In Fluent Forever, Gabe Wyner looks at the struggle for different language speakers when speaking a foreign tongue with certain sounds. He looks specifically at L vs R for Japanese speakers. What he says is that these sounds appear along a continuum spoken slightly different by everyone. In American households, the words containing the R sound and those containing the L sounds are sufficiently far apart for the English speakers to learn to tell the difference. In Japanese households they aren't. The sounds are closer to the middle and they correctly group them. The Japanese R doesn't sound quite like the English R, and it doesn't sound quite like the L and English-tuned brains can't differentiate between the sounds. You literally can't hear the difference. Carnegie Mellon did a study where the test Japanese candidates tried to tell the difference between 'Rock' and 'Lock'. Even after practise, they were struggling badly. A second group of candidates received instant feedback each time the got it wrong and suddenly began to pick it up VERY quickly. Instant Feedback is very powerful even for really hard rewiring.
Sometimes instant feedback isn't possible. Investing is a good example. If you buy a business and its stock price goes up or down, there is VERY little information about whether or not you made a good decision. Investing is a 'Fat Tailed' endeavour. Risks lie hiding in waiting and a few big events can make all the difference. There is no instant feedback. For the long term investor you have to wait 5-10 years to know whether you have been making good decisions, and even then it is very hard to strip out how much of that was luck and how much is skill. That doesn't stop you from thinking what you are doing is super smart because of the false instant feedback of stocks going up (or giving up on a smart approach because they are under performing).
Knowing when to look for instant feedback, and when to look at the big picture and use judgement when learning is the challenge if you are up for it.