> Learning from failures becomes cheaper than having to predict every possible failure mode up front. But this assertion is wrong. There's a minimum price of f

> Learning from failures becomes cheaper than having to predict every possible failure mode up front. But this assertion is wrong. There's a minimum price of failure in space. To have just a bare minimum of safety you must understand at least some failure modes of things you're sending into space. That requires testing and qualification. It doesn't matter if our hypothetical drill costs $10 or $100, to even think of launching it in the first place requires thousands of dollars in qualification for the environment. At that point the difference in launch cost is negligible to the point of unimportant. Safety regulations are written in blood. The Apollo I disaster occurred in part because the behavior of nylon (Velcro and cargo netting)[0] in the crew cabin in a pure oxygen environment was not understood. A material that was perfectly safe in an Earth-like environment had a catastrophic failure mode in a pure oxygen at high pressure like during the "plugs out" test being conducted. You can't just throw random crap into enclosed environments to tear them out in situ. That's how to write uncomfortable letters to next of kin. Real life isn't Kerbal Space Program. You don't respawn or reload a save game if your ship explodes. [0] https://www.space.com/14379-apollo1-fire-space-capsule-safety-improvements.html https://www.space.com/14379-apollo1-fire-space-capsule-safet...