thats true, if there were one sigle "coder" who was cloistered away from the team and came back when s/he claimed the requirements were met and the task was complete. Whereas, in FDA (and presumably FAA/DoD) regulated designs, a massive team is following a rigid SDLC including code reviews for best practices.
Now, technically the FDA doesn't force that, all they specify is that some standard procedure must be followed, and written evidence must be produced upon demand showing that the procedure is followed. So you could contrive an example where a cowboy coder made his own sqrt() and completely broke it, and that somehow led to the death or injury to a patient, or an elevator going into free-fall. or a fuze detonating. sure. Is that likely to happen? absolutely not.
> a massive team is following a rigid SDLC including code reviews for best practices
Regardless of how thoroughly a system is planned in advance, programmers are not automatons who mindlessly fill in the last few blanks. They must make choices about the code they write, and that is absolutely a part of the overall design. If that code fails a code review, well, then it gets rewritten, but that's ancillary to my point: programming is design (albeit at a lower level than the overall system design).