More about the article:
One of the best parts of touch based interfaces is they can change with context and can be customized/improved/fixed for the cost of changing the software. No need to ship a physical 2.0 for your POS when you want to change the button layout.
Also, with these blades, are you expected to carry around more than one and swap them out based on the activity? Then use the onscreen keyboard when you need to do any text input?
The thing is a computer. It has a USB port. One can attach whatever input devices they want. The only difference about a blade is that this doubles as a screen cover. In the medical and fast food examples he gives, the machines are stationary (and you probably want them to stay that way).
I'm all for dedicated input devices or customized interfaces to make jobs easier, but that's hardly a new idea.
I think they could have generated more interest and more money just selling them with more designs and sports team branding (The Surface is already used by the NFL, so some sort of relationship between them and MS already exists).
As for the knobs and sliders, I agree it would look 'cool' to see real LEDs light up for your fader or the active loops on your sequencer or whatever, but the reason actual musicians like physical knobs and sliders is because of the tactile response they provide. Thats why Serato's vynil controllers are still popular with a lot of people (http://serato.com/controlvinyl). With a touch keyboard like the one that the surface has you get the worst of both worlds. The (lack of) flexibilty of a hardware device and the (lack of) tactile response of a touchscreen.