I thought it was disappointing initially but, actually, it is brilliant. They already have a stock pile of the previous generation that are as good, if not better, than the competition depending on use case. Now, they released a product that basically costs twice as much with a feature that you can't test and benchmark until long after people buy it. That feature, to top it off, is basically useless because no one plays at 1080 when they are the type that purchases top tier.
So, they will probably have a limited availability because the dies are so large and yields will be poor but they get to completely test the arch and RTX AND make a large profit doing it because they priced it so high. Meanwhile, everyone that doesn't want terrible value to go with high performance will buy all the stock of the previous generation greater than MSRP. Following that, 10n, or 7n, whatever they end up using will be ready and they can release this same product on the node shrink and gain even more performance without having to develop another arch. Plus, the enhancements they will get by switching nodes and some enhancements they learn from this beta test will allow the RTX tech to be useful when 3000 series can go.
They will get to sell out all their old stock without having to drop prices, sell all the 2000 series at insane prices, and beta test RTX all in one go.