This actually sounds like it could make for a great system, in the sense that you still maintain the original quality and quantity of output, but you still get to avoid the burnout of that 20% completely, because after some time, it will be somebody else's "turn" to do more heavy-lifting. Obvious problems here are that 1) company/team structures in industry don't seem conducive to this, 2) skill levels are unequal, so to maintain the quality of the 20%, one would need to plan carefully to account for this -- but, it also seems like a good way to help lower skilled workers rise to meet higher standards -- and 3) hinted at before, this needs to be institutionalized in a tangible way, and that seems to require a difficult form of team member assessment/introspection that managers may not be capable of facilitating.
Wrt point 2 above, I've heard from sources I trust (but would have to bug them for specific sources on this) the following phenomenon has been observed in education: when you remove top performing students from sections of a course -- say, to join an honors section of that course -- the non-honors version of that course has a void for high-performance students. Interestingly, this void is often filled by the mid-tier students who would have typically performed at B/C level, and now some of them will rise to perform at an A level. It would be fascinating to find observations of this in the industrial setting as well. If I find a source for this, I'll update my comment.
This is definitely a phenomenon that can happen, but it has a dark side that I feel is rarely accounted for. While you want to have senior talent and empower your engineers to grow, you also need to have junior/mid level engineers to support them. All to often the reason you have 20% of the employees making 80% of the effort is because they have no supporting infrastructure to help them with their jobs, or they are powerless to leverage the other employees for help without interacting with, sometimes multiple, management levels.
When we started to expand the engineering org it became readily apparent what we needed wasn't more senior talent, but more, for lack of a better word, grunts to help with the mundane tasks that the senior engineers shouldn't be working on. This same thing happens in most other organizations outside of software development as well. Essentially one way to ensure that you don't have your engineers burn out, is to ensure they have a team they can fall back on completely outside of management. I personally appreciate getting recognition from my manager and the higher ups at the company I work at. However I appreciate the team I can rely on far more. I think generally HR/Managers should get out of the way more often, and instead of directly acting make it easier for teams to help/support each other.
Now I myself have long list of complaints about Spring/Microservice crap and its tendency to bury simple errors in 10 thick layers of abstraction. But that will be more relevant to IT related threads. It is because many other developers are dealing with it and still somehow make it work without spending days for an hour job.
Just yesterday at my workplace one dev spent whole day to set up Eclipse workspace for a basic Microservice project. It was about 10 min job. Now that person claims to be junior developer not a total novice and I see no reason to doubt that. But setting up similar projects has been explained and demoed in great detail at least couple times in last few weeks. Still anytime something changes he can spend days to do most trivial things. I would think this person would simply burnout if task is given with a end date. It doesn't matter how generous is the deadline.