Bytes: What’s new in tech
Beijing’s new censorship tools
“China’s already formidable internet censors have demonstrated a new strength,” said Eva Dou in The Wall Street Journal. Authorities can now delete images sent in one-on-one chats as they’re being sent, so that the intended recipient never sees them. Activists say the new capability saw its first widespread use following the recent death of Chinese dissident Liu Xiaobo from liver cancer, with Beijing moving aggressively to block supporters from sharing tributes online. In recent years, Chinese internet users have turned to sharing photos to avoid cyberpolice who use wordscreening software to filter politically charged messages. Censors initially responded by purging images from public forums and group chats, but one-on-one images are now being blocked midtransmission. “The speed is too fast for human intervention,” suggesting that censors are using an algorithm.
Facebook to test subscriptions
You might soon have to pay to get news on Facebook, said Gerry Smith in Bloomberg.com. The social network is planning to test a subscription service that will let media companies charge readers for access to their stories, with a paywall kicking in after a certain number of free articles. The new feature will be part of Facebook’s Instant Articles program, which hosts complete stories within Facebook instead of links sending users to a separate website. Publishers have been able to make money from ads sold through Instant Articles, but some have cut back on the feature after struggling to profit from it. The move could win over publishers, many of which have increased their focus on selling subscriptions to make up for shrinking ad sales.
Robo fashion designers
Clothing designers are getting a hand from Big Data, said Dave Gershgorn in Qz.com. Subscription fashion startup Stitch Fix, which already uses algorithms to pick out clothes its customers will want to buy, is now using artificial intelligence to create must-have outfits “a designer hasn’t made yet.” Stitch Fix uses what it knows about its customers’ preferences to predict and design blouses and dresses that aren’t hanging in their closets. Algorithms generate suggestions like “a different neckline or sleeves,” and add in “a little bit of randomness” with atypical patterns and other details. It’s still up to human designers to make a final product that fits “into this season’s fashion zeitgeist.” ■