If we would represent information exchange between individuals as a directed graph, then in the times before the print press it was a relatively regular graph where each person had mostly bidirectional exchange with a few other people around them, which likely knew each other as well. Written language was able to create a few links between people from the past to many people in the future, but most information exchanges were face to face talk.
Then print press and later TV and radio were invented and more nodes arised that broadcasted information to many other nodes. The diameter was still large though, because two random poor people very far away from each other could not easily pass information to a node with has a high reach.
Then internet became popular, and suddenly the average number of in- and outgoing edges that each person has skyrocketed. Whoever has a piece of relevant information that is likely to be shared can easily reach a huge network of probably more than half of the people alive on earth today.
Now big machine learning models become popular, and by generating images with dalle I indirectly consume information of every human that generated training data.