"π"
this post was submitted on 12 Aug 2026
418 points (97.3% liked)
Fuck AI
7909 readers
1115 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
If you actually paid $420 for 32GB of RAM, I'd call the police.
I hate this. It's a machine, playing at some sort of humanity doesn't calm me, it creeps me the fuck out.
I sincerely don't know how people "talk" to these things, much less fall in love with them? WTF?
Why do people want AI to type/sound this atrociously?
To be fair, I think a 32GB RAM should have costed $60 by now, if not for this AI shit.
I'm so glad the latest breakthroughs in AI are about doing more with less in the form of Quantization and Distillation efforts.
It's gonna make these AI focused companies like Meta realize how exposed and stupid they are for buying out more equipment then they had any plans to do with. And I hope that they fail hard without a government bailout.
So your saying that when OpenAi bought all of the world's memory wafers for a year, they had planned to stay afloat forever?
Iβm no expert, but Iβll bet that there is plenty of work being done on the ultra-high level to maximize usage of their resources.
Don't configure your software agent to generate first person prose. It's not good for you or the people you expose to it.
Donβt configure your software agent to generate
Say no more
if it was at all capable of maintaining anything, i would have it respond like hk-47 assassin droid.
too bad building upon ai is like building on sand in a bog on an indian burial ground on a fault line
Because in its market, the AI market, RAM is cheap β¦ if you buy in bulk.
Far more likely that the AI was just trained on people complaining about RAM prices. I doubt it has any concept of pricing bulk or otherwise.
The problem confronts itself.
I havent used AI in ages, why does it talk like a redditor now?
Because it ate reddit threads and is now spitting it out.
No. Because that is what the user wanted it to do.
Ew
Presumably, because either it was instructed to or the context had a bunch of a similar style in it.
At that very point, ChatGPT just acts like that random thief that says "I did nothing" to the police officers... While he has a bag full of money in one hand and a gun on the other.
Fake, it actually correctly inferred that 420 is more expensive than the 80-120 range
Doubly fake for not doing a 420 joke
Don't worry guys AGI is real! It'll just cost a few million to update the model with latest pricing data.
Tell it that it's the reason the RAM costs $420. Maybe it'll commit sodoku in shame.
I guess LLMs are notoriously bad at math and logic problems, but doing a sudoku doesn't seem like a particularly bad punishment ..
Why would ChatGPT use phrases like βweirdβ and βcall the policeβ and all these stupid emoji?
Itβs either from the context or a deliberately chosen chat style, I presume.
One thing that isnβt emphasized enough about the ChatGPT app is that it's βfarβ from the LLM. Thereβs a ton of stuff wrapping around it and getting injected into the context, tons of tracking and personalization and engagement maximization.
And the Chat models are finetuned to oblivion. They are deep fried.
It gives me creepy Facebook vibes.
It's definitely deliberately chosen. There's no way it'd randomly just go into Reddit speak; that's not how an LLM like this worse; it maintains that generic, stale toast tone unless you ask it otherwise.
Which also makes me question what the prompts before this were and makes me assume this is faked. Like "pretend to be shocked" faked.
More than a few people tune ChatGPT to have a personalized tone in this vein. This here is a mild case.
It's not a pure LLM though. It's ChatGPT. That's what I'm saying; users have no idea what else is in the context, and it could easily be something that pushes it to "Reddit speak."
When I've seen my parents use it, it brings up things it could not possibly know from the current conversation, so its definitely inserting previous context at the very least.
load more comments
(5 replies)
load more comments
(6 replies)
Because it's glorified predictive text which has scraped the shit out of Reddit.
Because its not ai, it's lookahead autocomplete and those are the kinds of things that would follow the previous thing in the training data
load more comments
(10 replies)
load more comments
(2 replies)
I guess the AI's price memory was frozen at pre RAM crisis (or earlier), so it thinks the current (inflated) price is abnormal.
(BTW, I don't think police will take the case about the pricing, they would just tell you to find another store(s) with more reasonable price. :) )
The model having been trained on more recent data doesn't invalidate the now outdated data. It could be the most recent model in the world, but it might still respond like a forum post from 8 years ago
Yeah I've used Hermes to check some prices for nvme drives it was ballparking stuff which sounded like a few years ago prices. When I got it to websearch to confirm it would find online stores and quote cheap prices from listings that where sold out so that wasn't accurate.
Please, call the police
view more: next βΊ