That said, WiFi's not a panacea. Technologically it's subject to interference, the protocols allow for NO control over the handset (at present), and its range is limited. Speaking from experience running several large muni WiFi networks, the protocol does not handle power asymmetry well - your phone's WiFi needs to be roughly as strong as the AP or performance is terrible. Compare this to licensed cellular technologies which can pull in a tiny weak handset signal from miles away.
From a business perspective, there's a breakpoint between licensed and unlicensed use - at the point that WiFi's shortcomings have less impact than the cost of deploying licensed spectrum, it gets used. There's plenty of talk about small cells in the industry and I think we'll see a hybrid deployment style emerge over the next few years, with WiFi handling the densest of deployments. Carriers want to offer plentiful data but they want to do it in a way they can control, for good reasons - people are paying them and expecting a service in exchange.
From a good-of-all-mankind perspective, what I'd like to see happen is a transition to all-IP on the handset, seamlessly decoupling every feature from the cellular network. Apple showed the potential here with iMessage, which works well if everyone's on an iPhone. SMS and voice themselves need to be carried over IP in a device-independent manner. When I can send an SMS from my Android phone to my iPhone over IP, and call a POTS number over IP, I'm truly carrier-independent and effectively paying them to be a WISP. That gives me (and the market) opportunities to build a service tailored to my needs.