None of them feel like tools, at best you have to force them to be that, but are still really bad at it, see vibe coding and “AI-assisted” or higher effort AI images.
It’s also like they skipped a few steps too. I think an actual synonym search could be way more useful than a sycophantic chatbot built into my text editor (wonder if something like that already exist). At least in case of animation I heard some frame interpolation tool actually made for artists got trashed due to their makers wanted to hop onto the still in-infancy full-on generative AI bandwagon.
However I have to give it to them when it comes to amazing outsiders. They often already thought digital art or composing on the PC was something like what the AI was, only they didn’t know what button to press.
So, I do LARP. And LARP means you roleplay as someone else. To roleplay as someone else, you need to know a few things. For example, I play a pirate, so I need to know what a forecastle is, what a capstan does, who the boatswain is and that a mizzen is the third mast. I am entirely capable of inventing a 10 minute story that involves lots of nautical terms about chasing and boarding a ship. That story will be entirely plausible and will impress anyone who doesn’t know anything about sailing a full-rigged ship. It might even be correct, but not on purpose.
And if you put me in a room with someone who DOES know about sailing anything larger than 4 meters, I’m going to fail as soon as they start talking. Because obviously I only know enough to pretend to be a sailor. I can sound like one just fine, but I’m not.
LLMs are larping at being experts. They can’t actually do anything, they can only sound convincing. They might even be right, but they won’t be right on purpose.
Yeah, that’s the big irony of this all: natural language is pretty much the worst way of conveying information for any task that’s even slightly technical. There’s a reason why architects don’t just tell the construction workers what the building is supposed to look like.
The one thing it does have going for it is accessibility, but unfortunately, it doesn’t lead to transferable skills. Like, if you do something like building with Legos, cardboard crafts, make collages from magazine clippings etc. and you want to start doing something more advanced, say, carpentry or graphic design, then what you’ve learned previously is still valuable. But if you just vaguely order an AI around until in produces something that “looks about right”, all you’ve gained is the skill set of a clueless executive and the limits of the model are your limits.
Personally, I don’t think there is a solid plan. Any plan. They are throwing very expensive spaghetti at the wall, mostly undercooked, and little sticks. They are the gamblers at the roulette table betting the house on double zero.
Proven usefulness is rare. Helping people with disabilities leading a more independent life is one of those. Are those the folks with deep pockets? I think film making will include this in one way or another. Does that justify building all these data centers and the costs attached? The only area where I would say to hell with the costs is if this can help in the fields of medicine or pharmacology but I don’t think they have proven usefulness there satisfactorily. So who will pay back the insane debt this pie in the sky is baked with? Answer: anybody, as far as the so-called-AI peddlers are concerned. But that wasn’t the plan. That’s probably already pivot plan T at this point. It’s not built for normies, but it’s marketed towards them, out of necessity at this point. To the detriment of overall ability to think critically and for the environment.
There is a plan, but it is not the plan people think it is.
There’s an initialization people at large need to learn: TESCREAL. It’s essentially a religious movement (though the loose association of people in it would not agree to that label) that comes straight out of the pages of LessWrong, which is itself effectively an AI religious cult, complete with its own eschaton in the form of a godlike AI at the end of time that will resurrect all of humanity in the form of algorithmic simulacra, eternally punishing those who did not work toward helping it come into existence. (But no, it’s not even slightly a religion. 🙄)
So all these AI techbrodouches? Their fanatical devotion to the “AGI” that’s coming just around the corner? Their utter contempt for any form of consequences their literally fantastic beliefs and actions may hold? There’s a reason for that, and that reason aligns with the neologism TESCREAL.
Oh god, I remember reading about roko’s basilisk years ago. Wondered if it’s actually a dangerous thought experiment and could seriously damage people.
It has damaged people. Look at the entire AI sphere. That’s just a collection of the most damaged human beings on the planet.
I was already thinking it is a kind of belief system.
Otherwise all the grandios claims what AI can do would be checked carefully, like, you know, in science. Trying to falsify hypotheses. And for extraordinary claims requiring extraordinary evidence.
But tech had already a tendency for cargo-culting.Things like “agile” practised like a religion. Tech companies acting like a cult. Free speech, critical thinking, or democratic decisions in a company? That would be communism.
lel yeah !sneerclub@awful.systems was a revelation to me. still not sure exactly how to feel about that whole thing.
if this can help in the fields of medicine or pharmacology
I suspect there are a lot of niche scientific fields that benefit and will continue to benefit from the enormous LLMs being built. Problem is, those will be used by a small number of scientists and technologists, applied to very specific problems. And for most scientific problems the existing open-source LLMs are good enough. But the AI companies want mass-market subscribable services so they can leech money off us forever.
However I have to give it to them when it comes to amazing outsiders. They often already thought digital art or composing on the PC was something like what the AI was, only they didn’t know what button to press.
I don’t understand, what do you mean?
A lot of people think that digital artists in particular (whether musical or visual arts) are just pushing buttons until they get a sound or picture they like. They don’t understand that actual composition (whether musical or visual) is an actual skill that takes actual cultivation so that you know which buttons to press and why.
