Revisiting this page now in 2019, I’d take more exception to this. For entirely unrelated reasons, I ran my own banner ad A/B test, and the results were far from harmless: https://www.gwern.net/Ads And this turns out to parallel experiments by both Pandora and Mozilla. Scuttlebutt has it there are more suppressed experiments also demonstrating long-term harm. (I’m running a followup experiment which I hope will show smaller effects but I don’t know what it is finding yet.)
Extrapolate the various estimates out to Craigslist and that’s a lot of potential global deadweight loss from sales/deals/rentals not happening.
Yeah, we successfully asked Jim if he would do it if enough users asked for it and he straightforwardly refused. He is very consistent in his desire to keep ads off of Craigslist.
Revisiting this page now in 2019, I’d take more exception to this. For entirely unrelated reasons, I ran my own banner ad A/B test, and the results were far from harmless: https://www.gwern.net/Ads And this turns out to parallel experiments by both Pandora and Mozilla. Scuttlebutt has it there are more suppressed experiments also demonstrating long-term harm. (I’m running a followup experiment which I hope will show smaller effects but I don’t know what it is finding yet.)
Extrapolate the various estimates out to Craigslist and that’s a lot of potential global deadweight loss from sales/deals/rentals not happening.
Yeah, we successfully asked Jim if he would do it if enough users asked for it and he straightforwardly refused. He is very consistent in his desire to keep ads off of Craigslist.