The emphasis here being on the word "should". I wouldn't buy just yet until some third party has published a sustained load/heat test. Apple have sucked at this for quite some years now, it's folly to touch a MacBook for anything remotely compute-intensive

Only when you compare apples to oranges, but this is an apples to apples comparison. Same underlying silicon, almost identical configuration, different chassis. It's impossible for the Pro not to outperform the Air when the only substantial difference for performance is TDP and we have zero reason to believe that heat-pipe + fan would be outperform by passive cooling of all things.

Agreed, the Pro's cooling should be better. However, I think it's worth waiting and seeing. There might not be that big of a difference in the end (e.g. both are uncharacteristically well designed and run fine; or conversely both are so badly thermally throttled that no serious intensive tasks can be performed) in which case other factors could end up being more significant.

Apple has always sucked when it came to cooling. The Apple /// and Lisa were plagued by “IC creep” where, due to heat expansion from inadequate cooling, the ICs would wiggle out of their sockets ever so slowly.

Yes, that is almost certainly the case. They didn’t mention clock speed at all though so we’ll have to wait for real-world tests.

They have said exactly that multiple times in the presentation. Mini & MBP will maintain higher clock speeds indefinitely.