Until you ship code that never fails in its primary function, you'll never get the point of TDD.
But the fiat market loves buggy code, so that's what it gets.
Discussion
bro, I'm in this shit for over 25 years.
I shipped so many fuckups you wouldn't believe 😂
And most of them were extensivelly tested. Automation and manual.
The truth is, that sometimes shit just hit the fan
Dude one thing I learned early on is to never trust someone who says how many years of experience they have. Cause the biggest morons I've ever met in the powergen industry had the biggest "years of experience" stat on their lapel.
I'm always going to look down on code that has exactly zero tests and assume it doesn't work. Cause most likely the developer doesn't even know what it actually does. Just what they hope it does.
I don't TDD because if your reqs and both code change too fast it's waste of time.
I'm deeply into functional programming and I do have "perfect" functions. And a test suites for them. Never worte them upfront.
I write a (hopefully pure) function for some data transformation taks. Test my program, then I write a test which makes sure even all corner cases are covered. And fix my function. Not TDD but still well tested. But only if I'm sure the function will be used.