PLUS: The latest AI and tech news.
![](https://link.wired.com/img/5cec23a9fc942d3adafdb950ee2if.17pz/4efd6caa.gif)
By Jennifer Conrad | 06.17.21 In February, Saturday Night Live pinpointed a new form of guilty pleasure for people in their late thirties: scrolling through real estate listings on Zillow. Who hasn't lost a few hours clicking through the options: How much more for a pool? Is that a fire pit in the back?! Part of the magic comes from Zillow's Zestimates, the company's proprietary measure of the value of a house. As Khari Johnson reports, earlier this year, as home prices soared, Zillow began making cash offers to buy homes based on the company's price estimate. Zillow says a new algorithm will produce more accurate estimates and allow the company to offer to buy more homes—in sales that can close in as little as a week. Zillow has been tinkering with its price-estimate formula for more than a decade. (WIRED reported on the company's early forays into AI back in 2007.) Since 2016, the company has used computer vision to evaluate photos of homes. The new system utilizes a single neural network, a form of deep learning that makes it easier to understand how certain factors, such as proximity to a waterfront, affect a home's value. Zillow hopes to bring the one-click sensibility of ecommerce to buying and selling homes, CEO Rich Barton told Johnson. "Eventually we want to have most houses in the country inside of that live-offer buy box," Barton said. Read more here. | Zillow's price estimator isn't the only innovation shaking up where we live and how we buy homes. Over the years, WIRED has chronicled how technology is impacting the real estate market. | |
Post a Comment