John Brunner and Alvin Toffler both made stark warnings wrapped in futuristic giddiness about things like it. People like Fuller no doubt thought polling on a large scale was terrific. There were old usenet groups and BBS subs (minus the money aspect) experimenting with the model. I do not believe they ever are or were good in a largescale model (money or not).

I mean wisdom of crowds, super-forecasters, calibration and pre-registration are useful tools that can result in better predictions, turning it into online gambling is where it went sideways.