this post was submitted on 11 Aug 2026
98 points (93.0% liked)

Fuck AI

7925 readers
1242 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
 

It looks ok but it just doesnt add up...

Wrote to the teacher...

you are viewing a single comment's thread
view the rest of the comments
[–] midribbon_action@lemmy.blahaj.zone 4 points 4 days ago (4 children)

That's a possibility. But it looks so bad. If it was generated by ai, why would they ask to make it look like a grainy scan taken from a book on a scanner from the 90s? Hopefully your note was friendly, as everyone can make mistakes, and teachers deserve big piles of gold delivered every week to their homes. Wouldn't it be funny if you made a mistake or didn't understand the instructions? I have no idea either way, I've never seen this kind of math problem, I just think it would be funny, especially if you were being an asshole to the person looking after your kid for 8 hours a day in your letter.

[–] Feddinat0r@feddit.org 18 points 4 days ago (1 children)

If i would post the whole page, you also would agree that its 10000 percent AI.

Yes i was nice. But you cant give that to your little students and didnt even did the tasks for yourself to check....

Thats the point i just cant justify

[–] midribbon_action@lemmy.blahaj.zone 8 points 4 days ago (2 children)

If you trust the source material, like the publisher of a math book, it's OK to not verify every single practice problem as a teacher. Editors are paid to do that.

If i would post the whole page, you also would agree that its 10000 percent AI.

Well do that then. Why should we take your word on it, there's no evidence here.

[–] mnemonicmonkeys@sh.itjust.works 4 points 4 days ago (1 children)

there's no evidence here.

Besides the fact that neither problem shown is actually solvable. Which is absolutely a mistake a clanker would make

[–] midribbon_action@lemmy.blahaj.zone -1 points 3 days ago* (last edited 3 days ago) (1 children)

Making mistakes is also a human problem, unfortunately. If you change the 8 to a 6, everything works, 300, 220, 350, 100, -50. My guess is a transcription error: the editors cleared the problem, but the typesetter fat fingered it when making a graphic.

[–] Azrael@feddit.org 2 points 3 days ago

OP mentioned that their Kid has not been taught the concept of negative numbers yet. So the -50 for the lower right part of the right "brick wall" wouldn't make sense either.

[–] gedfromgont@piefed.ca 2 points 4 days ago (1 children)

I also want to see the whole AI page please

[–] ZombiFrancis@sh.itjust.works 7 points 4 days ago

Why look like an old 90s textbook? That's also whats being used to train models.

Warehouses of textbooks are being fed into the machines.

[–] balsoft@lemmy.ml 4 points 4 days ago* (last edited 4 days ago) (1 children)

I think this is a plausible sequence of events:

  1. Someone (e.g. another teacher) asks chatGPT to generate a worksheet, prints it out on the 20-year-old school printer (which would already make it look bad)
  2. The teacher in question makes a bunch of carbon copies of that worksheet and hands it out to students, worsening the quality even more

That would explain the nonsensical numbers and the poor scan quality. Remember that not all teachers are comfortable with tech, so making carbon copies of a sheet of paper may be easier for them than obtaining the original digital file that ChatGPT spit out and printing it out.

[–] midribbon_action@lemmy.blahaj.zone 0 points 4 days ago (1 children)

Like I said, it's a possibility. I don't see any evidence for it though.

[–] mnemonicmonkeys@sh.itjust.works 1 points 4 days ago (1 children)

At this point I'm pretty sure you're being deliberately obtuse because you don't want to admit you're wrong

Lol ok, I think you need to chill... The difference between speculation and evidence is pretty clear cut to me. Evidence for a claim requires it to support your claim over other explanations. You can explain why a printout looks like a shitty photocopy either by using ai+shitty photocopier or just a shitty photocopier. Nothing in the comment or post preferentially supports the claim of ai use, other than a perceived math error.

[–] FiniteBanjo@feddit.online 1 points 4 days ago

Because AI is trained off of images that look like a book scanner from the 90s.