When Adam Selipsky quit as CEO of Amazon’s AWS cloud computing unit, it came as a big surprise. It may have been just as surprising that Matt Garman took over after him. Garman started working for Amazon as an intern in 2005. He started working full-time there in 2006 and worked on the first AWS products. The last job Garman had before becoming CEO was senior vice president for AWS sales, marketing, and global services. He knows the business better than almost anyone else.
In a chat I had last week, Garman said that he hasn’t made any big changes to the company yet. “Not much has changed in the company.” The company is doing pretty well, so we don’t need to make any big changes to what we’re doing, he said. But he did say a few things about where he thinks the business needs to focus and where he thinks AWS can do well.
Put More Focus On Startups And Quick New Ideas
Startups are one of them, which may come as a surprise. āI believe that as our group has grown…ā Many years ago, when AWS was just starting out, a lot of attention was paid to how to really appeal to developers and startups. This helped us get a lot of early success, he said. Then we thought about how to get bigger businesses, governments, and regulated industries all over the world to be interested in us. And I think one thing I just said againāit’s not really a changeāis that we can’t lose sight of the startups and coders. All of those things need to be done.
Besides that, he wants the team to keep up with all the changes happening in the industry right now.
He told the team, “I’ve been really emphasising how important it is for us to continue to not rest on the lead we have when it comes to the set of services, capabilities, features, and functions that we have today. We need to keep moving forward and building that roadmap of real innovation.” “I believe that customers choose AWS because we offer the best and widest range of services.” People lean on us because we still have, by far, the best operational and security performance in the business, and we help them come up with new ideas and move faster. That list of things to do needs to keep getting moved forward. Though it’s not really a change, I think I’ve stressed this point the most: how important it is for us to keep up the level of innovation and speed with which we’re producing.
He said that he doesn’t think the company hasn’t come up with new ideas fast enough in the past when I asked him about it. “I think the rate of innovation is only going to speed up, so it’s important that we speed up our own rate of innovation as well.” It’s not that we’re losing it; it’s just that we keep being told how quickly we need to keep up with the speed of technology.
Generative AI At AWS
Because of how quickly technologies change these days and the rise of creative AI, he said AWS also needs to be “at the cutting edge of all of those.”
Some experts thought that AWS might have been too slow to release generative AI tools, leaving rivals like Google Cloud and Microsoft Azure with a chance to take the lead. This was soon after ChatGPT came out. But Garman believes that this was more of a guess than a fact. He said that AWS had had good machine learning services like SageMaker for a long time, even before the term “generative AI” became popular. He also said that the business was more thoughtful about how it used generative AI than some of its rivals.
“We were looking into generative AI before it became popular, but I will say that when ChatGPT came out, it was like finding a new area of uses for this technology.” And I think it got everyone excited and pumped up, right? “I think a lot of people, including our competitors, rushed to add chatbots to everything to show that they were the best at generative AI,” he explained.
Garman said that the AWS team wanted to take a step back and see how its users, whether they were startups or big businesses, could best use their own unique data to integrate this technology into their apps. These people will want a platform that gives them the freedom to build on top of. They should think of it more as a building platform than an app that they’ll change. “That’s why we took the time to build that platform,” he said.
For AWS, that platform is Bedrock, which gives it access to a lot of different open and closed models. At the time, he said, it was a bit controversial to do that alone, let alone let people connect different models. He said, “But for us, we thought that was probably where the world goes, and now it’s pretty much a given that that’s where the world goes.” He said that everyone will want models that are made just for them and will bring their own data to them.
He said that bedrock is “growing like a weed right now.”
He still wants to figure out how to make creative AI cheaper, though. “A lot of that is doubling down on our custom silicon and making some other model changes to make the inference that you’re going to be building into your apps [free].”
Garman said that the next version of AWS’s custom Trainium chips will come out near the end of this year. The company showed them off at its re:Invent conference in late 2023. “I’m thrilled that we can fully bend that cost curve and begin giving customers real value.”
As it stands, AWS hasn’t always tried to compete with other tech giants when it comes to making its own big language models. I asked Garman about that, and he said that the company is still “very focused on” those things. He believes that AWS should have both first-party models and third-party models. But he also wants to make sure that AWS’s own models can add value and set them apart, either by using its own data or “in other ways we see opportunity.”
One area of potential is cost, but another is agents, which everyone in the business seems to be very excited about at the moment. Garman said, “I think there’s some innovation that can be done there with making sure that the models can use other APIs and do things in a very reliable and correct way.” According to him, agents will make generative AI much more useful by handling tasks for their users.
Q, A Chatbot Driven By AI
At its last re:Invent meeting, AWS also introduced Q, an AI-powered assistant that can learn on its own. There are two main types of this right now: Q Developer and Q Business.
Q Developer works with many of the most famous development environments and lets you update old Java apps with code completion and other features.
As Garman put it, “We really think about Q Developer as a way to really help across the developer life cycle.” “I think a lot of the early developer tools were very focused on coding. Now we think more about how we can help developers with everything that hurts and takes a lot of work.”
Garman said that Amazon’s teams used Q worker to update 30,000 Java apps, which saved the company $260 million and 4,500 worker years.
Under the hood, Q Business uses similar technologies, but its main goal is to collect internal business data from many different sources and make it searchable through a question-and-answer service similar to ChatGPT. Garman said that the business is “seeing some real traction there.”
Services Being Shut Down
In his speech, Garman said that not much has changed since he took over as CEO. However, AWS recently stated that it is going to shut down some of its services. This summer, AWS announced plans to shut down services such as its web-based Cloud9 IDE, CodeCommit, a rival to GitHub, CloudSearch, and others. This isn’t something that AWS usually does very often.
He said, “It’s a bit of a cleanup process. We looked at a bunch of these services and either launched a better service that people should move to or launched one that we just didn’t get right.” “Oh, and by the way, some of these we just don’t get right, and they didn’t have much grip.” We took a look and thought, “You know what? It turns out that the partner ecosystem has a better answer, so we’re going to use that. You can’t put money into every thing. You can’t build everything. That’s not something we like to do. People who put their money on us to support things for a long time are taken very carefully. We’re very careful about that, so
AWS And The World Of Open Source
AWS has had a tough relationship with the open source environment for a long time, or at least it was thought to have a tough relationship. That’s about to change, and AWS gave its OpenSearch code to the Linux Foundation and the brand-new OpenSearch Foundation not long ago.
When I asked Garman what he thought about the future of AWS and open source, he said, “I think our view is pretty easy to understand.” “Open source is great to us.” Open source is important to us. There’s a big difference between how much we use the open source community and how much we give back to it. We take that very seriously because I think that’s the whole point of open source: to get help from the community.
He said that AWS has put a lot of money into open source and made many of its own projects open source.
Also Read: Opensearch Will Now Be Run by the Linux Foundation Thanks to Aws
“Most of the trouble has come from companies that started open source projects but then changed their minds, which I guess is their right to do.” That’s not really what open source is all about, though. So whenever we see that, let’s look at Elastic as an example. OpenSearch, AWS’s version of ElasticSearch, has become very famous. … We want to lean into any Linux Foundation project, Apache project, or anything else that we can, and we do. We help them out. As an organisation, I think we’ve grown and learnt how to be a good caretaker in that community. I hope other people have noticed.
What do you say about this story? Visit Parhlo World For more.
