In nature, you don't necessarily have the needed genetic diversity or controlled conditions, in the right combinations, the way you can obtain them in the lab. And "several decades" is probably not enough time in evolutionary terms to adjust, if I had to guess.
Which holds that rather than a slow gradual change, species under go a rapid change and then are stable after that change.
If that hold true, then the driver would be that you have low impact changes going on all the time that do not take hold but once a beneficial change is introduced it spreads rapidly.
I think the notion of its status as a "prevailing theory" is overstated, and is an idea that exists more in the minds of lay audiences than in the profession (see criticism section).
I agree with the criticisms that Gould, unfortunately, spent much of his career attempting to position himself as an intellectual revolutionary, overstating the primacy of PE and was in many important respects an intellectual fraud. Gifted as he was as a critic of intelligent design, I think he set back our understanding of evolution almost as much as he helped.
If not nature.... where do you get diversity from?
It seems such hybris to believe that mere humans have a better handle on what 'nature' needs, than, er, nature.
The embedded idea is both guilt-ridden and egotistic. Amazing destructive humans are harming 'nature' (antropomorphised) but other, good humans are here to redress the balance by doing this or that. At which point, whatever is there is no longer 'nature' but some sort of themed garden.
That depends on the scale of the changes. Life can make a planet habitable, but our sun will eventually die, and the systems that ecosystems raised to control the changes in its favor will be unable to fix that. Only we could do something about it.
