>The "many pronunciations for one kanji" thing isn't a problem in practice, because 1) most kanji have _1_ on'yomi (Chinese) and _1_ kun'yomi (native) pronunciation, with it being obvious which one to choose from context, especially because when it's kun'yomi, the kanji are suffixed with hiragana to denote its correct pronunciation. There are however problematic kanji, such as the one you note, because there are several things it could mean, but this isn't really any different from _all_ languages.
What the heck...? It's most definitely a problem in practice! That problem becomes pretty obvious once you get a vocabulary of around 10k ~ 15k words and you read both fiction and non-fiction books. What you find is that every time you encounter an unfamiliar word you only really have a best guess at how to read it and you actually can't be 100% sure. This has repeatedly been my experience. I spent some time thinking about if I could shortcut this process by instead simply learning all the "exceptions to the rule" but what I discovered is that the volume of them is too large for that to be a tractable approach.
What the heck...? It's most definitely a problem in practice! That problem becomes pretty obvious once you get a vocabulary of around 10k ~ 15k words and you read both fiction and non-fiction books. What you find is that every time you encounter an unfamiliar word you only really have a best guess at how to read it and you actually can't be 100% sure. This has repeatedly been my experience. I spent some time thinking about if I could shortcut this process by instead simply learning all the "exceptions to the rule" but what I discovered is that the volume of them is too large for that to be a tractable approach.