[TeMPOraL]

That really depends on the context. I can think of many discussions in which "imagine the Earth is actually flat" would be an interesting counterfactual and a legitimate start of a useful line of reasoning.

But even if (imagine...) I was dealing with a flat-earther, I'd probably listen just to respect someone being able to work with counterfactuals in the first place. There may be something interesting on the other end, perhaps something I could help them rethink. And, if it was ultimately a waste, and then if I kept hitting on "imagine Earth is flat" people, I'd probably just short-circuit my response to "Is this about argument $X for idea $Y? I've heard it before many times, it doesn't work because $Z.".

But, funny thing, I don't think I've ever met a person capable of discussing counterfactuals who also strongly believed in something completely stupid. In practice, being able to work with counterfactuals during conversations - particularly assuming, for a moment, something they disagree with - has been an effective litmus test for me. If they can handle it, we'll have a productive conversation. If they can't, there's no point in talking to them.