His argument is mostly sound. To say that all predictions of doom are wrong simply because all prior ones have been is a logical fallacy .

Sounding a warning is still useful even if the apocalypse doesn't happen in specific cases. It's better to be wrong about a lion being in the grass more often than you are right, if you value survival.