Adjusting rewards down on overrewarded posts are good downvotes however, I'm sure you're aware of many "blind stakeholders" who vote things up way too hard. For instance in ocd we have rules in many initiatives where we don't vote posts unless they're at least 12 hours long to prevent these "blind" upvotes to happen to see them in the hot/trending list and overreward them - but others don't place these rules because they wanna be overrewarded - while to some it happens so seldomly that we let it be (it's fine now and then). Good downvotes can also come from certain authors being overrewarded time after time without giving many a reason to justify the rewards, for instance someone who is only a content creator and barely plays the social game, we may not know why they're getting so much rewards, why they're getting upvoted late in some cases even at the cost of the curator, etc, but downvotes may help send some of those rewards back to everyone else so it is distributed more fairly.
In terms of burn comments or hbd.funder ones, one could argue the reward pool is set to always give out a certain amount of stake but since we're not really seeing a wave of new users coming in and in general posts are deemed "overrewarded" by many other stakeholders, it is acceptable to not distribute those author rewards at all and use them for something else that benefit all stakeholders.
For instance I'm spending most of my daily voting power rewarding reddit shares lately as I think bringing traffic to our front-ends is more valuable than just creating content that very few consume.
If/when the time comes that that changes I hope people would join in on curating users again to help distribute stake, but I don't really think it centralises stake, one could argue that a lot of stake is going to authors for barely any effort lately and that this activity itself hasn't really helped bring in new users, not to mention authors who don't even care if it does.