There are several companies working on this, but Canesta is one of the better known start-ups and just landed more funding.
The technology is currently buggy and has a ways to go before being a mass market feature, but I think this is promising technology since it follows several of my rules of consumer electronics . Namely, it replaces moving parts (keys, input buttons) with electronic input and increases portability and integration. In addition, I think data input technology hasn't kept up with data output technology. We have flat panel displays and high resolution microdisplays, but the mechanical QWERTY keyboard is still the basic template for nearly all input.
There will be several consumer acceptance roadblocks, such as the "feedback" you receive from buttons as you type, and the increasingly important ergonomic issues with repetitive stress, but I think by the end of the decade this will be an accepted technology, most likely in applications we have no clue about today.