Tweet · 15 Aug 2018, 17:30 UTC

Replying to @jw_lockhart and @msalganik

I wonder to what degree one could require publication of metrics that would be meaningful across many recommendation systems. Eg. tendency to break down shared reality.