I agree with your analysis of management and the social media. However, I have a problem with 1 account = 1 vote: there are hundreds of thousands of accounts that are bots. So, how to differentiate a person from a bot? Should we apply the KYC rule? And how could Steemit comply with the GDPR law? I can not see a solution with Steemit's governance system. The code is the law but the code is made by humans where the consensus is difficult!
The proposal for authentication is stacked. But, in the case of the proposed SMT, the one on the video, there would be to sets of oracles filtering the users.
One set would be implementing 1 account 1 vote - this could be approached in a similar fashion that people to introduceyourself posts at the moment. However, who is to say that the account faucet could not also have IP restrictions and what not. Granted this part could be tricked, an Oracle could miss that one account created three more. But this is when the second account kicks in.
The second set of oracles determine if the user is acting ethically. If the user is spam farming, triggering upvotes thru automated curation trails, etc. The oracles could shutdown all those account's earnings.
So, to be clear. it's not 100% bulletproof but... The amount of work abusers would have to do, might be enough to curve it significantly.
The moment they get discovered, they lose all possibilities of making any income. Unlike today, were if they have stake, it doesn't matter if they abuse, because they got enough stake to live outside the ethical expectations we may have.
An abuser will use a VPN to counter an IP. Is it possible to use the IMEI? A person can have multiple devices but usually a person is single to use these devices. So IP + IMEI could be an account.
However, there is still the GDPR. Several law firms are studying the possibility of litigating large corporations and blockchains. In the case of blockchains, it will be miners and witnesses who will be prosecuted.
From my understanding an SMT would have free range to work this out in the way they see best suited for them.
There is some socio-economic factors to the game as well. Believe it or not, in Steem's white paper it predicts spam and abuse and it says that some of it is necessary to create the network effect.
Now that sounds counter intuitive, at least on the surface. But, when I think about the fact that no society has gotten ridden of negative behaviors, at least not entirely.
I will admit, I'm not aware of the sweet spot, but I'm convinced it does exist. Maybe if we cut down with spam/abuse to a third of what we have today. Maybe, that's optimal...