Perhaps the author assumes that a child with a deep interest in rust Unicode edge cases has likely been on the internet before, and may well have been exposed to the existence of sex?
On the other hand, if anyone has a good way to explain variable-length integer encodings with constant-complexity backtracking to a child, please let me know.
Every sentence starts with a capital letter, and ends with a full stop. Now, imagine that every sentence is at most four words long. It might be one, or two, or three, but it won't have five words.
Imagine you drop your finger randomly on a word. How can you find the start of the sentence it's in?
(After this, if the child were familiar with binary, I'd show the actual representation of UTF-8, perhaps colour-coded. It's really quite intuitive. No need to go for the abstract straight away: if the child can generalise, they can generalise, and if not, there's no point making it artificially confusing.)