I don't think he's saying all bias is bad. But bias finds support too easily, even when it's not accurate (prejudice, nationalism, etc). And that reasoning is easily repeated by uninformed people which just adds more noise to an already complex discussion.
When you're trying to have a rational discussion about which factors contribute to certain effects, it's better to exhaust all possible theories before attempting to prove that "some people are just better than others."
I guess I'm more in favor of putting all the theories on the table rather than deciding in advance which ones are permissible to hold. Let the false ones fall out because they are false, not because they cannot be discussed. What I deny is that there are true ideas so unpalatable that they must be taboo. I think this is what Barrkel is suggesting is necessary in the case of race.
When you're trying to have a rational discussion about which factors contribute to certain effects, it's better to exhaust all possible theories before attempting to prove that "some people are just better than others."