(Nothing new here if you've read my posts before but I just want to keep drilling these points since while subtle and even low level, they do impact user experience and first impressions in a big way)
To me this is like Windows 1.0 release, so the comparison would be what was the 1.0 of Apple and Google.
At this point the main concerns are:
Getting something out to show that MS is atleast in the game. Alternative strategy of "lets do 5 versions" before 1.0 (at which point the real user feedback is had) like I heard was the story for WPF didn't seem to go that well, it's still taken ~5 versions for WPF to get the basic features it needed in v2 ...**
the other main concern, as with WPF, is: are the foundations, fundamentals, building blocks etc solid? eg. If you wanted to do a musical game or process the camera or other data, can you do it at a latency that is better than the competition and if not, where is the problem? Technically I know audio at 1 ms latency without glitches regardless of the CPU use, all without any exclusive access to only single app at a time is doable, MS just needs to go to the bank and buy the experts and put it in the platform if they want to be taken seriously by serious developers. This type of stuff is the stuff that you design and think first when you start designing an operating system, APIs and spec'in hardware because getting low latency IO and low power use will drive everything. I haven't heard any "ooh aaah Windows 8 tablets are technically groundbreaking in this and that", so the only pull is, will there be such large amount of users that mediocre will be ok? Well as a user, I'm looking for those technically amazing apps that only Windows 8 tablet can do, and if Microsoft doesn't have a solid highly responsive platform suitable for low latency ( milliseconds from user touch or camera input to sound/visual response) then there's not going to be those apps that are fun to use, since you need that touch -> audio feedback loop for "instant gratification", which is max 15 milliseconds from touch to sound or visual action. If you don't get bleeding edge personalities/thought leaders excited, then getting the pull to get mass market excited could prove challenging.
Of course there's a risk that I'm a bit biased but I have superb track record for picking winning technologies. I haven't yet decided on if "Modern" is a flop but if app start time and IO latency aren't high priority well you might see another Kinect here, which is another thing where the end to end latency is absolutely the most critical element. Granted, Microsoft can't control LCD DSP latency*, but they can control latencies in this new platform. (*well they could, by giving out Xbox compatible logos for TV manufacturers and speccing out latency requirement, engineers have told me that LCD DSP processing could be done in parallel rather than sequentially in hardware, so there's no technical reason why LCD's have lag beside the pixel response time).
** (I think the WPF-directx-etc air space/interop things, app load time and such should have been thought about when Avalon was still just a thought on paper - atleast planned ahead if not fully implemented, these very obvious customer requests seemed to come as a surprise to MS and only SSDs have saved WPF and VS201x from irrelevancy for now)