It might be useful to keep percolation theory in mind if you’re trying to model the spread of the meme that LW exists and is worth checking out in a small population buffered by a larger population that is unlikely to find the message true enough to pass along.
Think of the social graph as a crystal lattice. Now imagine that some percentage, p, of the nodes in the lattice are likely to pass the message along because they find it useful. The message is like a fluid trying to spread through the medium via the nodes that will pass it along.
See here for an image of a percolation theoretic display of the square lattice in two dimensions with percolation probability p=0.51.
If p is high enough and/or the lattice has a lot of connections, then you get a sea of transmissibility with islands of isolation (because they are surrounded by buffer nodes) that need an unusual effort to reach. If p is too low and/or the network is too sparse then you have islands of opportunity and each separate island needs to be contacted by extraordinary means in order of the message to reach it.
Given the way you’ve specified the target demographic, I’d guess that we’re dealing with the island scenario. LW is probably an island. It may not even be the largest island there is. My naive guess would be that outreach to other islands would be tricky… like… how does one entire community Aumann update with another entire community?
This is likely to matter to our outreach strategy.
It might be useful to keep percolation theory in mind if you’re trying to model the spread of the meme that LW exists and is worth checking out in a small population buffered by a larger population that is unlikely to find the message true enough to pass along.
Think of the social graph as a crystal lattice. Now imagine that some percentage, p, of the nodes in the lattice are likely to pass the message along because they find it useful. The message is like a fluid trying to spread through the medium via the nodes that will pass it along.
See here for an image of a percolation theoretic display of the square lattice in two dimensions with percolation probability p=0.51.
If p is high enough and/or the lattice has a lot of connections, then you get a sea of transmissibility with islands of isolation (because they are surrounded by buffer nodes) that need an unusual effort to reach. If p is too low and/or the network is too sparse then you have islands of opportunity and each separate island needs to be contacted by extraordinary means in order of the message to reach it.
Given the way you’ve specified the target demographic, I’d guess that we’re dealing with the island scenario. LW is probably an island. It may not even be the largest island there is. My naive guess would be that outreach to other islands would be tricky… like… how does one entire community Aumann update with another entire community?
This is likely to matter to our outreach strategy.