Changing Software Development with Frontier Agents

Key Takeaways

Considering where AWS sits in the market and everyone that attends reinvent, probably the the the question most people would have is if I'm a large enterprise and I have multiple teams, multiple desks doing multiple things, how do I deploy three of these frontier agents? >> Yeah, let me start off with uh saying uh and here at AWS we are building the foundation for billions of agents uh and that's an understatement. So what we launched with these frontier agents uh is essentially a new category of agents because uh while so much of activity is happening in agents they are historically been more assistance to humans and while we deploy it even within Amazon tens of thousands of developers to help with software development we were actually wanting to push the envelope and being able to actually do lot more where they are actually autonomous where you don't need humans to constantly steer them and they're massively scalable and they can run longer >> to be autonomous. They need to have a an end goal. >> That's right. >> How does that work? So, it's a case of saying, by the way, you use the the the the term AI assistant. What we're now really talking about is AI coworker. >> That's right. And that is the key thing here, which is like what these frontier agents are doing is completely transforming how software is getting done in these teams. So much activity happens primarily in like software development but now these agentic teammates they show up as another teammate in uh for software development they pull up uh actually uh jobs from like GitHub uh task and start coding and clearing backlog and if you're in the middle of the night actually getting paid to deal with an issue happening with your website this agent first takes the first page and says oh let me look into what's happening and unpack and then find out the root cause so that you can actually quickly resolve it and go back to bed like I've been on call and I know I could use that help and be proactive so that those issues don't happen and third is actually a security teammate right >> so it shows up again and most of security has always been about afterthought one of the things we are changing the game here is actually meeting uh developers even before they write a single line of code when they're designing software we catch them right there and here are the things you should be aware of and when they write code we say oh don't do that do this and actually do penetration testing all before shipping. So now you can view it as like you have the best-in-class developer ops and security engineers showing up as virtual teammates for every software. You say best-in-class. What evidence does AWS have that in real world deployments running these three agents in parallel autonomously has a clear benefit? Either efficiency, productivity, what data can you share? >> Yeah, I'll give a couple of examples. One is uh in the DevOps agent that we launched uh we ran it across thousands of escalations that happened uh this year uh alone on like incidents and so forth. these agents actually correctly identified the root cause 86% of time right >> that is like really impressive state >> 86 86% >> 86% and the second one is Commonwealth Bank of Australia they actually have a huge uh cloud infrastructure running across,700 accounts and they put this DevOps agent test uh to test and uh for a very complex networking stack to debug it and what uh they said what it would take them like hours to debug this was able to do it in minutes and these are just few simple examples. Same thing what's going on with security agents. SmugMug is a great uh photo uh company and they are I'm a big customer and they are completely automating the security uh operations al together using like AWS security agent and these are all the beginning of how we this is going to change the game in a big way because what you want these agentic teammates to be is always be there and you delegates on the boring drudgery associated with like ops and security and backlog on development so that developers are empowered to be extremely creative and you will start seeing 5 to 10x improvement in productivity in a big way. To recap, it's three agents. Curo is software development. You have the AWS security agent, the AWS DevOps agent. They are designed to run for hours autonomously. To some that's disconcerting, right? And I go back to my earlier question on you know how uh much concern is there in the last segment. Colleen talked to me about observability understanding >> in real time what's doing but that sounds to me like it's human supervision. >> It's a great question. I mean this is the nuance and the art of balance we have to do. One you want these agents to be autonomous and massively scalable but you also want them to uh not go off the rails. So first we have built it with the right guardrails and second even as part of development workflow these agents can go actually take a goal from uh a developer to say hey go work on upgrading this uh piece of code to the latest SDK and then start upgrading but then as a final step it it might send it for a code review to the human and say take a look and make sure I'm okay and once they say I'm okay I think it's fine or they can even configure it to say if it passes all these tests you're good to ship So that you have built-in guardrails either through human or automated tests and these are the kind of uh things now we are able to actually innovate with the customers as well and even there we have done some pure innovation like you have introduced automated mathematically provable pro property based testing. So that means now developers can specify their goal and we can mathematically prove with things like Ko that whatever it is generating can be tested exactly that way and it's accurate. So that is like another game changer.