Amazon sent shockwaves through the internet when it announced that its highly anticipated drone delivery service would be launching “within months.” This program, tentatively named “Amazon Prime Air,” promises speedy delivery times of 30 minutes or less — with packages carried by aerial drones that take off, fly, and land without human controllers. When it launches, Prime Air would be one of the biggest efforts in history to bring robots into the mainstream and a monumental achievement for retail and logistics.
That wasn’t all that Amazon announced at its landmark re:MARS conference, however. The future-focused event brought a number of speakers to the stage to discuss developments in space travel, robotics, and environmental sciences. Between these keynotes, the company unveiled a number of new projects that bridge the gap between artificial intelligence (AI) and our daily lifestyles.
The two biggest announcements both focused on Amazon’s latest advances in AI. Alexa, whose natural speech and comprehension is already turning heads, is set to see even more improvements in natural speech. And StyleSnap, a newcomer to the Amazon family, plans to harness AI to help people discover new fashion items and clothing. Both of these projects bring ambitious amounts of research and development, but will they be able to win over fans and customers in the coming years? The future at Amazon is about to get wild.
Amazon announces Alexa AI advancements
Since debuting the Echo system, Alexa has been a cornerstone to Amazon’s strategy for marketplace domination. The popular virtual assistant not only lets users order items with nothing more than their voice, but also allows them to control their homes, get information, and carry conversations in real-time.
But conversations, traditionally, have been a weak point for Alexa (and nearly every other virtual assistant up to this point). Current AI technology works primarily in “workflows,” or processes that the computer understands one step at a time. This makes it difficult to naturally change topics without telling your virtual assistant — breaking the illusion that you’re actually “talking” instead of “commanding.”
Amazon’s new Alexa project aims to tackle this problem for good by changing the way the virtual assistant handles conversational workflows.
During the Alexa keynote at re:MARS, developers showed off the new method the system is using to create conversations that sound and feel natural for users. The demo showed a user ask Alexa which movies were showing, select a movie, and order a ticket without needing to switch between Alexa skills or break the natural flow of discussion.
If you’re not a coding expert, you should know this sequence is extremely complicated for a computer program to perform. Fortunately, Alexa has the help of a neural network that predicts likely responses the user will say. These responses are then queued up with relevant skills to allow Alexa to switch seamlessly between functions behind the scenes.
According to Amazon, this advancement will increase the accuracy of Alexa’s responses significantly, and thanks to machine learning technology, it will only improve as more users continue to talk with the system.
According to Amazon, these Alexa improvements are scheduled to roll out in “the coming months,” so you won’t have to wait too long to have full conversations with your favorite artificial intelligence service.
Amazon reveals “StyleSnap” — a “Shazam” for fashion
If you’ve ever been browsing on Instagram or Facebook, you’ve probably seen friends, connections, and influencers dressed in outfits you’d love to get your hands on. The only issue is figuring out where they got what they’re wearing.
Outside of awkwardly asking these people where they found their top, your best bet is to eyeball the outfit and go off your best guess — hardly a way to acquire clothing.
Surprisingly, Amazon has paid attention to this common scenario. It’s announced a new service called “StyleSnap,” which functions much like Shazam does for music.
All you need to do is upload a picture of an outfit you like, and Amazon’s artificial intelligence machine learning algorithms do the rest. Using data from Amazon Web Services and the Amazon Marketplace, StyleSnap will serve you a list of items to purchase. If it can identify the brand or designer, it’ll present that item for you to purchase. If it can’t, it’ll recommend something similar looking.
If StyleSnap is able to accurately recognize designer pieces, it could revolutionize the way people share their outfits — and potentially give Amazon’s already profitable marketplace an even bigger boost in sales.
As for now, there’s no official launch date for StyleSnap. Hopefully, we’ll see it launch in time for the holidays — although I doubt it’ll be able to register all the ugly sweaters floating around.
Amazon has big ambitions, and with artificial intelligence behind it, the sky is the limit (literally, in the case of Prime Air). All we can do is see where the future takes us, but from the looks of it, it seems like our leisure time will only get more convenient with every update Amazon delivers.
Whether it’s the first group to crack the AI code is another story, but I wouldn’t put it past Alexa to become the next Hal 9000.