Much of this knowledge is highly specific, and isn't captured in the literature. We can call this 'tacit' knowledge. To this end, science isn't really the literature - the literature is a partial representation of what we know. The full range of what we know lives inside the minds of researchers.
Still, as Baszucki made clear, the goal is still actual social networking: surely that will always be better than interacting with an AI! Or will it? It seems to me that perhaps the most important constraint on the web — to actually interact with people as if they are, well, people — disappeared a long time ago.
On the internet, websites are like buildings, hyperlinks are like roads, traffic comes in the form of people and information. It is a complex system with many interacting parts.
the problem is with the interface of scientific literature. We cannot expect the wider population to be scientifically literate if they are not given the capability of understanding the artifacts we use to communicate about new scientific developments.
DAOs are novelty search engines which can more efficiently explore a search space by enabling many cooperating teams to collect and integrate stepping stones.