Have to wait on two things there, see if MS have reversed their trend of having no idea who their demographic is, and seeing if the proposed improvements are any good.
Windows 9, I know they've backtracked and tried to appease desk top users with 8.1, but I wouldn't be surprised if their belief is that time and familiarity will mean they eventually push out another Windows 8 touch/mobile feel because that is where they are reading their market.... incorrectly. Companies that completely screw over their existing user base(desktop/server/laptop) to try and win a market they are ill suited to(mobile) winds me up. Make a mobile OS, sure, but don't try and force your existing user base into a new meh OS, make both. If you want to compete in both markets commit to it, don't try to merge them because that has been a shockingly poor decision. MS just consistently makes these wrong calls of late "lets put the money into kinect and edram and end up 50% down on gpu performance..... for a gaming system". I'd actually like to see them get back on track but they've been this way for so long now.
6 months before Mantle was released I was talking up my hope of PS4/AMD together making a push for openGL(improved) to take over and offer us gaming on any platform we want. Mantle is merely another giant step over what i wanted and thought could happen in the first place.
If Mantle was never released my hope was for DX to effectively die(over time) and give the industry control of the API, not MS, and allow us a choice of Windows/linux/android(further into the future)/even OSX.
So while I'll take a better DX while I'm still stuck on windows for gaming, my wish has always been for a cross platform API. I'd prefer Mantle over openGL as frankly most software tends to do better when it's both highly focused and starts fresh every now and then.
Carmack has spoken repeatedly about openGL driver nightmares, getting extensions approved being a fight to the death(too many parties involved), about the openGL group refusing the chance to start fresh on several occasion, a huge amount of legacy code and many dev's, Carmack included, has asked for a low latency, low overhead and low level access API.
That is really all I want except I also want a industry controlled standard, not an MS locked down one.
I also said when Mantle was announced, probably worst case scenario was it actually made MS commit more resources and stop holding DX back. If DX12 is great and we're still locked to windows, we're still WAY better off than a year ago, presuming DX 12 delivers.
Hell, if gaming wasn't locked to windows and MS wasn't getting, automatically, new Windows sales every year from their locking DX to a new version....... MS might actually put some work into Windows, creating a lightweight uber performance system. With less ability to force people to upgrade they would actually have to focus on features and performance to persuade people they should upgrade.... win win win, win. okay I lost count of the wins but it's good any which way you look at it.
The thing is, so many people talk about Win8 being bad etc., but I remmeber a number of people (LtMatt is the name that comes to mind, on a number of occasions) has said how Win8 is better than Win7. So if Microsoft follow this trend then it stands to reason Win9 should be even better. Sure it's an added expense but while we're stuck using Windows, what's the problem with Microsoft improving it again?
I do agree that Microsoft probably aren't the best people to control the main gaming API. The OpenGL style of control seems to have definite drawbacks also though, so that's probably not the way to go either. So the best option is for each vendor to create and maintain their own. This way it can make best use of their hardware and nobody can block new features being added. This could well divide the market even more between GE and TWIMTBP titles as you'd have to imagine that even if both APIs are supported, the game would likely be designed around one and then have the second/third added in.
To be perfectly honest this outcome wouldn't bother me too much for now as I'd just put AMD cards in 1 PC and Nvidia cards in the other. I appreciate not everyone has or wants 2 PCs though. Some people may be happy to do what I believe LtMatt plans to do (I may be wrong, if so I apologise I'm going from memory) and just buy every Mantle game and avoid any TWIMTBP games (unless they have a benchmark

) or vice-versa. Intel would suffer, but if they want to compete on the gaming front they have more than just this to address!
I think if AMD were going to have Mantle controlled by a neutral party they should've gotten them involved much earlier. The longer it seems to be an AMD API the less likely Nvidia are to go anywhere near it. I think that's quite understandable and reasonable. I wouldn't expect AMD to do any different in reverse.
If Mantle had already supported at least the newest generation of Nvidia cards at this point then I think it could've made real progress. While it's AMD only I think the industry will be considering alternatives such as OpenGL and DirectX 12.
I mean sure, everyone could add in a Mantle render path as well as everything else, or they could pick one that supports all vendors. If the performance improvement is only minor would that be enough of a reason to use an API that only supports some of one vendor's cards? True, DX12 may support even less cards when it's launched, but we don't (do we? 100%?). I agree it's unlikely knowing Microsoft but it may support any hardware that supports DirectX 12 (or close to). That would give it a massive edge, IMHO.
If Microsoft want to add a DX12-like API to XB1 then the current hardware would have to support it surely?
Will this generation of consoles last as long as last gens? It seems they may have worse hardware compared to the PCs of their time than the last gen?
What happens with the next gen? Who will be involved? How will this affect PC gaming?
Basically there are so many hands yet to be played (including Mantle I feel) that we're only guessing. If things were easy to predict you'd expect, well everyone, to make less mistakes!
EDIT: I realise I've used LtMatt in there regarding 2 points. This isn't supposed to be a covert attack on him getting people to say he's wrong. He's just the first name that sprung to mind for both examples. He's that memorable!
EDIT 2: Sorry to Humbug who doesn't like large walls of text, quoting DM made this a double whammy! I didn't realise I'd written so much!