I think Intel's value based on performance isn't good compared to Ryzen, but the product isn't inherently flawed, and that combined with AMD's shaky history is enough to keep myself, and lots of people I know from switching. (Right now stories of RMAs for the Risen GCC crash bug taking ages are popping up)
Higher base clock on the 7700k, biggest reason for entertainment alongside:
Higher OC potential (back in the 4790k days 5Ghz on air was something else, for the 7700k not getting 5Ghz is considered "losing the silicon lottery")
Cheapest 7800X motherboard is 100$ more than cheapest 7700k motherboards (which are still quite good) because X299 is a HEDT platform
Most entertainment machines won't benefit from quad channel over dual channel in a meaningful way.
The 7700k is generally regarded as the fastest gaming CPU available, that's why so many benchmarks use it. I'm pretty sure AMD used it to showcase Vega because even if it becomes a bottleneck, no other CPU on the market can be "less of a bottneck"
And I'll add, this is all with my entertainment/gaming hat on.
For a dev machine where I'd expect to work with compiled languages or compiling OSes I'd definitely go with more cores.
Frankly I've never heard of a 6950X overheating and throttling, the lid is soldered on those processors. Sounds like a personal problem.
https://www.bit-tech.net/reviews/tech/amd-ryzen-1800x-and-am...
https://www.pcper.com/reviews/Processors/AMD-Ryzen-7-1800X-R...
> Why the cynism?
Because in performance and efficiency Ryzen is basically equivalent with Haswell-E/Broadwell-E. The sole advantage it has is that the 8C Ryzens are cheaper than the 8C and 10C Intel SKUs (the 6C have been competitive for a long time), and the tradeoff is you are moving to a much more immature platform.
Claims like "oh my Ryzen was so much cooler than my 6950X" set off my bullshit detector real hard because that's not what the people who are actually measuring things are saying. They usually end up being made on the basis of the TDP figures, which AMD's marketing department has massaged significantly downwards from their actual measurements.
Ryzen is basically on par with Haswell, it's not the second coming of Jesus unless you've been in a coma since Bulldozer was released.
Later iterations of Ryzen are undoubtedly going to fix some of the teething issues (segfaults, memory clocks, etc), and probably get the clocks up too. But they will also be going up against Coffee Lake and others - Intel can iterate on their products too.
Consumer Ryzen is basically just Haswell with an AMD sticker on it, and it's just so transparently fake to see all the fanboys cooing over a product they wouldn't even give the time of day to when it was released under an Intel badge. It's an incremental improvement on pricing, but not really that drastic except in a few special cases (the 8C and 10C Intel processors had very steep pricing).
As it's on a newer stepping with a better platform and fully-validated ECC support, Epyc is much more interesting product to me. But it's also a market that doesn't care about throwing money at the problem if that's what it takes to get performance/stability - there are plenty of companies who are not even using COTS hardware let alone off-brand.
Also, bear in mind that "nm" measurements are pure marketing. It's now the spacing between features that determines total density, not feature size. Intel is at least a full node ahead of everyone else in density, even if others claim to have "14nm" nodes.
Stuff like Intel unable to perform below 80 degrees on 4 cores, makes people switch with ease.