Regarding the FizzBuzz myth, I think it's selection bias that makes us think it's just a myth. After all, everyone I work with would find the FizzBuzz problem child's play to the point where we, like you, don't trust that it's meaningful.
But having recently had the experience of going to a local job fair with the goal of looking for students to fill internships, I met with over 70 individuals that day. While we didn't administer the FizzBuzz test, I'm confident that less than a handful could have completed it during the interview process.
This is by no means concrete proof. But, I think for many of us, we suffer from selection bias, because even the worse people we work with are light years ahead of what the great majority of "programmers" can do.
Even if the stories are taken literally ("199 out of 200 applicants can't program!"), that number is itself inflated by selection bias the other direction - people who can't program don't get the job and apply at the next place. It's obviously not going to be reflective of the programmers on the other side of that wall.
But having recently had the experience of going to a local job fair with the goal of looking for students to fill internships, I met with over 70 individuals that day. While we didn't administer the FizzBuzz test, I'm confident that less than a handful could have completed it during the interview process.
This is by no means concrete proof. But, I think for many of us, we suffer from selection bias, because even the worse people we work with are light years ahead of what the great majority of "programmers" can do.