> Let's say: if you can just open the manual and start building, then it is trivial.
> We've built many rockets, there are numerous resources about building one, so building rockets is trivial.
Except that you're wrong.
Because it's very common for the first launch of a company's first orbital rocket to fail to make it to orbit. So you can't "just open the manual and start building".
> It is also very common for the first pancake to be a total failure.
I guess, it depends on what you mean by "total failure".
It's very rare for the first pancake to be inedible, or basically anything except a little misshapen. As someone who's primary interested in pancakes is eating them, that's a far cry from "total failure".
But if that is your criteria, there's a very simple and effective solution: ring molds. They cost a couple of dollars on Amazon and guarantee that your pancakes will be perfect circles every time.
In contrast, there is no known way to ensure that a first rocket launch will be a success. If there was, the companies launching them would do it since failed launches are extremely expensive in time, money, and reputation.
Building a new Falcon 9 is trivial - spacex as built a lot already and knows how (or so we assume). However that is only true if you use the existing design as is. Change anything about the design (which we can assume spacex is doing from time to time) makes it non-trivial.
We've built many rockets, there are numerous resources about building one, so building rockets is trivial.