00:00Halo, saya Surya Perkasa sedang ada di Amazon Web Service Rain Man 2025 di Las Vegas, Nevada.
00:13Ada apa aja si DXO yang diselenggarakan oleh Amazon Web Service ini?
00:30Pada saat ini memang membawa topik khusus yang lebih banyak membahas tentang agentic AI ya.
00:50Nah, ada sekitar 600-an dan juga yang berisi teknologi mulai dari software, hardware, hingga teknologi terapan fisikal seperti robotik yang digunakan oleh berbagai developer-developer yang ada.
01:07I think it's pretty obvious, this new generation of AI-driven applications is going to drive a ton of innovation.
01:16So, and I think there will be new kind of experiences, new kind of services that will be different than the ones we have today.
01:23Because if you talk to an engineer today, you say, if I code it, they usually go, they look at you like, yeah, whatever.
01:29It's like, if I draw something with my hands, and let's say you ask your engineer to draw something, the initial reaction of people is like, ah, it's not as good as mine.
01:42But that's not the point.
01:43Like, the point is not to make it like mine, it's how you can make it, right?
01:48So, it will create these new kind of applications that are so specific or so different that will have their own marketplace, their own ways, and they will be judged differently.
01:58And that's important, right?
01:59That's important.
02:01One of the things that are still hard is, when you create traditional software, it's very hard to scale it and make it reliable and secure, like the thing we're doing in AWS.
02:13Doing that through web coding is really hard, right?
02:16Because it requires all this, like, all this, it's a very specific set of coding and building set for set.
02:26Enterprises is very specific, right?
02:28However, I don't think that this is going to replace that.
02:31It's just going to create a new kind of thing, right?
02:33I was having my sister, she's based in Canada, and she's helping people with disabilities, little kids.
02:40Then they have a hard time, like, speaking.
02:43So, she just like coding an app that helped them when they put something, the app just sees that with machine learning and vision, and it speaks it back.
02:51She's great, and she did it, and she has, she's a chef, right?
02:55She has no technology, she doesn't have a computer, she did it on her phone.
03:00And she just feels great, and she felt those people feel great.
03:03And I think we will see more of that use case.
03:05So, my name is Neha, I'm a Director of Applied Science in AWS.
03:09Coming up nine years, I have a PhD in Automated Reasoning, worked a lot in the space with IM Access Analyzer, Block Public Access, CEDAR, Verified Permissions.
03:19And I also work in the space of, like, cloud governance, and helping secure proactively.
03:28And so, the security agent is one of the frontier AI agents, a sophisticated class of AI agents that are autonomous, are scalable, and can essentially work for extended periods of time without requiring human intervention.
03:45And the security agent is one of those agents.
03:48What it does, it bakes the security process right from the get-go, into the development process.
03:57So, you don't, so now, security doesn't have to be an afterthought.
04:00It doesn't have to be after the fact.
04:02Because when you discover a security design flaw late in the process, the cost of fixing it goes up a lot higher.
04:12While if you fix the security design issue early on, when you're writing the design flaw, then it just gets fixed at the root.
04:21That's the goal of the security agent, to be a proactive security agent.
04:26That's the goalcier agent in the process, but again...
Jadilah yang pertama berkomentar