For most other products, Google prefers to frustrate some users (cf. https://plus.google.com/117598418867899518106/posts/Zr4vhv6p...) by disabling all logos' links while users get familiar with clicking the products' titles to go to the products' home. But once users will be used to this and all those products will be fully integrated with Google Search (cf. my previous comment), then they'll add a link to the all those Google logos in products' header because they won't be lost in a service (Google Search) that is independent to the product they come from (Gmail / Docs / Reader...). If they succeed, I think all users will benefit from this new navigation concept.
I don't like the current situation (unusable product logo) but it's certainly temporary: they've reverted to the dark navigation bar a few days after the release of the big toolbelt navigation menu, so they're probably still looking for a good solution.
In the meantime, it's a difficult POLA dilemma (http://en.wikipedia.org/wiki/Principle_of_least_astonishment): 50% of Gmail users might expect to be sent to Google's homepage by clicking the Google logo in Gmail, while the other users might expect to refresh Gmail or go back the inbox by clicking the only logo in Gmail's header. Where should Gmail's Google logo point to?
My point is, all this requires me to think. But before I could just swing my mouse into the top left corner, click "Gmail" and always end up at my inbox, no matter where I was in the application. This consistency has gone.
I should have phrased my question: "What possible advantage does this offer, to the user?"