I was very surprised when Adam Selipsky stepped down as CEO of Amazon AWS Cloud Computing. Perhaps just as surprising is that Matt Garman followed in his footsteps. Garman joined Amazon as an intern in 2005 and became a full-time employee in 2006, developing early AWS products. Few people know business better than Garman, whose last position before becoming CEO was Senior Vice President of Sales, Marketing and Global Services at AWS.
Garman said in an interview last week that he hasn’t made any major changes to the organization yet. “There is no change in the organization. “The business is doing quite well, so we don’t need to make any major changes to the areas we are focusing on,” he said. But he pointed out a few areas where he believes the company should focus and where there are opportunities for AWS.
Re-emphasize startups and rapid innovation
Somewhat surprisingly, one of them is startups. “I think we have evolved as an organization. … “In the early days of AWS, we focused on how we could really appeal to developers and startups, and that’s where we got a lot of early traction,” he explained. “We then started looking at how we could attract large corporations, governments and the regulated sector around the world. One of the things I just emphasized again is that we mustn’t lose focus on startups and developers, not really on change. We have to do all of that.”
Another area he wants the team to focus on is keeping up with the current whirlwind of change in the industry.
“I have been emphasizing with my team how important it is for us to continue to move forward and not rest on our laurels when it comes to the suite of services, capabilities, features and capabilities we have today. It’s about building a roadmap for true innovation,” he said. “We believe customers use AWS today because we have the best and broadest range of services. People rely on us today because we continue to have industry-leading security and operational performance and help them innovate and move faster. And we need to keep pushing forward with our roadmap of what needs to be done. It’s not a change per se, but it’s probably what I emphasized the most. That is how important it is to maintain the level of innovation and speed of delivery.”
When I asked him if he thought the company hadn’t innovated fast enough in the past, he insisted he didn’t think so. “I think the pace of innovation will further accelerate. Therefore, I can only emphasize that we must also accelerate the pace of innovation. We’re not losing it. “It just highlights how much we need to continue to accelerate to match the pace of the technology that exists today.”
Generative AI on AWS
With the advent of generative AI and how quickly technology is changing now, AWS needs to be “at the cutting edge of all technology,” he said.
Immediately after the launch of ChatGPT, many experts questioned whether AWS was too slow to launch generative AI tools itself, leaving an opportunity for competitors like Google Cloud and Microsoft Azure. But Garman believes this is more perception than reality. He noted that AWS has long offered successful machine learning services like SageMaker, even before generative AI became a buzzword. He also noted that the company has taken a more cautious approach to generative AI than some of its competitors.
“We were looking at generative AI before it was widely accepted, but I would say that when ChatGPT came out, we discovered a new area of how this technology could be applied. And I think everyone was excited and energized by it. Right? … I think a lot of people, our competitors, have competed to put chatbots on top of everything and show that they are at the forefront of generative AI,” he said.
I think a lot of people (our competitors) have been competing to put chatbots on top of everything and show that they are at the forefront of generative AI.
Instead, the AWS team wanted to take a step back and look at how customers, whether startups or enterprises, can best integrate this technology into their applications and use their differentiating data to do so, Garman said. “They’re going to want a platform that they can actually think of as a building platform rather than an application that has the flexibility to build and adapt. So we took the time to build that platform,” he said.
For AWS, that platform is Bedrock, which provides access to a variety of open and proprietary models. Just doing that, and allowing users to link different models together, was a bit controversial at the time, he said. “But we thought maybe that was where the world was going, and now it’s kind of a foregone conclusion that that’s where the world is going,” he said. He said he thinks everyone wants a custom model and will bring in their own data.
Garman said Bedrock is “growing like a weed right now.”
But one problem with generative AI he still wants to solve is pricing. “We are doubling down on custom silicon and other model changes to make the inference that it will be much cheaper to build (something) into your application.”
Garman said the next generation of custom Trainium chips, which AWS debuted at its re:Invent conference in late 2023, will be available later this year. “I’m really excited that we can really shift the cost curve and deliver real value to our customers.”
One area where AWS hasn’t even tried to compete with other tech giants is building its own large-scale language models. When I asked Garman about this, he said it’s something the company is still “very focused on.” He believes it is important for AWS to have its own model while also continuing to rely on third-party models. But he also wants to see whether AWS’s own model can differentiate and add unique value through its own data or “through other areas of opportunity.”
Among these areas of opportunity are costs, but also agents. It seems like everyone in the industry is optimistic about this right now. “I think having models stable at a very high level of accuracy and actually being able to call other APIs and do things is an area where innovation is possible,” Garman said. He said generative AI will be of much greater use by having agents automate processes on behalf of users.
Q, AI-based chatbot
At the last re:Invent conference, AWS also launched Q, a generative AI-based assistant. As of now, there are basically two types: Q Developer and Q Business.
Q Developer integrates with many of the most popular development environments and, among other things, provides code completion and tools for modernizing legacy Java apps.
“We think of Q Developer in a broader sense, as providing practical help throughout the developer lifecycle,” Garman said. “I think a lot of the early developer tools were focused on coding. It was more about how they could help developers with all the things that were painful and difficult.”
At Amazon, the team used Q Developer to update 30,000 Java apps, saving $260 million and 4,500 years of developer time in the process, Garman said.
Q Business uses similar technology internally, but focuses on aggregating company-wide data from a variety of sources and making it searchable through question and answer services like ChatGPT. “The company is seeing real traction there,” Garman said.
End of service
Although Garman noted that not much has changed under his leadership, one thing that has happened at AWS recently is that the company has announced plans to shut down some services. AWS has announced plans to shut down services like its web-based Cloud9 IDE, CodeCommit GitHub competitor CloudSearch, and others this summer, although it’s not something it traditionally has done very often.
“When we look at these services, frankly, they’ve either launched a better service that people should switch to, or they’ve launched a service that hasn’t launched yet. “Make it right,” he explained. “And by the way, there were some of the things we didn’t get right and their traction was pretty light. We looked at it and said, ‘You know what? There are actually better solutions in the partner ecosystem and we will rely on them.’ You can’t invest in everything. You can’t build everything. We don’t like that. We take it seriously when companies invest their business in supporting us for the long term. So we are very careful about that.”
AWS and the open source ecosystem
One of the relationships at AWS that has long been difficult, or at least perceived to be difficult, is with the open source ecosystem. Things are changing, and just a few weeks ago AWS brought OpenSearch code to the Linux Foundation and the newly formed OpenSearch Foundation.
We love open source. We rely on open source. I think we’re trying to leverage the open source community. and Become a significant contributor to the open source community.
“I think our view is very simple,” Garman said when asked what he thought about the relationship between AWS and open source going forward. “We love open source. We rely on open source. I think we’re trying to leverage the open source community. and Make a significant contribution to the open source community. I think that’s the whole point of open source: benefiting from the community. So we take this seriously.”
He noted that AWS has made major investments in open source and has open sourced many of its own projects.
“Most of the friction came from companies that started the original open source project and then decided to de-open source it. I think this is their right. But you know, that’s not really the spirit of open source. So whenever you see people doing that, you can use Elastic as an example, and OpenSearch (AWS’ fork of ElasticSearch) has been quite popular. … If there’s a Linux (Foundation) project or an Apache project or something we can rely on, we’d like to rely on that. We contribute to them. I think we, as an organization, have evolved and learned how to be good stewards in that community, and I hope others have noticed that too.”
