want to know if I'm right about shit. Maybe I could finangle one with prompting but I doubt it, its too ingrained in their very design.

EDIT: Oh, I should have noted what com I was in. Oh well, I wont delete it. Its my honest thoughts.

[–] IzzyScissor@lemmy.world 1 points 1 hour ago

There is no 'maybe' there. LLMs have no metric or ability to tell if something is 'true' or not. Any 'finangling' you'd have to do is so much of the work that the LLM is doing the same thing it always does - "That's a great point!"

You'd have better luck asking the paint bucket in MS Paint what a rainbow looks like.

[–] trackball_fetish@lemmy.wtf 1 points 2 hours ago

@grok please make sense of this rambling

[–] shani66@ani.social 5 points 10 hours ago* (last edited 10 hours ago) (2 children)

Why do i hate this art style? Not that it's bad, although i think it kinda is, but it's got a real negative association in my head. It's like seeing a rockthrow comic and immediately getting icked out by the Nazi that made it.

Edit: wait, i think it's because it reminds me of some gamergate comic

[–] mfed1122@discuss.tchncs.de 7 points 4 hours ago (1 children)

Yeah, even though pizzacake takes stances I generally agree with, her work has that "this comic only exists for me to express my social and political views in a way that amplifies my voice online" type of quality. This type of art usually is:

  1. Flat or pillow shaded for simplicity, because the beauty of the image isn't relevant to its purpose, and the artist needs to produce it quickly to respond to current events
  2. Depicts characters, even protagonists, in a mockingly exaggerated way, because this is the main form of stylization the comic must support in order for the author to make fun of strawman versions of positions they don't like (because anything beyond a strawman, regardless of position, does not fit in a digestible 4 panels)
  3. Simple and repetitive poses with no dynamicism, often because the artist is not a very skillful artist, because they only developed the skill to make comics like this in the first place

Basically, it's visually unpleasant because the only reason this has any visual component is to make it more eye-catching compared to just text. Unlike most art, it does not need to be visually appealing, and in fact even benefits from being visually unappealing, as an art style that pisses you off a little is more likely to make you feel a vague sense of there being a problem to resolve - which the comic tells you is the issue it presents. Even advertising has to be eye-catching first, but memorable and compelling after that, because you need to take it in very quickly. Whereas this has the expectation of reading built into it, which frees it from what little need to be pleasant it may have left. The visuals are nothing more than lubricant to help the text go down.

[–] LetThereBeNick@lemmy.zip -4 points 3 hours ago (1 children)

The art also serves to promote her onlyfans. Do you think her lighter brunette friend is on there too?

[–] mfed1122@discuss.tchncs.de 2 points 3 hours ago

Honestly I have no intrinsic problem with people doing onlyfans. It can be gross or problematic depending on how dishonestly it's approached, but I often feel like its pretty honest and harmless? I guess with her, it probably relies a lot on sad weird redditors who are in love with the simplified yet agreeable entity her comics portray her as. That's a little bit in gross territory I guess, but I'd rather she made money with porn than with putting out this "slop"-tier circlejerky junk that reduces a complex world full of discussions worth having down to:

"Me and you (Smart and right and based)

People we dislike (dumb and wrong and cringe)"

In my opinion, her comics do a lot of harm to the state of political discourse here and even more so on Reddit. It would be so much better if she would make cleverer, or funnier, or at least more critically thought-through comics (but that wouldn't be as popular). Barring that, it would be less harmful if she just did onlyfans. But also I just don't see an issue with selling that kind of stuff anyways.

[–] adaveinthelife@lemmy.ca 7 points 11 hours ago (3 children)

Maybe I'm using it wrong, but every task I give copilot comes back with empty documents. It shows me a preview of the work it did, and it looks okay, but the documents are always empty or lack the actual content while full of filler.

And it always responds with oh yes you're right there is no content, here you go this time with content. No content.

This has to be a joke, right?

[–] YourMomsTrashman@lemmy.world 9 points 6 hours ago (1 children)

You're absolutely correct - There's no content in your documents. 🚫📃

Would you like me to format your documents again, without actually leaving them empty this time? ✏📫

Notice: Your free trial has expired. -> View upgrade plans

[–] adaveinthelife@lemmy.ca 3 points 4 hours ago

Please I have PTSD

My guess is that the software or website needs access to a local temp folder to write in before handing you its work. It probably doesn't get that from the server in basic versions.

[–] Canconda@lemmy.ca -1 points 5 hours ago* (last edited 5 hours ago)

Maybe I'm using it wrong, but every task I give copilot comes back with empty documents.

I dunno about copilot but I paid for GPT to do a bunch of document migration from legacy pdfs/papers for work. Saved me dozens of hours. Still it was also prone to returning blank or ridiculously bare documents that did not match the preview.

IMO AI would more aptly called IA, or Intuitive Automation. Basically the classic genie in a bottle granting wishes literally trope, but with python.

I'm a beginner>intermediate coder. But thinking about projects and prompts in terms of, "how would I code this" helps a lot. I also try to prompt in a way that "grammatically" translates well to code concepts.

Complex projects inevitably will require multiple prompts. You'll want it to generate a downloadable document at every step and if you're working with data you need it to generate a master copy for reference.. Sometimes you need to go back several prompts and change directions.

[–] laranis@lemmy.zip 64 points 18 hours ago (3 children)

So fucking relatable. Recent experience: I called up my brother who is a car guy to get some help getting an old car started. He's worked on cars, farm equipment, and sold car parts. He knows his shit.

I gave him what I was seeing and what I had done and he sends back a screen shot of what ChatGPT said about what I explained to him.

I would have rather had a 30 minute conversation of us going, "I don't know. Have you tried this? Yeah, oh, well, then, um, hmm. Maybe it is the defanglinator. Or the wires to the pulsometerizer." We're exercising our brains, enjoying each other's conversation, solving a problem together... which is more important than actually solving the problem and why I called you, you dope!

Oh, and the LLM response ended up not being relevant or helpful. We're all fucking cooked.

[–] ayyy@sh.itjust.works 1 points 4 hours ago

Kinda sounds like he’s tired of being g a free consultant.

[–] Demdaru@lemmy.world 10 points 10 hours ago* (last edited 10 hours ago)

which is more important than actually solving the problem

Error. <_> I do not compute. xD

Alas still, being an expert and yet sending someone LLM response is a low blow. xD

[–] Shayeta@feddit.org 5 points 10 hours ago

If you send me uncurated AI output, I will just pipe it back into ChatGPT and reply the result.

[–] Luisp@lemmy.dbzer0.com 36 points 17 hours ago

You are right! AI is shit and do wastes a lot of resources -- do you need help with anything else?

[–] violentfart@lemmy.world 76 points 23 hours ago (3 children)

One of my coworkers is a proud advocate of LLMs. What I’ve noticed is she’s become more scatterbrained and less confident when put on the spot.

[–] TRBoom@lemmy.zip 4 points 4 hours ago (1 children)

Our new COO came into my office the other day and started asking some prepared questions. When we got to our MDM for our iPhones she pulled her phone out and started feeding my answers into an AI to get follow up responses for me.

Not sure how well she'll do... she is quite on the ball for a lot of stuff but I'm certain that the LLMs will rot her brain. :(

[–] violentfart@lemmy.world 3 points 2 hours ago

“Why the fuck are you even here?”

[–] Zink@programming.dev 1 points 4 hours ago

That's also how I act when I haven't been doing my job, lol.

[–] HalfSalesman@lemmy.world 1 points 5 hours ago

Being confident on the spot died the moment you could look something up on the internet. LLM's did not do that.

And it was dead for me at birth, if I'm 0.1% unsure about something I'll stutter and hesitate to pull the trigger.

[–] Gork@sopuli.xyz 83 points 1 day ago (1 children)
[–] lena 55 points 1 day ago
[–] jadetoffee@lemmy.blahaj.zone 20 points 20 hours ago (1 children)
[–] Newsteinleo@infosec.pub 12 points 18 hours ago (2 children)

My boss has become the person in blue. Can't get any work done now because follow up questions are to hard for him now.

load more comments (2 replies)
load more comments
view more: next ›