吉 is pronounced き↑ち↓ but 不吉 is ふ↑きつ

Because Japanese is a spoken Japanese pizza topped with Chinese writing on half of it but a few bits snuck on the other side, too. Or some better metaphor.

Single kanji -> stand in for whatever the Japanese word is, read it like the Japanese word, probably. Two kanji -> oh shit, maybe, if you're lucky, it's the Chinese reading of both. But sometimes it's not, sometimes it just gets slapped on the Japanese word. And if you're really unlucky with a word, they mix. Which is first, I don't know, you don't know, the Japanese might know, or they might just add the pronunciation right on there. Four kanji -> I dunno, ask a linguist.

Also, thank you for introducing me to that video. Why indeed.

I don't know, but thanks for the great video :D

It's Onyomi (The Chinese-based phonetic way) vs. Kunyomi (the Japanese own phonetic way) of pronunciation.

Like 心 kokoro vs. 心臓 shinzou. The latter in simplified Chinese that this is based off of is 心脏 (Xinzang), which sounds similar.

Commonly, Onyomi is used when multiple kanji are used to describe a single "word" or concept, and Kunyomi is often used when on its own or is a verb with its own trailing character conjugation (okurigana).

Many exceptions apply but I hope this rule of thumb helps you.