That's just what commercials are like.

Ah. Well, I'm glad I don't watch television, so I'm not exposed to that.

Side note: this pattern is apparent in LLMs and image generation AI. Some models won't do what you want at first, but if you yell at them or threaten them, they will.

I've seen even official instructions say that you have to repeat yourself if it's something important. Makes sense in a way, if each input results in some sort of weight, then double the instructions will have double the weight.