Mantle could be great, what is disturbing about it is it could lead to even more fractionalism or, if Nvidia's counter doesn't get adopted by a similar number of developers, the death of Nvidia in the PC gaming space.
If Nvidia starts pushing their equivalent of Mantle, we'll start seeing GPU purchases becoming much more complicated and more of a guessing game. A high end card is no longer high end in all games, just the ones that support it's API. A customer making a decision about a card that they want to last 2 years might get completely screwed if all of the games their interested in that come out after a year end up using the other company's API.
If you buy a high end card, you expect it to run games well, not some games well and others like a mid range card from the other side. The only way this is good for consumers in general is if both sides use the same API that has better access to the graphics hardware. Developers are not going to code for Mantle, DirectX and NvAPI.
Port optimization is the best thing that can happen to PC gamers. We should rejoice, not mourn. We would not need a super expensive and super powerful gpu to do the same job that a console can do more efficiently. This is great news!
But, it also sounds too good to be true. As I said in another thread, I am withholding judgement until I see some real results.
Let's all hope that AMD can pull it off this time as being the leader of change for PC gaming, instead of a laggard.