One argument against 100% test coverage that I've never heard articulated in my decade career as a professional software engineer is that we need to learn to just trust trustworthy software engineers. Yes, there will sometimes be mistakes because we're human. But I'm convinced 100% test coverage is a form of extremism. Once an engineer has proven their ability, and as long as they continue to prove their ability, then a test suite becomes a new kind of tool: not one to prove to management or stakeholders that the code works, but one to prove to the senior engineers that a certain spot of the code works as they intended it.