I think this makes sense. We need to distinguish between something like "obvious current comparative advantage" and "less obvious potential comparative advantage." In practice, the heuristic "stick to your comparative advantage" may optimize excessively for the former at the expense of the latter.