neus
I would say it’s in between. It’s always 100% better to lose an engine in a test where no lives are at risk.
But to compare the loss to how SpaceX loses rockets is not quite the same. SpaceX is built to iterate on it’s rockets in a way it can much easier handle a test failure. Blue origin is much more traditional space development where it can actually be much more impactful when a failure happens. See what stevecrox is saying above.
A lot of buggy releases often aren’t attributed to lack of number of QAs (which helps you find rarer bugs). The simple fact that crazy common bugs any QA probably immediately found and filed, yet was triaged and determined lower priority to other worse bugs the dev team was busy with. Then deemed not stop ship worthy to some exec trying to hit a date.