Simulated opinion polls to drive public policy — what could possibly go wrong!
In the name of saving costs and producing more polls (i.e. the double benefit of increased efficiency AND productivity), statistical truth (traditional polls) is being replaced with its derivative, producing statistical insights via a second layer of statistics and agents simulating known, pretrained demographic behaviors & preferences, instead of interviewing actual people. Yay!
It seems they're blindly following Wheeler’s Law, i.e. "All problems in computer science can be solved by another level of indirection" and ignoring the corollary: "...except for the problem of too many layers of indirection."
Queue future presidential approval ratings based on a simulated population of agents... Too dystopian?!
https://ngmx.com/@pathunstrom/116565726550598652
#LLM #AI #Gallup #Statistics #Poll
In the name of saving costs and producing more polls (i.e. the double benefit of increased efficiency AND productivity), statistical truth (traditional polls) is being replaced with its derivative, producing statistical insights via a second layer of statistics and agents simulating known, pretrained demographic behaviors & preferences, instead of interviewing actual people. Yay!
It seems they're blindly following Wheeler’s Law, i.e. "All problems in computer science can be solved by another level of indirection" and ignoring the corollary: "...except for the problem of too many layers of indirection."
Queue future presidential approval ratings based on a simulated population of agents... Too dystopian?!
https://ngmx.com/@pathunstrom/116565726550598652
#LLM #AI #Gallup #Statistics #Poll
1
