I completely agree that they need to go away, and that new lens tech needs to appear.
But I think that we are -just now- starting to get to the point where we can have truly excellent mirrorless cameras that actually can replace (and will surpass) DSLRs.
A lot of it has to do with on-chip computational photography. All the camera companies are coming up with custom image processing chips, and these are what makes the difference.
Lag time needs to be in the single-to-mid-double-digit ms range. I don't remember what the number was, but the company that I worked for, did some research, on what the maximum allowable lag time on a viewfinder could be, and I think it was like 50ms. That means that you need to expose the chip, scan the image from it, do the A/D conversion, demosaic, gamma correct, color correct, denoise, vignette correct, distortion manage, remove chromatic aberration (because cheap lenses), and all the rest of the image processing, then, display it on the screen, continuously, within 50ms per frame, while also not draining the battery. Also, pixel density has a big impact on this performance. More pixels, more milliseconds; more battery drainage.
Not a simple task. You can do tricks like scale down the image early on, and eschew some of the finer image processing tasks, but you still need that viewfinder image to be right on the money, and to the standards of a very picky photographer.
Also sensor physical size has a lot to do with the image quality, bokeh, and what-have-you. A big issue with smartphone cameras (and point-and-shoot cameras), is that small sensors mean that there's little control over DoF (Depth of Field). Apple has introduced artificial bokeh, with their software.
Sensor size, and the length of the optical path, are what really drive lens size. They discuss that, in the article. Hasselblad and Phase One (if they still do that stuff) cameras have these monster sensors. They still need big lenses.