Mobile has shown the need for new systems of dealing with identity, application management and sync. Essentially, you need 'cloud services' to get the most out of mobile but anything you buy into is typically hampered in some way (iCloud, Google etc). What would be great is if everyone had their own place in the cloud which could run basic infrastructure (e.g mail, contacts, calendar) and then new applications could be installed there to provide additional services (an analogy might be that you MITM yourself). To do this well involves reassessing how we build software for distributed systems. I'm working on the underlying tools to make this possible - http://nymote.org
its interesting to look at keystone components in mobile. AFAIK it looks as though identity is the predominant keystone API . if you own identity then you are leveraged across both mobile apps and web.
I'm not so sure identity in that sense is that important. It's important to advertisers for sure, but where's the value to Jane or Joe consumer beyond being able to claim a piece of cloud data as her or his own?
The mistake, in my opinion, is to insist on connecting identities across devices, and across meat space and cyberspace. I don't have any answers, but do know that Kim Cameron's seven laws of identity[1] are the best story I've heard on this.
I'm not sure I should fear or applaud Apple's TouchID. They just made a massive Sapphire investment: how long until every Apple Device comes with TouchID? How long until third-party devs can target it as an API?
Apple's fingerprint integration is convenient enough to be dangerous: once you enable it, it is hard to go back.
Identity is key, but whoever controls the identity with a credit card connected to it...that's a useful one.
(Although actually to take a step back...owning the credit card channel still leaves the field wide open for the billions of people who don't have credit cards)