If it's that easy to replicate then what's to stop them from copying you and eating your market share once you've launched?
> because customers won't get the idea without a mvp
If that's true then you have to ask yourself if you're actually solving a problem people have or designing a solution in search of a problem. Which then comes back to the GPs advice about speaking to your customers.
> because revealing your idea might reduce your impulse
I guess that's possible but often I see the reverse: engineers spending forever refining their idea because they're scared to release anything until it's perfect. At least by revealing your idea you're committing to a release early and often model that would see actual users on your platform.
> because customers will ask for all kind of features and change your idea into something else, more practical, like "get a job" instead of dreaming in your great idea
Sorry but I don't really understand the point you're making here.
If you're saying customers will ask for more practical features then surely that's a good thing? The opposite of that problem is spending months building something users don't want nor need. Which again, comes back to the GPs advice about speaking to your customers.
If you're concerned that people will tell you to "get a job" then that could happen regardless of whether you announce your project or not. I don't really see how secrecy would affect the opinions of those kind of narrow minded individuals. Plus why should those comments even matter in the first place when they're clearly not relevant to your task at hand?