AI is a subject I actually know a great deal about, because I've been studying it for over 40 years.
And we have made no progress on strong AI in that time. "Strong AI" is the ability to manipulate concepts, for example, to have a conversation, listen to a story and then answer questions about it.
The idea that an AI would "teach us" is as far away now as it was in the 70s when I wrote my first LISP programs.
If you doubt me, try to find a program that you can type at and will actually have a conversation where it understands what you say even in the basic manner a four-year-old does, and can answer simple question.
---
When you talk about the "insane growth" of AI in the last five years, what exactly are you pointing at? There's been a lot of hype but what has actually worked and changed the world?
My guess - a huge amount of "machine learning".
ML is a worthy field, one that had some fine initial successes (games, language translation) but is not in fact "artificial intelligence".
An ML system is a massive statistical model of an existing corpus of data prepared by humans. It also doesn't give you any explanation about its answers, there's no way to validate or invalidate the result, and paradoxically, you can't "teach" a machine learning model its errors - you have to throw away the whole model, change your code in some way that you hope will give better results, and train the whole model on the whole data again, taking weeks and hundreds of thousands of dollars in computation, hoping it will do better.
(And there isn't even a theoretical way to get around that last one.)
This is nothing like "artificial intelligence".
---
Science is not infinite. You just haven't addressed that statement, but that's the key point. There isn't going to be breakthrough after breakthrough because there aren't an infinite number of significant truths. At some point, we'll either have discovered nearly all the truths about the fundamentals of science, or have reached the point of diminishing returns where it's impossible to get further.
And we might already be in that spot.
And we only have a few decades to get it done, because after that our collapsing biosphere is going to occupy almost all of our attention for the rest of history.
We've just experienced almost two centuries of sustained exponential growth, and that has led everyone to believe that exponential growth will go on forever, but this is impossible in a finite world.
People want to believe that this exponential growth will continue forever, so that's why they believe that strong artificial intelligence is "right around the corner", as I was told in 1980, or that a self-sustaining Mars colony will happen in Elon Musk's lifetime, when any realistic evaluation of a such a colony shows that it would take quadrillions of dollars and most of a century to set up, even if it were possible.
I tend to see things from a very large historical viewpoint and from that view, I see us as "almost at the end of this society". We've gone too far with the CO2 and we aren't turning back. Whether we have twenty years or a hundred, we're not going to get too much more science done before new science becomes impossible.
----
Thanks for a civilized conversion! (No, that's not intended to put the kibosh on further chat. :-D)
