It's an interesting departure from my 1970s highschool days... when we were all assured that the path to strength and endurance was through breaking ourselves down to rubble first.

Okay... so going forward I will keep in mind that the primary goal of efficient training shall be to provide the necessary stimulus to trigger the desired response... at the minimum possible cost in fatigue. As a process engineer, this is an appealing intellectual model.

Unanswered question: do the long Z2 (Friel definition) efforts have to be substantially all Z2?

There is nothing on earth so boring as 90 minutes of Z2 on a trainer. I've been doing a fair number of Z2 trainer rides lately, and one way I deal with the boredom is inserting a few 10 second all out sprint efforts (say every 5 - 10 minutes). This has been good for my sprint power (I need that!) & doesn't seem to come with a big cost in fatigue.

It's interesting that the 8 minute intervals were found to be so effective. Actually, it's confusing as hell. Is the training response of 32 minutes (4 x 8) @ threshold plus a 5% really that much superior to a 32 minute session at threshold plus 2%? Perhaps it's all about volume: the four minute efforts don't allow enough time in lactate accumulation mode to get the desired stimulus/response, 30 minute & longer threshold intervals allow too much and invite over training.

Can't argue with a 70 ml/kg/min VO2.