I think this would certainly be an interesting project - sounds a but like an open source boot to Qt.

A big advantage would be certainly the relatively small amount of work needed to have something running and being able to basically run Qt apps on Android as first class citizens and without the Android bloat & spying crap. That sounds kinda like Firefox OS done right (Firefox OS does not use libhybris but basically gutted down Android with Firefox running fullscreen).

I can also imagine porting my apps to this platform as long as Python 3 is available as they & their deps already run on Android (after a substantial amount of work).

On the other hand I can also see some potential disadvantages and open questions:

• If you want to do something Google is not interested in doing/supporting (or does not want you to do) you will be on your own and might need to write stuff from scratch that is easily possible on normal Linux distros.
• Also any changes in the underlying Android base system might require to carry the resulting patches forever as Google would not be likely to merge them & to make them work with any changes Google does in the future.
• Loosing control of your base system - Google dictates the direction of the platform and the project could not reasonably expect to change it, even if it is heading in a detrimental way. Also note that while this is highly unlikely, Android (minus kernel) is under BSD - Google could pull another Android 3 if it wanted (not releasing source or giving access only to partners).
• Every non-android library you would want to use would need to be rebuild against Bionic - this is sometimes easy, but sometimes not trivial thing due to Bionic being a broken mess. So by default (until someone ports the needed libs) this means loosing any remaining compatibility with non-Qt based software.
• If someone wanted to make this closer to "normal Linux" he would basically need to "unbreak" the system vs modifying normal Linux to be suitable for mobile devices. Still this could be eventually possible if the project is popular (eq. slowly replacing Android stuff and converging back to "normal" Linux).
• What about multitasking and task switching/windowing ?
• How would software distribution be handled ? - The existing Android packaging mechanisms probably could not be used.
• Not normal but embedded Linux - which is a difference many seem to underestimate. (example: try to unpack a tar archive on a stock Android device, not to mention using ssh - good luck ).