Also, there's the rumored 18c32t Ryzen server chip. So, I'm still not sure what that poster's point was regarding going back to Intel over a $1000+ 12c24t server chip vs Ryzen.
Well, the biggest unknown at the moment is how will Intel cores compare to AMD's, and not just in number (that's pretty straight forward) but in single threaded performance which is a function of IPC and clocks (and HT implementation details possibly).
So far the results of Ryzen 7 were somewhat strange as it's not clear why a CPU which shows lower per thread performance is able to beat an Intel CPU with similar number of cores in multithreaded workloads. There are some hints that Zen's HT implementation may be helping a lot here but it also looks like it's just working on higher clocks there.
For games this is even more important as many (most?) games aren't actually scaling that well beyond four-six threads so chances (and expectations) are that a higher performing per thread Intel CPU would be able to beat Ryzen here - but with the factual data on how Ryzen scales in synthetic benchmarks this expectation has become rather moot.
Right now I fully expect Ryzen to perform on par with Intel in most games (talking about i7 and 7 CPUs at the moment), loosing to Intel in those which aren't doing a lot of work outside of 4-6 threads and winning in those which are able to load 8+ threads - this isn't at all different from the earlier expectation of course but now I expect Ryzen 7 to actually win over a 6900K in those latter games as well which wasn't the case previously, simply because it seems that Ryzen 1800X is able to work on noticeably higher clocks than 6900K in multithreaded workloads.
This will surely put a lot of pressure on Intel to recapture the "halo performance" crown thus I think that if there will be changes they'll come in the form of 6900K and 6950X being updated in one way or another.