I write columns on news related to bots, specially in the categories of Artificial Intelligence, bot startup, bot funding.I am also interested in recent developments in the fields of data science, machine learning and natural language processing ...
I write columns on news related to bots, specially in the categories of Artificial Intelligence, bot startup, bot funding.I am also interested in recent developments in the fields of data science, machine learning and natural language processing
Interviewed at the Gartner Symposium yesterday afternoon, Amazon Web Services founder and CEO Andy Jassy said that Amazon Web Services had three major areas of focus: serverless computing, machine learning, and the Internet of Things.
On serverless computing, Jassy said people will use full instances for a long time, but increasingly companies will transition to the event-driven service model pioneered by AWS's Lambda services. He said not having to deal with servers is very appealing to customers.
On machine learning, Jassy described three layers of services fundamental work done by experts on the bottom; a middle layer where traditional programmers can create their own services; and a top layer consisting of services and APIs, such as Amazon's Lex speech recognition service. He said Amazon was investing in all three layers, with the vast majority of work done by experts working on Amazon's P2 instances. But, he said, very few people have the ability to hire expert practitioners, so more has to be done to make it much easier and much more accessible for everyday developers. In this middle layer, where he said everyone has services, "no one has made it easy enough."
On the Internet of Things (IoT), Jassy said its a long journey, but that in 10 to 15 years, most of a company's on-premises footprint will be connected devices. He noted that big applications to date include devices within oil fields, cars, planes and ships; and said most big IoT applications use cloud services, and most are pretty significant users of AWS, including Illumina genetic sequencing machines, John Deere's tractors with analytics, and Major League Baseball's Statcast.
But, he said, many devices don't have the time or connectivity to get back to the cloud, so its important to have the same programming model used in AWS on devices. To that end, he pushed the company's Greengrass services, which has Lambda built in. This is the area that is delivering the fastest results, he said.
Interviewed by Gartner Fellow Daryl Plummer, who pointed out that AWS is the most profitable part of Amazon and now has a run rate of $15 billion in annual revenue, Jassy spent a lot of time recapping the history of the firm and its growth.
He said that one of the major reasons the company started AWS was to provide Amazon's consumer business with a way to build applications more cost effectively and to allow it to innovate more quickly. Jassy noted that Amazon gave AWS its brand and credibility. AWS treats Amazon as "one important external customer," because its a very demanding, big power user with high standards, and he said AWS gets a lot of feedback from Amazon.
No one inside the company would have had the audacity to think AWS would have grown so quickly, Jassy said, and he would not have guessed it would have been a $15 billion run rate business, growing at 40 percent, at this point. When he presented the proposal to the board he had no profit and loss statement, saying we had "no idea how fast it was going to grow."
Jassy spent a lot of time talking about the company's customer focus and the firm's "working backwards process." He explained that every time they produce a major new feature, they begin by creating a press release, because that makes them think from a customer's perspective about what would be interesting to them. They then create a FAQ (frequently asked questions) document, because that contains all the how's about using the product, pricing, etc. He said its important for the company to be long-term oriented, because they don't always get it right on day one, so it's necessary to get a product out there, listen to customers, and iterate quickly.
Asked about notices the company has sent to customers suggesting ways of saving money on the service, Jassy said the company sends millions of such notices, which has resulted in $500 million in savings for customers and thus less revenue for AWS. But, he said, if you understand customer focus, you'll understand that AWS doesn't want to make money unless they are delivering value. "If you do right for customers, it will pay off over the long haul," he said.
Plummer noted that in an informal survey he had done, customers said the hardest thing to understand about AWS is the bill (which was mentioned by 95 percent of customers), followed by the number of new features. Jassy said that last year, AWS introduced over 1,000 services and features, and was likely to introduce about 1,250 this year, so they have plenty of things to keep working on. He said the company has received feedback that its confusing to be billed for things in so many dimensions, and said "we think we've made progress, but we still have work to do."
Jassy said he believed the transition to the cloud will take 10 to 15 more years, and that because most enterprises are starting from on-premises data centers, most will be in hybrid mode for a long time. He said he expected there would be "a handful of very successful players" in the market, and said that while most customers think they want to split their workloads relatively evenly among multiple players, most end up picking a single provider, because it's much easier for development teams.
Asked about the biggest threat to AWS, Jassy said he is most worried about being nimble and keeping up the pace of innovation, as well as hiring the right people as the company continues to grow.