I mean, think of it this way: it comes down to how often you come across words in any language including English (even in ENG: you may forget how to spell words correctly if you don’t use or encounter them often), kind of the same logic with Kanji: a Japanese person doesn’t know all Kanji in the same way English speakers doesn’t know every single word that exists in ENG.
There are over 5000 Kanji but only about half of that is used in Japanese or closer to 2136 while the remainder consist of ones only present within technical jargon (medicine, science, politics, etc.). or certain Kanji only has limited uses in some words (but mostly written in kana). That is also accounting for grammar being “straight forward” more than English or Euro languages.
The “real” hard part is numerous readings (depending whether it’s paired with kana or another kanji, reflected from kunyomi & onyomi plus nanori when applied in people’s names). What I hate about most online translators is that it often gets lost in translation (like words used in the wrong context but on their own it’s correct, however not right for the situation or topic at hand).


Depends on what you mean by “similar”. If you’re looking for the right thing, there are often a lot of phonetic hints in kanji.
For example, 矢 and 夫 sound completely different, totally unrelated. These are pretty basic kanji though; I think it’s analogous to how O and Q are totally different, or i and j.
If there’s more complex similarities though, then yeah there can be an implied phonetic relationship. Look at: 同胴洞銅恫桐粡. These can all be pronounced the same: どう (dou). Though, many of them have kunyomi readings that are totally unrelated.