[caconym_]

This comment is nonsensical.

Just like anything else, assuming SpaceX is able to build a functioning example of their Starship concept, they'll fly demonstration flights and iterate on the design and perform safety analyses until they're satisfied that it meets (at least) industry-standard safety margins (or until they see that it can't). I don't understand what you (apparently) think is fundamentally different about this particular technology vs. all the other machines humans put their trust in that are theoretically capable of killing them if something goes wrong.