Beyond Code Generation: AI Agents Grow Up

Hey everyone, welcome back here to techstrong TV. When we scheduled this, I was a little bummed out that we took it this long to schedule because I wanted to just pick up the phone and call him right away. But I finally got David DeSanto, CEO of Anaconda, on here to talk about the news this week around their acquisition of Kilo Code.

And we're going to talk about more, but for those who've watched techstrong TV, David's a frequent guest, friend of mine. I love having him on here. Let's welcome him.

Hey, David. How are you? I'm doing great.

Thanks for having me on. All right. com about the Anaconda/Kilo acquisition, and I do encourage you all to go read that.

But David, I'm going to skip the usual who's David DeSanto stuff. I think people know who you are, former GitLab CEO at Anaconda, and I'm going to skip who Anaconda is, right? The people who bring Python to you, but now so much more.

And let's jump right into the Kilo acquisition. Why now? Why at all and why now?

Yeah. Great question. Right before that, I just wanted to say, too, my heart goes out to everyone impacted by the Canada wildfires.

Oh, yeah. Yeah. I live on the East Coast.

It is definitely hitting Pennsylvania, and so I can only imagine how bad it is. So yeah, my heart goes out to them. Same thing with my wife.

Yeah. Let me take a moment on that. So I spoke to a friend of mine.

I had an eight o'clock business call this morning. A friend of mine up in the Boston area, and he said it's terrible there. Beautiful sunsets, but yet knowing the pain that those beautiful sunsets, what that's causing is tough.

And while we're talking about it, I got to tell you, I've spoken to my friends in Colorado, and it's not pretty out there either. Yeah. I've heard from Anaconda employees across the country, how it's impacting them, and we're just feeling the after effect of what's actually impacting people really bad in Canada.

So it just felt like the right thing to call out at the beginning here because I know you and I both- I agree. Yeah. But, to answer your question, so why Kilo?

Kilo, I've been following since it launched. I've been very impressed by how they've been focused on open core, source available. They took an approach that's very similar to Anaconda in that Anaconda's always been agnostic, whether that's the cloud provider or the infrastructure, the development environment that you're trying to build within.

And now they've done that with AI agents, AI models, providing access to over 500 AI models across 60 providers, giving you the ability to have full transparency into how the agent is working, and honestly, how much it's costing you, too, at the same time. And then why now? Anaconda, we're building out our AI native development platform.

We've come a really long way with it, starting earlier this year with the efforts we've built, plus bringing in Outerbounds. And now with Kilo joining, we can now actually provide full visibility end to end from the beginning of you building your AI native application all the way through to it being in production, and then taking that knowledge and feeding it back over to the beginning. And agentic development, we've been talking about it, not just you and I, but the industry for several years now.

But it's only really been this year with some of the breakouts with what you could do with agents, the models that can run, running things sometimes within your own perimeter. All of that is now going to be part of that Anaconda experience. It's really going to help people really build some really awesome apps.

Absolutely. David, I feel like we'd be doing a disservice to our audience if I didn't-- I went to law school. You know this.

I got to build a case. Yeah. And building a case means building a foundation.

And so I want to take them through this. Right? Pre-David Anaconda.

Anaconda's a powerhouse. They are the source of Python in many ways for most people on the internet. Right?

The leading repo and all of this is where people go to get their Python. You came in right around the time of AI just changing everything, right? And you immediately incorporated, internalized what AI means to Anaconda's business model and started building, let's call it an AI first, if you don't mind.

An AI first new Anaconda. Explain that piece of it, then we'll do the first acquisition, then the second acquisition. Yeah.

Sounds great. Yeah. So, for those who are not familiar, Anaconda has been providing secure Python packages for over 14 years now.

Those started off as data science, moved into machine learning, and now it's all about AI. And realistically, if you look at Anaconda as a company last year, and before that, really a software supply chain security company, because you have secure packages fully built by Anaconda, and you know their lineage, and they can be trusted. And so last year, when the company started looking at bringing AI models in as another type of artifact that people could trust Anaconda with, it was this pivot for the company to start thinking about more than just the local developer environment.

That's where Anaconda's always lived, whether it's a data scientist or a software developer. And when I started thinking about the trends I was seeing in the industry and the things that I know Anaconda's technology could do, it became very clear to us that we should build on top of that success and give people a larger foundation they can trust. And so that's what got the Anaconda platform launched.

Actually, my first talk with you joining Anaconda is the week that the AI platform went live at re:Invent. And so- Yes ... it's been an amazing ride so far.

I'm 10 months in, and I know we just have so much more we could be doing for users. Just a quick plug. That interview is available on Techstrong TV.

If you check out, I guess that would be 2025, our AWS re:Invent 2025 coverage, it's in there. " I don't want to use the term fold in cheaply, but there's technology, people, processes we could bring in here that are going to help me achieve this faster than we could if I had to do it all organically. Yeah.

Let's go to that first acquisition. Yeah. So shortly after launch of our platform, we realized one big thing, Anaconda is great at the development side, but we've never really gone over to the production side of that process.

And so, as we start to build out a roadmap, we realize, hey, there are companies who are doing what we want to do that we haven't started on really yet. And it was great to meet the founders of Outerbounds. Actually had a really great conversation with one of them, Ville, at re:Invent, about what they were building, what we were building, and it became obvious that there's very little overlap between the two.

They wanted to get closer to the development side, we wanted to get closer to the production side, and we realized it would make a really good combination for the two to come together. And for those who are not familiar with Outerbounds, just to kind of give you the sense of its DNA, it was started by developers at Netflix who were trying to improve AI pipelines, machine learning for Netflix. And Netflix allowed them to open source it and then support them leaving and starting the company Outerbounds.

It runs right on top of that. And so now you have- Very cool ... that trusted development side that could really scale for you to make sure things are working as expected.

And now the production side to know that it's gotten there securely, it can scale, and you can monitor it. And so that really gave us that end-to-end component. But to your next question, the big gap that we were still trying to decide what to do was, how do you get agents in there?

Because everyone talks about coding agents, and Kilo started focused on coding agents. But really the world has changed beyond that. Coding via AI is becoming more and more commoditized, and the power of agents is not can it write code, it's can it help you architect something securely, can it help you document what's going on, can it help you troubleshoot a problem.

And the last nine months for Kilo was launching a code review agent, a security agent, an architect agent. And if you bring those into that same environment that's at AI first, as you called it, and what we usually call AI native, all of a sudden, it's not just about fixing and writing good quality code, it's actually helping you get things into production that are probabilistic, because that's the big thing with AI is it's probabilistic, it's not deterministic. And you get that out there with the highest level of confidence possible.

And it's only been a couple of days since the acquisition closed. We've already shipped integrations between Anaconda's desktop and the Kilo desktop experience, and our goal is just to make sure that everyone can really benefit from the trusted foundation that is Anaconda as you're using agents that were created as part of the Kilo offering. Excellent.

Thank you for doing that. David, a question when I was writing the article, was a question in my mind I didn't have the answer to, and I said I'll ask him when I talk to him. One of the greatest accomplishments, I think, of Kilo, is how quickly they've been adopted.

Mm-hmm. I think off the top of my head, was it two million developers or something like that? Over three million.

Three million. Have you looked at the overlap, right? Is this an adjacent population, or is there a big overlap between the two, the Anaconda community, the Kilo community?

I'm sure there's some overlap. Our community is primarily focused around Python, data science, and AI. They have been supporting multiple programming languages.

Some of those are Python. And so there is some overlap, but I think the uniqueness is that it's not complete overlap. And so if you can take those three million users that trust Kilo every day, I even saw some amazing articles of what people have done with Kilo, people have shared over the last couple of days, and you now give it that same foundation that Jupyter Notebooks have because of Anaconda, our desktop experience where you can build your own agents, all of that, and now you bring in that technology and that group of people, that community, you really take Anaconda's community from 52 million to almost 55 million.

And when you think about open source and how important it is to Kilo and how important it is to Anaconda, that's a really large community of people that you can help change the world and help people be more secure and more effective with AI. Excellent. You probably saw this week, right, IBM's stock got hammered.

They had a bad quarter, right? Yeah. Or they didn't actually have a bad quarter.

They pre-announced a bad quarter. They didn't actually report. They pre-announced.

But they got hammered, and it had shockwaves through software. Part of that was Starbucks, the coffee people, said they're going to spend $400 million less on software this year because they're rolling their own, basically. Right?

Mm-hmm. That kind of stuff, does that play into the Anaconda strategy or something you've got to keep your eye on to worry about potentially as well? Yeah.

I think Anaconda has a very unique place in the market. We sit between the physical hardware and then the application that's built to use that hardware. And if you think about it, we're the largest provider of NVIDIA CUDA packages, so you can build on top of NVIDIA GPUs.

Mm-hmm. We're sitting in that level, so when people are building these new applications that maybe, say they don't buy another commercial product. com.

Yeah. Exactly. Right.

We're sitting right below that, and so in some ways, we're helping people build their ideas. In other ways, we're helping companies be more effective at how they're spending their money. The thing that brought me to Anaconda was the support of the regulated market, and how a lot of that is heavily regulated, also sometimes they're offline and air-gapped.

And those companies are trying to do what you're talking about, build the applications they need for either internal or for their own customers, and we can be there to help them along the way. The best way to kind of visualize that for people, and they can go check this out, but I'm doing a blog series right now. " And in partnership- I've actually seen it.

Yeah. In partnership with NVIDIA, I'm using a DGX Spark Box, and the first two blog posts are live. The first one is me using it as my supercomputer.

The second is me using it as inference. And I'm working on the third one, and check that out. It'll really give you a sense as to what Anaconda is doing, how we're different, and how we're helping people achieve the goals they have for their company.

Excellent. Let's turn back to the Kilocode acquisition. So just a little bookkeeping.

You announced the deal. Is it closed yet? When is it expected to close?

It is closed. It's closed. Yeah.

It closed, and we announced it publicly two days ago. Yep. Well, by the time people watch this, because it's Friday, we recorded this on Friday, they may not get to it till Monday, Tuesday.

Last week. It was announced last week. Yeah.

That's a better way to put it. Yep. So that takes that out of the way.

I assume all the employees basically are staying on here, and this is not a "I'm just buying the technology and cutting beta everywhere else," right? No. The majority of the team joined, and they're going through their Anaconda onboarding this week and getting computer set up for their house.

We're all remote, so- Right ... we help people make sure they're set up for success with that, and yeah. So they're here.

They attended our all-hands, our first one since they joined yesterday. And the teams are starting to integrate in and start planning out accelerated roadmap based off the fact that some of the things we're going to build in the future here are now part of our platform. So you mentioned all remote, and it triggered a brain cell here in my head, and that's the Sid connection, right?

The GitLab heritage here. Of course, David, you were six years Chief Product Officer over at GitLab, and Sid Sibrandi was the CEO founder. I believe he's still Chairman of the Board at GitLab.

Yeah, he's Executive Chair. And he's also, I believe, the founder and probably, I guess, maybe lead investor with Kilocode. I'm not going to ask you to give me behind the scenes about whether it's how close your and Sid's relationship had to this deal.

We're going to leave that in the mystery. But that GitLab culture of an open book, of all remote, of everybody knowing what everyone's doing, there was that... It was really a very special culture, right?

And I interviewed someone from GitLab recently when they announced sort of the CEO put up the blog post reorging and what they're going to-- building it for an AI agentic future. I hope they don't lose that kind of special... It was a little crazy when I first learned about it from Sid himself, but I came to appreciate it.

So it sounds like you bring that to Anaconda. Kilocode had that, so it's a good cultural fit off the bat. Yeah.

I agree with that. And so for those who don't know, when I joined GitLab, the culture was a little bit of a shock, if I'm being honest. I remember something Sid said to me in my interview, though.

" And when I started thinking about that, what it really told me was that if you can be open and as open as possible with a community, they get excited. You build together, and really great things come out of that. And one of the reasons why I joined Anaconda, is that I could see that in Anaconda's culture as well.

It started as an open source project, eventually became a consultancy, and then a product company. And we're very fortunate to have Peter Wang, one of the two co-founders, still on the team, as Chief AI Officer in the leadership team. And I would say that, yeah, looking from the outside in and now from the inside out, Kilo had those same values.

And so if you talk about culture overlap, there's a lot of overlap there. I think people are appreciating, even at the size Anaconda is. We're nowhere near as big as GitLab, but way bigger than Kilo that we can still have transparency.

I post regular videos on Loom for the company to kind of follow along with what's going on and from my point of view. And so I think they really appreciate that, and I think it's really helping us build not only a healthy culture internally, but with that broader 52 plus million community. Absolutely.

Hey, David, we're well past 15 minutes, but I knew we were going to blow past that anyway. But we do have to kind of wrap up. First of all, congratulations.

Thank you. Actually, I got one more question for you, if you don't mind. Sure.

And I put it in my article, too. I did a little speculation. Are you done?

Are you done acquiring companies here? Do you have everything you need, or what could we expect to see without getting in trouble? Sure.

What I would say is that we have a very ambitious roadmap of where we want to get to, and we want to get there as fast as possible because we know organizations are struggling today with AI native. And so I would always say there's never not something that could happen, but there's also times where you're definitely not doing anything. And so I'll shroud that in mystery for you and- Okay ...

and the viewers, but I'm just really excited that the Kilo acquisition worked out. We closed it probably the second fastest any acquisition I've closed in my career, and they're already interconnected with the former Out of Bounds team and the broader Anaconda team. And so, yeah, I'd just say stay tuned to everyone as we deliver on the promise of our platform.

Love it. Hey, congratulations, my friend. Good work.

Keep it up. We may see you in Black Hat in a couple of weeks in Vegas, yeah? Yes.

Yeah, I will be there. We'll catch up for sure. David DeSanto, CEO at Anaconda, announcing their recent acquisition of Kilocode.

You're watching Techstrong TV. We'll be right back.