2 Comments

Only loosely related: I wonder what the population pyramid looks like filtered for high-status people (or filtered for many other relevant dimensions like income or education). The ability to predict the the share of the population with certain attrubutes seems valuable.

Expand full comment

Are you thinking more about categorical, indexical or continous data? But anyway I agree that it would help. And with more data we could move beyond simplistics hypothetical models like Scott Alexander's Zebra stripes:https://slatestarcodex.com/...

Expand full comment