Maybe you have seen this post shared by @exyle, including witness 'voting' on some proposed changes to the reward system. Look at the image further down the article.
Or this post, which initially stirred the new talks.
I gave some thought what I would do in case of each of those proposals coming to effect.
In essence, it would all boil down for me to limited posting in the future, focusing on curation while undelegating from curation projects to increase my available SP (which may still happen in this system) and delegating for ROI (if better than curating), especially after SMTs are out.
I don't believe the proposed changes would help the ecosystem. While I can live with some changes and adapt, others will try for a while than probably quit if it doesn't work for them. Which doesn't mean plenty haven't already did that with the current system, which is not perfect.
Some changes, like a superlinear reward curve (however light it may be) or 10% free downvotes, were largely dismissed by top 20 witnesses, so the possibility that they will be implemented in the foreseeable future is low.
But even a 50/50% reward distribution, especially combined with a reduced curation window, will make human curators focus on known authors with high payouts, because it induces a FOMO in the small curators. Time goes by and you haven't voted yet on that post what everyone will vote on? Who will have time to even see, not to mention vote on smaller authors, or undiscovered ones?
Maybe I'm wrong about this. But I believe the most to lose will have the unsuspecting minnows/red fish authors, and the most to win will have those who understand the system best and have enough SP to spread their curation and well-known authors.
As for the downvotes proposals (either of them), I believe I would act the same way as I do now. Meaning I will rarely downvote something, and I don't have a problem 'paying' with an upvote for that.
Although I understand why this change would make an impact to those using downvotes more regularly, it would probably be either ignored or confusing for the regular guys.
Some people are unhappy with the current reward system. They might have their reasons to be.
It doesn't work perfectly for me either, considering my posts generally receive mostly automated upvotes (some 'paid for' with delegation, and I might start fixing that).
But with all the vote abusing, it seems like a stable system. Do we want to make these changes when we don't even know how this ecosystem will look like a year from now and what its needs will be post-launch of various SMT projects?
If reward system proposals will be considered at any time, they should be properly tested on a close-enough to mainnet conditions testnet, with many of the DApps working there and incentivizing regular users to participate.