I'm not sure there needs to be a 'why', or that the devices need to be useful. As George Carlin put it, "If you nail two things together that have never been nailed together before, some schmuck will buy it."
"FitBit etc. are interesting but seem kinda niche."
That's true, but those types of devices are at least addressing a specific market need that exists now, and would appear to have a huge potential as the technology improves. "Smart" watches or "iWatches" or whatever seem to be a gadget in search of a market.
The term watch itself is really old. Wrist computer might be the new term to define this segment.
As someone said, medical applications are the first to come. For example, a wrist computer (smart watch) will now watch your heart rate, ecg, blood pressure, etc.. and inform it to the smartphone. An app on that smartphone can make the information useful. Thus it is obvious that the company that can integrate more number of sensors on the device will be more successful.
Beyond these, there are certain other interesting things you can do with a wrist computer. For example the motion sensors on the device can monitor your hand movements and count the number of times you smoke per day and will warn you when exceeded a certain limit. With motion sensors, you can also take control your home appliances or smart TVs by simply gesturing your hand in the air. No need of kinects and video cameras.
It is easy to say many applications like this. Implementing them is quite different.
"Watches have been disappointing to me because their functionality is too limited to make them compelling," Muller said. "But we think jewelry could be a platform that's interesting," he said.
I totally agree. The truth is that many people don't wear watches any more. Rather than trying to cram every mobile phone function onto a watch, it's time to rethink what needs to be communicated for what purposes from which wearable devices.