Looking into optical computers. If we want personal AI, or AI in a human formfactor, silicon is almost certainly not feasible. The technology is some years away from the early adopter phase, I'd say mid to late 2019 at the earliest, but I know jack squat, and it's difficult to find up to date information on this subject.
Anyway, what this means is that it's mostly pointless to develop a replacement to GNU/Linux at this point in time. G/L will need to be ported to RISC-V fast and get some decent development for the system there, because this will be a slight window to get the average person, not just in the highly developed Western countries, but potentially across the world, into GNU/Linux. Tech companies are already investing heavily into architecture, but no one but Google with Fuschia OS has anything already developed to leverage the technology. It's inconceivable that not one of those heavy hitters is thinking of the consumer personal computer market in addition to whatever else they see potential in RISC-V. With Microsoft dissolving the Windows division, it's unlikely that there will be a port of Windows to RISC-V. Which, all of this means, is that companies are going to need to rely on UNIX and UNIX-likes to push out market-ready desktop operating systems.
Of course there's still this stigma surrounding Linux that it's not functional for everyday just werks desktop computing tasks. If I'm correct about the above, then I think we'll either see investment in the already existing corporate distributions like Ubuntu or Red Hat, or each company is going to hack their own distribution together---they might even collaborate and come up with a standard. The benefit of this move would be that the companies could more easily acquire support from device manufacturers for drivers and software developers for ports (such as games or heavy-duty applications for professionals like CADs). Yet the state of Linux isn't looking too good, given that something as basic as the Desktop Environment is still buggy and fractured. All of this is pure speculation, so dwelling on it is not very productive. What is important is that GNU/Linux will be here to stay, but it won't be able to extend itself much further unless this opportunity is seized.
Meanwhile, we have the wait for optical computers, which will be magnitudes faster than electrical computers, and so more powerful/efficient/cost-effective. As soon as this comes out, developers should be rushing to get the foundation down for the next generation of "personal" computers.