When studying Japanese about 10 years ago, one of the big realizations was to notice that I learn much faster if I fail fast and check the answer, as opposed to try to think about the word for over 10 seconds.
Some of my friends studied more systematically all readings of a Kanji, but I mostly focused on real world words and wrote down everything I encountered. Studying only about 30 mins every day, I could study about 100-200 words.
I realized later that I had learned the same thing when I was studying the board game Go. In about 2 years I went to 3-5 kyuu level, but it was because I played a lot of fast games, with short reviews afterwards.
I have been using that to study other things also, but I think it has become kind of motto for my working style.
If I need to do something, I draft something fast to see if it might even work or not, prepared to throw it away. To be able to do that I do spend long time learning and optimizing shortcuts and other practical things. That allows me to get feedback fast.
This same advice is also very sound for business environment. You should aim to get something out to real people fast. See their reaction and adjust. Yet, often the release schedules are months or years.
What if you could build a very simple prototype in a day or a week, and ask someone for real feedback before committing to do it for long term?
Fast feedback loop is very important. On the other hand, so is good feedback. For somethings you can do it with software, for other you can do it with money, other times co-workers and friends can give you the feedback. Sometimes you can only get feedback from yourself.
But more importantly, my advice is to try to become fast first. If you write a lot, spend time to become faster at typing and every time you save few minutes here and there. If you write code, learn all the shortcuts, refactoring concepts, try to imagine ways you can do tedious editing with single keypress or command, and practice them actively. Make it so that every time you save a file, you see the changes.
When you are fast, you can make more mistakes faster, and the accuracy and correctness come only with experience. If you are slow, you cannot make enough mistakes to evolve.