You can be wrong on every step of your approximation and still be right in the aggregate. E.g. order of magnitude estimate, where every step is wrong but mistakes cancel out.

Human crews on Mars is just as far fetched as it ever was. Maybe even farther due to Starlink trying to achieve Kessler syndrome by 2050.