i came across a rough version of this idea by myself and then someday i found a guy who has extensively written about this and a few other things. it is called do hundred.
when i started to learn to program, i tried to optimise for my fluency in syntax before understanding the intuition
for days and nights and months and years, i longed to perfect printing patterns using loops
in hindsight, i know it was a stupid thing to do as pattern printing sucks.
but then someday i started atcoder and codeforces and churned through an few hundred easy problems like a machine while asking gpt for correcting my syntax errors and understanding them and repeating them less and less.
i got good (fast).
whenever learning anything hard, there comes a point when it becomes too hard to understand. there are a few ways to tackle this issue.
first, just stop and start watching anime. (which is somedays the best option )
second, start from beginning. if reading a physics textbook or some book, there comes a point when things become too much to contain. the best thing is to start rereading and rerereading. i have underutilised them. i think everyone does. the brain has a strange way of processing stuff. ifyou just keep going at something hard. it will eventually make sense of it and it gets good faster. so by the third or fourth reading i can easily understand most of it better than i did the first time.
when learning something which needs practice like learning a new language, olympiad number theory, combinatorics, c++, bike riding, writing. the best thing to do is to repeat it a hundred times. it gets me past the (this sucks cause i am bad at it to i can become very good at it)
i experimented this with speedmath (thanks to Matiks), c++, full stack dev and currently doing this with deepML.