User-side tests are the only tests that really matter.
Everything else you do is there just to reduce the odds of users tests catching anything. But you don't get any certainty before that step... that happens after your software is on production and people depend on it.
(Of course, that's a worldview that can be either very beneficial or incredibly harmful depending on what you are creating. It's not good to see it applied to bridges, but I believe the OP did it in jest.)
Everything else you do is there just to reduce the odds of users tests catching anything. But you don't get any certainty before that step... that happens after your software is on production and people depend on it.
(Of course, that's a worldview that can be either very beneficial or incredibly harmful depending on what you are creating. It's not good to see it applied to bridges, but I believe the OP did it in jest.)