The training argument should be better specified IMHO.
What changed exactly, and what are the expected positive changes in key physiological variables?
I for one believe that training evolves and this is driven by competition, individualisation and experimentation. This will have positive performance effects. But improvements in this domain are very often overstated. Arguments are also often wildly anecdotal. Sometimes they highlight particular coaches such as Sola now and Kerrison back in the days, as though there are some secrets.
At least in running the so-called double threshold method that is supposedly behind improved performance is discussed very explicitly and openly. Yet it comes down to very simple, logical and time-tested principles: maximisise load in a manner that is sustainable and specific enough to competition needs. Innovators started to tinker around this model in the 90s already, mind.
So quite often a sort of simple past fallacy is deployed. Athletes were not clueless yesteryear.
Re: Pog there has been a claim that he previously did not do "long intervals" and, exaggerating only slightly, including them has been the key. Assuming this was the case, I would find this a positive change, particularly in relation to GT performance. But I'm not buying that his performance jump could possibly have hinged on this. Threshold can be built with Z2 volume too. Allegedly that was his thing previously.
More generally, what I am skeptical of is that improved training could explain how roughly 6w/kg became roughly/almost 7w/kg in GT climbs. How seasoned top professionals made this jump after COVID-19 and 2023-24 in particular remains the puzzle at hand.