🤣
Let me help you a bit.
Solving problems in extrapolative domains gives a base that can be translated into bounded/specific ones.
Solving equations is part of how the atomic bomb became possible. Nobody started with “let’s solve the atomic bomb.” Scientists were trying to understand the structure of matter, nuclear fission and chain reactions.
Once those underlying principles were understood from first principles, they could be translated into an entirely different, specific application.
Same way Terence Tao wasn’t trying to solve MRI. He was working on a general mathematical problem whose solution later became useful for reconstructing MRI scans from fewer measurements.
Some of the leading AI profs in the world are Olympiad Gold Medalists in maths and physics, they didn't solve a real world problem back then, the mathematical equations they were grappling with are the foundation of the most important technology of the Millenium.
This is the point of abstraction.
You solve the underlying problem first, and its applications can extend far beyond the problem you originally had in mind.
If every problem has to look like its eventual application, you’re missing the entire value of solving problems at a more fundamental level.
those “engineering” exam questions on the tl just dey make me sad. shows how flawed our system is.
not a single problem solving question. “show that” “assume” “derive”
we’re so cooked