On Monday, residents of Seattle will have the chance to shop at Amazon Go, the online retail giant's brick-and-mortar grocery store, becoming the first people outside Amazon to try out the cashier-free shopping. Amazon employees started using the convenience store in December 2016, and mastering the technology of using cameras and sensors to charge people the correct amount for their purchase proved harder than expected. Issues included differentiating shoppers with similar body types and dealing with children eating items in-store or rearranging them on shelves, Reuters reports.
Shoppers pass through a turnstile to get into the store, scanning a smartphone app that links them to a credit card on file. Cameras and weight sensors on shelves determine what customers buy, and they are charged for whatever they still have with them when they walk out through the turnstiles again. Reuters correspondent Jeffrey Dastin tried out the store, and he got in an out with a bottle of water in under 30 seconds.
Since customers like speed, Amazon's checkout-free technology could upend retail stores more than its online store already has. But the company says it has no plans to introduce this technology to Whole Foods Market stores, which are bigger and more complicated than Amazon Go shops; Amazon purchased Whole Foods last year for $13.7 billion. Peter Weber
On Wednesday, Google DeepMind's AlphaGo notched a new milestone in artificial intelligence, beating a champion player in one of the world's oldest and probably most complex board game, Go. The Go master, Lee Se-dol of South Korea, had initially predicted he would beat AlphaGo in at least four of the five scheduled games, arguing that the 3,000-year-old Chinese game requires "human intuition." Some AI experts had agreed.
"I am very surprised because I have never thought I would lose," Lee said after falling to the AI software in Game 1. "I didn't know that AlphaGo would play such a perfect Go." He still put his odds of prevailing in the best-of-five series at 50-50. He'll have his next shot on Tuesday, for Game 2. The victory of Google's DeepMind over the human mind is being compared to IBM Big Blue's defeat of chess grand master Gary Kasparov in 1997. Demis Hassabis, the founder and head of DeepMind, called the win a "historic moment," adding, "Really, the only game left after chess is Go."
Go is a two-player strategy game where you use black or white stones to try and win territory on a board. Instead of considering the seemingly infinite options, AlphaGo had already taught itself to improve and pick the best moves by playing millions of games against itself and the numerous Go games you can play online.
Matches between man and machine are almost always lopsided in one direction. "But every once in awhile, a technological moment comes when the man-machine match-up gives us a fight worth watching," says Erik German at Quartz. "(Just ask John Henry.) DeepMind vs. Lee Se-dol is one of those moments." Game 1 took three and a half hours to play, but BBC News sped that up to a brisk 25 seconds. You can watch below. Peter Weber