I'm not sure if it is that what the canon guy meant.
I don't think we will see a Canon EVIL-Camera in 2012. And I also think that some of the systems that are on the market now, won't be developed any further in 5 years; there are just too many at the moment. And then there are some problems that can't be solved with todays approach by most manufacturers.
Camera-size means also includes lens size. So for the image quality (and a shallow depth of field) a big sensor is needed, which means big lenses. Just look at a NEX with a telephoto lens, the size of the body is becoming less important, because it's the lens that makes up the bulk. Then there is the handling issue for serious work with a camera. I think most people would get a cramp in their hands, if they have to cover a sport event for 2 hours with a NEX+telephoto-lens.
On the other hand, these systems are nice to go hiking; just take it with you with a 35mm or someting like that, and you have a powerfull little camera. But for that I could also take a Fujifilm X100 with me, or a future Canon G-series.
If we go down in the sensor size, the image quality decreases, and it will do so even with better sensors in the future. Diffraction will remain a problem with small sensors. The DOF-problem could be solved with software (like other shortcomings of the lenses, like it is done more and more), but I'm note sure if this will be accepted by all users.
Downsizing is not easy in the camerabusiness. I think there is a market for small sensor EVIL-Systems, but this market is quite crowded already (if we consider 4/3 as small). People looking for image quality will at least buy a 4/3 sensor-camera, and I think this won't change much (as mentioned diffraction and DOF won't improve...). People who do not care that much about IQ, will buy a smaller camera, be it an ILC or a fixed lens camera. But I see the advantage of a fixed lens camera, because these people do generally not like to think about focal length, and which lens to take with them.
I think there is room for an APS-C or even FF EVIL-camera in the rangefinder style. SONY is close to this, but the ergonomics of the NEX sucks. The camera should offer very good IQ, and low light capabilities. Some nice primes, and a quality standard-zoom would be nice; something like the soon to be announed Fujifilm Pro X1, just a bit cheaper.
If I were Canon, I would take some time to find the ideal sensor-size for a future EVIL-System. It is not an easy task...