Companies trying to break new ground in real estate search by moving beyond the old parameters (location, price, home features) and into territory traditionally reserved by agents — figuring out what homes are a good match with a prospect’s taste — take note: the “chief scientist” at music matchmaking site Pandora admits the company can’t always predict user’s tastes.
“I do not believe this is a solvable problem,” Pandora’s Eric Bieschke tells PandoDaily’s Erin Griffith. “But that’s what makes it so interesting to me.”
Pandora still believes that “only an algorithm can create the best, most personalized listening experience for its users,” Griffith writes, and has largely shunned the move toward “curation.”
A self-described “wanna-be music snob,” Griffith says the problem with algorithms is that she detests some songs she should like, and loves others in a way that “confounds logic and rationality.”
The prospects for divining the tastes of homebuyers may be brighter. In most categories, Griffith thinks, “algorithms can get pretty damn close based on our past preferences.” Music is a “whole different beast” because we “place a huge emotional significance on songs we like and hate.” Source: pandodaily.com.