I see where you’re coming from, and I agree democracy doesn’t necessarily guarantee truth or even the best outcome for everyone. Often the majority can vote against their own best interests without even realizing it. That’s where I see machine learning coming into play. By being able to identify when democracy fails and verify information against various sources the ML can tag it and either alert a moderator or even show the article as flagged.
For example, let’s say something gets upvoted by millions of people rapidly. Well, that’s a sign it’s being brigaded or manipulated. The ML can then learn from instances like this so that it can tag said content in the future.
Not to mention, much like Reddit you’ll see content based on communities you’re interested in as well. I don’t want to speak for Travis, but I don’t think the intention for DeepSee is to be the front page for the Internet like Reddit, but rather a portal for seeking the truth, and community.
Governance and human involvement are still going to be major factors in any kind of aggregation system like this, the ML is merely there as a tool to help those processes. So democracy will factor in as a form of governance but not be the only system in place for content and community management.