I think you're overlooking the idea that tablets only have to be better at one or two things in order for people to adopt them. Once you have one, the additional functions don't have to be objectively better in order to displace single function devices because they have two unassailable advantages - zero additional hardware cost, and zero mass.
For example - just being able to browser, keep up with light email, and read ebooks is enough for me to carry my iPad in a lot of situations. I used to carry a backpack full of books and a laptop to get the same functionality. Once I'm carrying the iPad, the fact that it can become a music keyboard for another $5 is just a bonus. I was never going to buy and carry a small portable keyboard. Likewise the GPS and maps. I wouldn't have bought and carried an independent device for that. For specialists, dedicated devices will better for some time to come, but for most people that doesn't matter - the tablet is enabling activities that were simply not practical before.
As far as UI conventions go, most non-geeks just aren't very good with the desktop metaphor anyway. Sure there are experts and for programmers or digital content creators by profession it pays to become one, but most people just aren't.
My 87 year old father (who's last computer was analog!) was able to take my iPad out of my hands and operate the browser without me even telling him what the device was let alone explaining UI metaphors.
I think the key is that tablets can work in situations where netbooks etc. simply can't, and they can make functionality ubiquitous where a collection of special purpose devices would be too much overhead.
And then it's not going to take long for them to evolve to outpace the development of independent specialist hardware because all those separate makers don't stand a chance of putting the same level of resources into the development.
And for all the things that need custom sensors, why not simply have them be extensions of a tablet?
Take something really specialized - e.g. a metal detector. Why not make that into just a Bluetooth sensor that communicated with an iPhone? Setting could be adjusted on the touch screen, but more importantly the phone could log the output, put it on the map with GPS, and upload it to the cloud. All this extra capability, and the sensor could probably be lower cost than a standalone device.
So once tablets enter the equation, previously devices can become cheaper and more capable.
We aren't going to see tablets fall to standalone devices, altough other hardware isn't necessarily going away either.