Roman Malov comments on Alignment remains a hard, unsolved problem

[Roman Malov]

but there’s no similar fundamental reason that cognitive oversight

This might not be a similar reason, but there is a fundamental reason described in Deep Deceptiveness