I think the topic of the next HF and core protocol changes should be focused on SMTs.

I agree that we should consider SMTs and other methods for supporting tokens in Hive in the next HF. But even more important, I think we need to focus on how we should facilitate creation of useful Hive apps (token creation is only one aspect of this).But all of that is a big task. I got very tired of how Steemit always tended to push off discussions of doing anything else, no matter how easy it was to do, while working on SMTs.

The promise of EIP was authors would have more votes with 50% rewards, compared to 75%.

That's kind of true, but not an exact statement. To be exact, the promise was the quality authors would get more rewards (because of better curation). I think that has worked out to be true, other things being even. Of course, the dominant factor for author rewards has always been and will always likely be the price of the Hive token itself, since all inflation is a percentage of the current market cap. So the best way to increase author rewards is to figure out ways to make the coin more valuable.

For example hive.auto made some changes discourage non-managed and forgotten auto-voting. We can closely work with auto-vote services to make improvements there.

This is definitely helpful.

I like your Solution 2: make it easier to find good posts.

I definitely plan to create proposals for doing this in the future, if no one else takes it on. One problem for the past months is that most all of the front end developers at BlockTrades have been tied up on projects that were begun long before Hive launched.