While on one hand, HF19 was a necessary change, it made the bad incentives 1000x more problematic. They would be wise to make changes I will be making, that will mitigate this. Specifically, using reputation as a vote power coefficient. I am uncertain as yet whether the normalised form of reputation, or its native logarithmic pattern should be used, I'm inclined to think the native value would be better. It would effectively give a limitation effect like the old, geometric curve used in vote weighting before, except using a social regulation scheme. If zero reputation also caused accounts to have severely limited bandwidth, it would allow humans to flag bots down to zero, and basically solve the bot problem.
With the explosion of blockchain size in the last 6 weeks, this change is now past being necessary, and going towards 'too late'.
Steem definitely needs course corrections more than new users at this point.
I think they are waiting for somebody to build the new,...
Let's hope that is not the sum of the indicators.
This sounds like an amazing idea to not only fix the bot problem, but also to make reputation much more meaningful while limiting bandwidth for those that don't deserve it.
On the other hand, it seems that you're saying those with high reps and high stakes will essentially be gods on the platform, with even more say with what goes on than on here. There must be enough at the top to keep others in check, and I haven't the slightest clue what number/percentage is ideal, but if the top weight/rep is concentrated in few hands, the probabilities of an oligarchy forming are quite high IMO.