There isn't a huge difference between "laptop" parts and desktop parts. Mobile GPUs usually just lag behind the desktop versions by either a process shrink or a number of function units (as long as the work is dynamically dispatched to the units present that doesn't take loads of work to scale). The CPU package are also similar. The Mac Pro has NUMA issues but NUMA issues are going to show up inside the CPU packages in a couple of iterations.
I was thinking long term though, not immediately (or just the next couple of years).
At first, I wouldn't expect much, if any issues, but as time progresses, this is likely to change (i.e. Apple ceases to use ATI or nVidia GPU's, in favor of more powerful IGP's for cost reduction for example, assuming they're up to the performance level Apple sets as their target). And say, in 2 - 3 architecture changes (say 6 - 8 years), NUMA issues will likely surface. There's also the possibility of ARM moving into the laptop domain in the not too distant future, and Apple makes the decision to switch to ARM exclusively (once the 64bit versions reach the market for example).
The other major issue being overlooked is that laptops will have PCI-e expansion via Thunderbolt. An increasingly diverse set of "stuff" will be attached to these laptops. Way too much attention here is being paid to form rather than function. "boxes with slots" are not necessarily the top of the food chain going forward.
In general, I don't disagree.
I only formed the post in terms of the MP, as that's the section this thread is located (i.e. presumes the people interested, are curious about boxes with slots having a future in the Hackintosh project or not).
For the rest of the world that wants laptops, there will be a future so long as people can figure out a way around any future security/validation implementations to determine if the system is an Apple product or not.
There is a bigger threat that Apple will go back to ROM dongle approaches once the overall PC market switches over to EFI boards that don't hide behind "BIOS" capability adapters anymore.
It's quite possible they will resort to a hardware implementation.
But if a software solution can be done that doesn't actually steal Apple's IP (i.e. just bypass the hardware implementation, not make a dump of it's contents and try to sell it), it's likely non-commercial Hackintosh systems will still be around (don't want to get into the EULA issues between different nations, and assumes the implementation is not being used to generate a profit).