How Amazon’s GenAI tool saves developers 4,500 years of work and $260 million a year
Developers don’t want to spend a lot of time on repetitive, mindless tasks. GenAI helps them get out of this situation.
At Amazon in particular, developers previously reported that they spend 70% of their time on boring and repetitive tasks rather than coding, so they’re using new tools like the AI-powered software development assistant Amazon Q Developer to take the grunt work off their hands.
The tool helps developers increase their productivity and code more confidently, while allowing junior developers to ask more questions and learn from examples provided by the 24/7 assistant.
The proof is in the air. Amazon CEO Andy Jassy announced on the Q2 2024 earnings call that using Q Developer’s agents for the code transformation feature, Amazon has migrated 30,000 product applications from Java 8 or 11 to Java 17, saving over 4,500 years of development effort and $260 million annually in performance improvements. It has been available to the general public since April 2023. A little over a year later, it’s easier to understand the impact it’s had on developers.
So how exactly does it work and how does it improve workflow?
Amazon Q Developer Features
Amazon Q Developer generates highly accurate code and can have conversations about that code. These conversations include filtering out code suggestions that may be biased or unfair, implementing new code generated from developer requests, and debugging and troubleshooting.
“A big part of it is that a lot of the work a developer does on a daily basis isn’t that interesting,” said Doug Seven, GM and director of AI developer experiences at Amazon Web Services. “It’s a lot of mundane work. It’s all necessary, but it’s not particularly cognitively interesting or challenging. If AI can take over some of that for the development team and get them to put their cognitive skills into something more interesting and novel that has more business value, that’s great.”
Early indications are that Amazon Q could help a company’s employees increase their productivity by more than 80%. When Amazon Web Services ran a productivity contest, developers who used Amazon Q Developer were 27% more likely to successfully complete tasks. That’s partly because the software is easy to use and doesn’t require training to understand it.
Seven compares it to another developer sitting next to you, watching your work and making suggestions. He says it’s like super autocomplete. But the developer can also ask questions about where the potential problem lies in the code.
“A big part of it is that a lot of the day-to-day work of a developer is not that interesting. It’s a lot of mundane work. It’s all necessary, but it’s not particularly cognitively interesting or challenging.”
Another way to leverage Amazon Q Developer is to use an AI agent. For example, converting code from Java 8 or 11 to Java 17 is tedious and can take up to two days per program. Instead of doing this themselves, the developer can delegate this task to Amazon Q.
“In some ways, he can be used like another engineer on the team, doing the work and then coming back for peer review and checking that it meets the requirements,” Seven said.
And when a developer truly understands what work is best delegated to AI, it can help them achieve real success.
“I like to say that every single developer can now become a development team,” Seven said. “I can distribute my work to different agents and do other things while the agents do that work.”
The ripple effect
Given the huge productivity gains this new tool brings, some may wonder what more developers can do. With AI booming, this question has long been asked, but most argue that AI will allow workers to perform more fulfilling and important tasks. Plus, there’s always a person who needs to be on top of things.
“Asking an AI assistant to write code is not much different than asking it to translate English text into French with the correct semantics,” said Jason Andersen, vice president and principal analyst at Moor Insights and Strategy, where he covers application development and platforms. “In both cases, the requester needs to be in the loop and provide the context of the situation to request the work.”
“Given the complexity, an AI agent will also not truly understand the context of previous or potential decisions and constraints. Therefore, a developer must be brought in to review the AI recommendations and decide whether they are useful.”
There are also other gaps that developers need to consider when using AI. For example, Andersen says that AI today is not smart enough to fully understand the nature of an application that consists of thousands of files spread across many systems and locations.
“Given the complexity, an AI agent will also not truly understand the context of the previous or potential decisions and constraints,” Andersen said. “That’s why a developer needs to be brought in to review the AI recommendations and decide whether they are useful.”
Although Amazon Q Developer has safeguards in place to ensure that an AI agent acts responsibly, not everything the AI does is always right. This is where human oversight comes in again to make sure everything runs smoothly.
“I think I have to say that the role of the developer has been evolving for a long time,” Andersen said. “The range of skills, including collaboration and coordination, is becoming more sophisticated. So AI would need to not only keep up with what developers are doing today, but also evolve with the future evolution of the role. I think that’s in contrast to other jobs that AI could one day replace.”
