Searching for failure modes is what all good developers do. If this guy is finding ways to make FSD fail that Tesla was not already aware of, they should hire him. The more failure modes are found, the better the fixes will be, and the sooner they are found, the sooner they can be fixed.
Tesla makes no claim that "FSD" is a driverless system yet. They claim that it's good enough that the driver doesn't have to intervene very often. I say more power to anybody who can find more situations where driver intervention is needed, so that Tesla can address them and reach true driverless operation sooner. Hopefully while I am still alive.
If Tesla were to advertise, they should show the failure videos and say, "We are bringing Tesla FSD closer and closer to reality by creating and studying situations where it still falls short, and then developing systems that do not have these failure modes."