Good Judgment Project apparently started in 2011 as part of an IARPA tournament. I first learned about through The Economist, which frequently mentions the crowd consensus on forecasts related to Economist articles.