LLMs are a proxy fight against end-stage capitalism and a coming to a technofeudalist society.
Being able to combine the worlds' knowledge into a massive array that can probabilistically answer *known* things is amazing indeed... Even if you know how it works. This is Leibniz' dream, of calculus of words.
Most of us lament the Library Of Alexandria's burning. An Open data LLM is like the largest free talking library, that uses calculus to provide predictive text.
But its not about advancement of humans. Corporate LLM token dealers are being used AS weapons against humans. Why? Cause LLMs have the appearance in solving wages - that part of capitalism that the owners are always looking to cheat.
And also, Marx was right, as was Ned Ludd. Its all about controlling the means of production. And in Ludd's time, he wasn't able to gain control, so he chose the second path- destruction. Soon after, that was a death penalty offense.
Are LLMs the final answer? Obviously not. They can't train themselves. They use waaaay too much energy in training and inference. The algorithms work, but are really terrible. (Studying chess algos is the similar path I see with LLMs)
But as a tool, they are useful, especially in fuzzy searches. But capitalists will capital, and try to destroy labor wherever they can.
Don't let tool hate misdirect the real actors are. It always has been capital. Always has.
