While I agree with your admiration of the versatility and power of CJK languages, I don't think the particular aspect you're commenting on is specific to those. You see this in many western languages too. Here's an example in English. Take the word "pot". It can be a container, it can be a plant, it can be a drug, it can be a dish (pot roast), it can be an electronic component (short for potentiometer). Things can "go to pot". Someone might be told to "shit or get off the pot" when they are indecisive. The meaning depends on context. Some of these meanings are related, some are not. Yet it's all the same word. It has some meanings that are shared across different anglophone cultures, and some that are specific to one or several. There are many other words like that, and you find things like this in almost every language family.

Can you explain better?

There are a lot of false pairs in Latin based/Romance languages.

As an example if you can read the sequence b-u-r-r-o you get "burro", that means butter in Italian but ass/donkey in Spanish.

That's the key difference between ideograms (based on pictures of combination of pictograms) and sound-based writings (alphabet and the like).

Ideograms have a major drawback, however. To represent most of things you need in daily life, you need thousands of them.

I found their bewilderment a little puzzling as well, but I believe it's due to the simplification of the last character. If you watch the older Canto/Traditional version of the video, the Japanese folk get the ramen part immediately.

Japanese uses an alternate simplification (麺) for the sense of noodles, and doesn't merge it with 面. Generally you wouldn't expect Japanese person to have exceptional difficulty recognizing 拉麺, though.