Recently I have been changing the way I learn things. I used to read textbooks and then solve problems in them, but this always resulted in me learning more than required for some topic and also learning a lot of things that I may not use anytime soon, if ever.
I enjoy learning things, but the problem is that sometimes just learning is not enough. It can even be boring, I don’t want to become wikipedia, nor I want to be unable to do new things by being restricted to always wanting to learn something new, which for many fields you can spend your whole life learning and will die knowing you still have so many more things to learn.
Because of that, I thought I was wasting my time learning things just
for learning. Obviously, learning is never a waste of time, but as
Goethe once said Es ist nicht genug zu wissen, man muß es auch
anwenden; es ist nicht genug zu wollen, man muß es auch tun
and so I
decided to learn in a different way, that is, learn as doing. I only
learn something if for some reason, what I’m doing requires knowledge in
that something.
Obviously, one problem that appears from this is that sometimes I may not know what is the thing I should learn to solve the problem I’m facing, but thanks to the use of AIs like ChatGPT, I can formulated my problem in a way that the AI can easily identify what I need to know to solve this. In the cases it can’t, I can easily restrict the problem (either by changing parts of the question or simplifying it, for example) until either the AI can identify similar problems to it or I can solve it and use the knowledge from that.
One must use AI for the cases where you learn something from it, and avoid all the cases where using it doesn’t make you better at something. Use AI as IA(Intelligence Amplification), not as AI(Artificial Intelligence).