Daily reminder that LLMs do not process words as words but as tokens - basically number IDs. There is absolutely no room there for any kind of understanding to take place. The one thing these things have been optimised to do is output text that is statistically similar to other text
A way to explain how AI works, is that it was access to all the puzzle pieces you could possibly want, but it has no way to differentiate one puzzle from another, so you’ll have pieces from a lighthouse in your puzzle of a cafe
I like how the second one contradicts itself within the slop paragraph.
This actor played a character named Harald (but not Hardrada) in Vikings but was never on Game of Thrones.
This actor is Danish and has nothing to do with Finland.It’s an improv machine that will “yes, and…” your search. It gives me different answers when I refresh
Well duh, you didn’t tell it to be smarter first
Writing a 250 word prompt full of technical jargon, context and conditionals
to ask chatgpt5 a question about why my poop is green and slimy like a duck’s
Will never forget the time I was trying to find a video clip of King of the Hill and the Google AI invented an episode of the show whole cloth and presented me with a outline of the plot that included a section called Hank’s Reaction to the hate crime
Hank's reactions to events around him are an important part of the series and in this episode his reactions are important and drive the narrative forward
: “You WILL use the heckin’ innovoooshun, and you WILL like it!”Even AI is advising people to take the Centrist position, get owned tankies
This wasn’t even political! A game has 3 options presented as left/middle/right and I want to do the option the least people are likely to pick because I’ll have less competition for resources in that area.
You just dont GET it man. Obviously both are true because anyone who doesnt choose middle spontaneously combusts and so the only choice ever made is middle. Hence it is chosen both least, and most frequently. duhhhh
I’ve experienced similar when googling for stuff and it pops up, Google AI makes up stuff on a scale that makes ChatGPT hallucinations seem like a non issue.
This is my favourite one so far
Look, figuring out the difference between the largest number and the smallest number in a set of 3 numbers is really, really hard, ok? Just invest a few more billion dollars into the project and Google AI will be able to get it right at least 70% of the time in about 5-10 years.
There’s a 2/3 chance it was correct once, I suppose…
they’re not sending their best.
Its philosophical, can’t you tell? /s
I grew up hearing that, in general, people tend to choose left more. I don’t have any citations or anything, it was just a random “weird fact” that I heard more than one person say when I was younger.
