I think a better term for it would be, applied compute. We're applied compute too. Just a lot more complicated & intricate. People are overlooking a profound lesson from the rise of these large models; sufficiently large systems lead to increasingly unexpected emergent properties/behaviors.

If you put enough compute in there, and make the system intricate/large enough, unexpected things come out the other end. It tells us a lot about us & the emergence of intelligence. A magical step didn't have to occur for sentience to arise; anything sufficiently complicated would have sufficed.