Most relatable thing I've seen an AI say so far, to be honest.

@lizardbill
AI even thinks today is 2024…
But hell yeah, let’s stop thinking for ourselves.

#artifificialstupidity

@lizardbill

Generative AI is getting more and more "human" all the time!

😆

@lizardbill Just tried the same with #ChatGPT (GPT-5).

But given an #LLM produces statistically probable human-like answers, and a lot of people make this kind of mistake, maybe one could see it as a correct response? 😉

@lizardbill It's just missing the "holy shit" at the end of the reply.
@lizardbill math is hard even for gemini
@lizardbill "Im sorry Dave, but i'm afraid i can't ... oh wait!"
@lizardbill Reminds me of this conversation I had with GPT-4o mini
@Jonas @lizardbill So, it pretends to run Python for the proof but then just fakes the output? (I see this correctly, don't I? It's a dictionary with multiple instances of the same key.)
@Jonas @lizardbill This isn't just "hallucinating the response", it straight up manipulated the output in order to gaslight you by adding an illegal index that would add to 3 ns. What the hell.
@lizardbill If you imagine it shocked at the realization (as I was) this series of sentences is absolutely plausible.
@lizardbill brain considered harmful
@lizardbill "nah, can't be that long ago, see, if you subtract 1995 from 2025 you get... Oh..."
#feelingoldyet
@lizardbill I went to a thirtieth anniversary screening of Clueless a month ago!
@lizardbill google has fixed this, which is unfortunate, since I’d like a crisper screenshot
@lizardbill But that's the objective, to make AI speak and think like a real human being. 🤪
Alex 🏴Ⓐ🔆 (The Anarchist Turtle) (@theanarchistturtle.bsky.social)

This is real try it yourself

Bluesky Social
@lizardbill AI is senile – that’s it. It has ingested so much of everything that it has gone bonkers from the start. It is senile.
@lizardbill AI coming to terms with the fact that it is now middle-aged.
@lizardbill @thomasfuchs that sounds like some people I have regrettably met. So hey, credit to it for being relatable at a moronic level.
@lizardbill is it showing AI is becoming more humane.
@lizardbill
These quick-fix patches they're slapping over the viral gotchas show EXACTLY how the tech is gonna evolve - a huge crufty patchwork of deterministic catchers overlayed atop the nondeterministic lying machine
@jaystephens @lizardbill we're going to reach a point where the majority of compute is needed for the 5 million line switch statement between you and the LLM
@lizardbill The hotfix they made was just for english:
@lizardbill Copilot came up with this fantastic answer today…

@EduEdu @lizardbill

So not only they tell you sh*t, now they also waterboard you with useless information regarding your question?

I see interesting times ahead of us...

@lizardbill didn't work for me, but it easily eats whatever bullshit you feed it

@lizardbill
This is a Monty Python skit.

"Are you the brain specialist?"
"No, I am not the brain specialist.
.
.
Yes, I am the brain specialist!"

https://www.youtube.com/watch?v=M68GeL8PafE

Monty Python - Gumby Brain Specialist

YouTube
@lizardbill the 90's was 10 years ago and no one can change my mind