I'm a huge proponent of testing. I also believe the 20/80 rule seems to apply here. You gain enormous benefits from just a few tests, and I feel that benefit reduces as you head towards unrealistic goals of 100% coverage.
Either you're writing tests or you're manually running the software (manual testing) or you're letting your users test your software in production. It might be that all of those might be the correct option as some point. However, in my experience, success in shipping code that works and finding and fixing bugs takes considerably less time when there are at least some automated tests available.
Either you're writing tests or you're manually running the software (manual testing) or you're letting your users test your software in production. It might be that all of those might be the correct option as some point. However, in my experience, success in shipping code that works and finding and fixing bugs takes considerably less time when there are at least some automated tests available.