The people voting for funds to the DHF would benefit the same for voting for anything else. The curation rewards are a non-issue.
They are a non-issue. There was an accusation that I was somehow gaming curation rewards by voting for the hdbstabilizer comments. But I could pick a random curation trail and I would likely get more curation rewards (because transisto downvotes these comments, so I lose some curation rewards).

Anyways, this seems to be a moot point now: initially there was FUD on the curation reward issue, but I think that claim has been dropped, and the outstanding claim by Kenny is that funding the DHF from post rewards is fundamentally against decentralization. Personally, I believe it will ultimately lead to more decentralization, because I think more functionality will attract more users. So instead of the current case where increasing post rewards mostly just get distributed to the same existing social media posters, we'll end up with post rewards going to more new users.

As to whether Hive stake is becoming more decentralized over time or not, I saw a recent post (based off data from arcange's statistics posts) that strongly argues that it is:
@taskmaster4450/hive-the-growth-and-change-in-2021

Since we are almost into our 3rd year and the steem witness votes for witnesses are still in use, I take this as you are reluctant to give up the power you've enjoyed.

We've already coded in the change, and it takes effect some time this year (I don't recall when exactly).

But as you've rightly pointed out, it is unlikely to impact my top 20 rewards. And other than the rewards, there's not much in the way of special power for ranking in the top 20 (there is some minor extra ability of the top witness, but I've never used it and don't even remember what it is) even if it I did lose the first spot.

To be honest, I doubt it will have much impact on the relative positions of a lot of the witnesses, since my guess is that most stake is active. But we'll find out.

Or at least, those placed in positions to help YOU achieve your goals with Hive.
I don't honestly get this part: do you somehow think specific witnesses are helping us to develop the code and that if they lost top 20 positions they would stop doing so? Are do you just mean that we would gain new top 20 witnesses that would be opposed to hardfork changes that my team makes? If so, it is worth pointing out that for the most part, I'm working towards a platform where most changes can be made at the 2nd layer, without requiring hardforks.

I myself want to hear how we'll further decentralize the platform, but it seems few of those capable of doing so, yourself included aren't having that conversation.

I'm currently focused on building a 2nd layer contract system, which will in turn enable easy token creation among other things, and which I think will resolve some of the current disputes around rewards. I think this will also allow for more experimentation in economics without requiring hardforks.

But I don't think simply splintering off into separate token groups is the absolute best answer in the long term for organizing cooperation of human kind. It is more of a stopgap measure, IMO.

I do plan to spend some of my time starting to address that more challenging problem, after the smart contract platform is done. But while that work will utilize Hive capabilities such as cryptographic signing and maintenance of immutable data, it will exist for the most part outside the economics of Hive.

Which brings me to the fact that the top 20 consensus is the biggest obstacle to ever reaching decentralization.

I think the real answer is to move more of the decisions to the second layer, which reduces the impact of decisions by the top 20 witnesses.