There's this video series on some social media that's just, like, clips of the folk (okay, it's always bros), clips of the bros who run AI companies, saying just, absolutely ridiculous things.
It makes it clear why folk would think LLM output is meaningful and reasonable, because these bros are considered reasonable (because they have money), and they say shit like
"We should make meat out of oil, because you can make anything out of oil. Gels! You can make gels. If you wanted to make something from scratch you wouldn't start with a plant, you'd start with oil! You wouldn't make a gear out of plants! you wouldn't make rubber out of plants!"
...We wouldn't make rubber out of plants?
*That's where rubber comes from*
And I was actively, as I heard this nonsense, 3d printing a gear made from corn starch.
Like each *word* in the sentence this bro said, made sense.
But only if you ignored 99% of what the words could mean, in favor of the 1% they HAVE to mean, grammatically, based on sentence position, and ALSO ignored that if you tested the truthiness of those sentences, they'd still end up failing!
And that is *stunningly* close to how AI itself decides what word to say next, and *is also literally how Americans and most Anglophones are taught to assemble sentences in school*
And so while it is OBVIOUS when someone is talking from a position of absolute disconnect with reality, it's logical to assume the same flaws in reasoning and speech are still there for everyone raised up the same way, and it's just less obvious when a person isn't a billionaire.
(Which I am refraining from making this a longer post that talks about how there are forms of communal leadership that basically exist to serve as drag performances of the social values around them, so that people can interrogate how they feel about those values.)
I just can't wait for the day when folk realize that if they think so similar to an AI, and they hate how AI think, maybe they should try thinking in totally different ways.



