Thank you very much for this podcast, but I am afraid you are underestimating the devastating effect of the failure caused by the transition to HF21. It was a textbook example of how testing in production, the wrong way, looks like. Now, it seems, the most important thing is to win back the confidence of users, which was lost. It certainly does not help when you say, that the fact that the blockchain halted was "actually, a good thing". Sorry for the parallel, but it sounds like the "propaganda of success" in the former communist countries. It was not a good thing. Actually, it was a very bad thing which undermined the confidence of users, not to mention the investors, of course, and you have to face it. To say the truth I am pretty scared that the next hard fork could happen already in February next year. If I were you I would not plan it at least until 2021. Give yourself more time to test things properly.
The other thing you mentioned is content search. It is certainly not possible to solve properly now, when you got only five tags and they are mostly abused for things which have little to do with the contents. On the other hand, everyone knows that tags suck and the right way is to use controlled vocabularies. While this may be beyond capabilities of avarage user, it certainly is not for even a semi-skilled curator. If only a small fraction of posts had proper subject identifiers it would already greatly increase the number of meaningful queries.
The reason for the rush is that others are developing blockchain-based social medias and Steem is no longer the only fish on the pond. I still think they have an enormous advantage because of the amount of great developers and the endurance the core of the Steem community has.