The response is the same now as it has always been in the past: YC knows it can't predict in advance which companies could be home runs, so its approach is to "bet on the field" (which, thanks to power-law returns, actually works in startup investing in a way that it doesn't in horse race betting).
Remember, Airbnb and Dropbox were roundly panned when they first appeared out of YC, and in both cases were very nearly not accepted at all (Dropbox was rejected on the first application. Airbnb was accepted even though YC thought their home-sharing business was stupid and would most likely fail. YC accepted them because they were impressed that Airbnb had bootstrapped themselves by selling breakfast cereal).
Plenty of other companies that looked like winners early on have failed, and others that looked ridiculous have gone on to become very successful.
But nobody knew which would be which at demo day.
To understand more about YC's m.o., read these PG essays: