Makes one wonder, as well, if they can get it done in a cell phone that costs $600, why not in a DSLR that costs 3-20 times as much.
Indeed an intriguing contrast.
Some pessimistic options:
(a) The far smaller sensors of phone-camera modules make it easier to move the photosite signals at the high rates needed, due to shorter signal paths.
b) Designs that support higher read-rates compromise on still image quality (e.g. dynamic range), a trade-off that is more acceptable in phones than in SLRs ... at least amongst customers who think that the most important review of a new SLR is the one from DXO!
Some more hopeful thoughts:
(c) SlR sensors and processing chips have been designed by and /or for companies with a heavy still-photography priority, who have simply not put as much effort into video support ... yet.
(d) Companies like Canon and Nikon are tied to their separate, proprietary, in-house processing chip designs, that cannot keep up with the technological progress in the far larger and far more vigorously competitive world of ARM processors and UNIX-based operating systems. The OS options include not only Android (which ironically, was originally developed with cameras in mind, before Google bought it) but other Linux-based options, and derivatives of Free BSD, like Darwin, the open source guts of iOS. This too could change: Samsung and Nikon have tested the Android waters in some compacts.
P. S. in case I worried anyone, I certainly do not mean that SLRs should adopt a phone's style of user interface or apps; just using those core technologies as a platform, "under the hood".