I have to ask a buddy at work who has 2080 Ti, but I don't think he uses RTX very often. If that's the case, it doesn't inspire confidence, regardless of which API is being used to deliver it. I think that it's too immature of a technology to start wanting it on everything and calling it a con when it's missing. AMD even said themselves that they'd rather pursue an option that doesn't require dedicated hardware which makes a lot more sense, at least to me.
Personally, I think the tensor cores made a lot more sense than adding RT.