Ok, so twitter recommendation code is out. But it doesn’t have the weights of th

Ok, so twitter recommendation code is out. But it doesn’t have the weights of the NN, and there aren’t any graphs that display results.
It’s what I expected but more complicated than I expected. And requires more processing than I expected.
What I can see off the bat, is that it’s going to gravitate to a sort of echo chamber centered around content producers. And there is no hierarchical separation, so that’s why it produces a race to the bottom at times.
Fascinating.
Without some metrics on the result I don’t know how to evaluate what we’re seeing.
What I”ll gamble on, and it’s the reason for ‘the problem’ is that it’s far more complicated than it needs to be, because in general, these algorithms drive toward deterministic results with far fewer criteria.
Can’t wait until we see more.