It’s a legit risk to be basing anything on a Google product or service. Know too many stories of people or companies getting burned when Google wakes up one morning and decides to abandon another product.
YTM (last time I checked) lacks feature parity with its predecessor, has a questionable migration path, and seems more interested in music videos that you can incidentally play as songs. It integrates messily with its host platform (hn comments on the last thread I read about it indicated that Google's algorithms were inadvertantly tagging unrelated content as 'music'), and dilutes focus on the prior project's core feature-set.
Just because another product's abstract contains some of the product I'm using's core doesn't mean I want to switch over to it because Google's got a short attention span, and wants to reimplement the product I'm using in the most approximate manner possible.