Right, but it is not true that Intel is having absolute lower temps per core or working on the same task. Per CPU perhaps, but not in any normal, real life workload that loads the cores with variable loads.
What that linked temp from TPU tells you is what the CPUs will run at on full continous loads. Let's not speak of the resulting performance
While I agree with the premise that Intel's 14nm is just a very strong node that manages to last, its also way beyond any semblance of normality in terms of behaviour. What you're getting now is a peak burst followed by a drop to abysmal clocks, IF you like to do full continuous loads. Therefore it loses everything against a slightly warmer productivity scenario on Ryzen. At the same time, Ryzen seems to have found a sweet spot between temps/load and clock behaviour where it doesn't go into extremes but still performs admirably under all circumstances.
You, yourself said these variables are all linked and bring great arguments for it, but they certainly don't apply to Intel's current offering. Intel's spec is about playing benchmarks, not real scenarios, for their performance parts, and the customer suffers by getting something that always performs way below expectations unless you slap a massive heatsink on it and take off all safety measures (long term power limit and max current draw etc.).
The fact that Intel is even today marketing K-CPUs with massively extended TDP budgets (on the spec list, nvm they're a straight up lie, but ok) against almost similarly clocked non-K versions is the writing on the wall. They still sell the 'OC' moniker for things that simply won't OC because they're already pushed over the edge.