Nix Brings Reproducible Software to the Age of AI Agents
Reproducible software used to be a niche concern for build engineers. AI agents have changed that. Teams now generate more code than ever, and every package must run the same way everywhere.
Flox CEO and co-founder Ron Efroni examines that shift with Alan Shimel at KubeCon + CloudNativeCon in Salt Lake City.
From Meta to Nix
Efroni ran developer products at Meta, where slow builds held engineers back. That work led him to Nix, an open source project that makes code run the same way anywhere.
His co-founder, Michael Brantley, had already brought Nix into D. E. Shaw. The platform he built there became Flox, which spun out in late 2021. Efroni is also president of the NixOS Foundation.
Community First
Efroni sees Flox as one of many contributors to the Nix ecosystem. Flox builds on top of Nix rather than competing with it. Its open core layer is free, while enterprise features cover security and scale.
His advice to newcomers is to start with Nix itself. The community needs contributors and growth. Enterprises with harder problems can turn to Flox.
Agents Need Determinism
About a year ago, Flox shifted much of its engineering to an agentic-first workflow. An internal tool called Forge scaled agent use across the company. Efroni measures the payoff in customer results rather than lines of code.
Externally, Flox packages any software or agent with everything it needs to run. That package is locked, so it behaves the same on any OS, cloud or customer site.
Answering the CVE Question
AI is shrinking the time between disclosure and exploitation. Efroni argues the real danger is attackers using AI, not the rising number of CVEs.
Building with Nix produces a precise software inventory. It answers two questions fast: do we have the vulnerable component, and where is it running? No scanning is required.
That is why reproducible software matters more as agents write more code. Efroni suggests starting at nixos.org and trying Flox for free at flox.dev.
Transcript
Hey everyone, we're back here live at KubeCon, continuing our coverage on Techstrong TV. My next guest is Ron Efroni. Ron is the CEO co-founder of Flocks.
If you couldn't tell from your shared or the bottom third on your screen, you see it there. Hey Ron, welcome to Techstrong TV, man. Thank you for having me, Alan.
It's a pleasure here. So before we get into Flocks and KubeCon, give people a sense of your journey. Yeah, definitely.
So my journey into what we're doing today began back at Meta. Really? I was running developer products there, pre-AI, pre-agents.
We were really trying to solve being able to make developer lives a lot better and not having them have to wait 40 minutes for a cold build. And what we did is we actually replicated the entire iOS tool chain, hooked it up to Linux server farms, and really reinvented the wheel so that our developers can build things really quickly and do their job that they're being measured on. And that brought me to a technology called Nix, and a community called Nix, and a project called Nix, that already solved this problem of taking software, taking code, and making sure that it runs the way we want it on anything.
And that blew my mind, obviously. And that's where I was also introduced to my now co-founder, Michael Brantley. Okay.
Who used to be the head of build and release engineering for one of the largest quant funds called D. E. Shaw- Sure ...
for north of 25 years. Wow. And he actually found Nix way before I did.
So he found Nix in 2017, 2018, brought it into D. E. Shaw for very similar reasons, built a platform around it.
That platform ended up becoming Flocks, and we partnered up and we spun that out of D. E. Shaw in late '21, early '22.
So this was a match made in open source project heaven, huh? Match made in open source project heaven. Yeah.
And now I also find myself, among other things, also running, I'm also the president of the NixOS Foundation. So I think today we're the fifth largest open source project in the world. Really?
Based on contributors. Good for you, man. Yeah.
So we have north, I think now, of 11,000 monthly actives. Wow. Good for you.
That's helpful. Thousands of different companies and a lot of scale and that's just super exciting and things are rolling. Good for you.
Good for you guys. Yeah. So is it part of CNCF or no?
It's a separate foundation. Nix is its own foundation. It's not CNCF- That's great ...
which is a Dutch foundation. I love it. Yeah.
That's a great story, man. Now people always wonder, okay, so how do you separate your open source responsibility, obligations, supporting the community- Yeah ... with your commercial opportunity, and where's the line drawn, right?
Some people say, "Well, we're open core, and so we always make sure we maintain our open core, and it's enough. It's a good product. As a project, you could download it, it's usable.
You could make it what you want, all of that. " Is that the relationship between Flocks and Nix, or? So it's pretty much the fact that Nix is its own open source project.
There's a blazing and amazing community that's built around it, and Flocks is just one of many contributors to that ecosystem. That's how I view it. Mm-hmm.
And the other side of how I view it is that Flocks builds on top of Nix. So in the past, we've had a lot of instances where open source projects had competing commercial products. Oh, my installer is better.
Oh, my fork of this- Right ... is better. You should use me over that.
Like that's going to be the reason, yeah. Yeah. And the way we've approached Flocks from the get-go, we're always adding value on top.
And that value on its own has its own open source layer, an open core layer, so anyone can use it free for life. But once we're able to provide enough value on the commercial side, Nix at enterprise, Nix security, Nix at scale, things like that, that's where we differentiate and we really focus and hone in on. So that's what we do, and then everything below that, community first, right?
We just had a Nixcon in Kraków in Poland, and I was asked a very similar question, "Should I try Flocks? Should I try Nix? " Start there.
We need as many contributors, we need as much growth as possible. You want to solve enterprise problems? Come to us, that's our specialty.
Enterprise, right. Right. Yeah.
And it's not that open source projects can't do enterprise scale. No. But at some level, from an enterprise perspective, the economics are go with something a little more polished, a little more enterprise ready, because what you make up in getting things done versus what it costs you to start from pure open source- Yeah ...
it's hard, right? And that's always been the... Not always, but for a long time, that's been the open source model, right?
Exactly. And we have dozens of partners that are not customers, but they're partners and we help them with their Nix setups, and we help them with ecosystem interaction and just celebrating their wins that they have on Nix inside of the broader community. And I think that's just part of being actually a good citizen- Yeah ...
in an open source stack. Well, especially it's your community, right? Right.
So you got to... Unfortunately over the years I've met a lot of friends who started open source projects and built commercial companies out of it. And some of them stayed true to that community, some didn't.
Yeah. I had some people change license. You read about it even now.
Yeah. We saw it, right? Changing licenses and all of that kind of stuff.
I want to move now into the era of AI. Yeah. It's changed everything.
It has. How has it changed Flocks? So internally, a year ago, everything changed inside of Flocks.
Literally, I think it was like- It was just about a year ago. It was. That release right then, about November- Exactly ...
is when stuff got real. We completely moved a large portion of engineering resources to build up and maintain an ongoing agentic-first development and company operational flow. So one of our engineers built a product called Forge internally that allowed us to scale agentic use across the stack, knowledge base, knowledge graph, all that intrinsic stuff.
And that has really just enabled us to catapult the amount of value and the velocity that we're able to put out there into the ecosystem and to our customers. So that's internal. I think externally, it's just that what we solve is that we allow you to take any software, any agent, and then package it with everything it needs to run and lock that package so that the next time you run it, it doesn't matter which OS, which type of metal, customer site, cloud, local, that thing is reproducible and deterministic when you want to have it run.
And I think that is a huge fit because for us, that's what we've always done. And now with agents, right, we've scaled the amount of software that's being created, the amount of criticality of that software, and amount of software that needs to be maintained and managed, and that's our special sauce. That's where we actually come in, right?
So software maintenance and software security are the two things that we solve for organizations like Nvidia CoreWeave and then a lot of others that are public on our website. Absolutely. I got to ask you a question, a CEO question.
Yeah. So one of the big things, all of our surveys and the research we do, and the analysts we've talked to, how many companies are seeing measurable ROI from their AI use? From what you're telling me with this Forge project, you're one of the people who actually are seeing ROI.
For us, it was an internal use. We use Flocks, we use Forge. Forge we don't sell, it's just something that we use, we've given away to friends and partners and stuff like that.
But for us, I think it's generally been a multi-Xer, but how do you measure that, right? Is there a measurement? At the end of the day, because of this move, we've solidified this culture of results and customer oriented results.
And now the question is, how many results do we get to a customer in a period of time? And that has just become much faster with much better customer reception. So there's your ROI.
So yeah. So I'm not going out there and saying, "Oh, we ship 10x more code," but is that good? No, I don't know.
It's not the amount of code, it's quality, it's what they use and everything else. Is it maintainable? Can we actually reuse it?
It's the things in that. Agreed. But it's a great success story.
And you know what? Look, I'm a big believer in AI, right? I wrote this book, it's out there.
I don't necessarily agree with everything we do with AI. Yeah. But I do think the kinds of use you're describing is game changing.
It's game changing. Not that you couldn't do stuff like this before, but the way you can now, the velocity you can, the scale you can. And I think the biggest thing that we care about today is just the amassing of the security problem that's coming in with AI.
Yeah. Right? Folks are talking about CVEs are being discovered faster than ever.
There's an ever-growing amount of CVEs, but I don't think that's a real problem. I think bad actors and malicious, if you have malicious intent- With AI in their hands is the problem ... You had-- CVEs were not...
You always had CVEs. I started a security company in 2001. We had a vulnerability management solution.
Exactly. And- We never got to them all. We never did.
Exactly. And, I think the real problem is what you're saying. It's like, what happens when AI is helping exploit those CVEs, right?
We're seeing the time go down from seven days to three days, now it's maybe even a couple of hours. Minus seven days is what they're thinking now. Exactly.
And, for us, that's like a big boost because when you're using Flocks or Nix, in the most non-marketing terms possible, right, it's like you get a pure software inventory that can answer two simple questions just by the way you build your software. Just by the way you tell your agents to build your software, which is, A, do we have it? Should we care when a new CVE comes in?
B, where exactly is it running? Mm-hmm. No scanning, no tokens involved, just like straight up, systematic computer engineering, that can answer those problems.
No, I get it. We get it. Yeah.
All right. People who say, "Hey, I like what I hear. " What's the best on-ramp for them?
org. That's our community website. Lots of- Spell that?
Nix, so N-I-X- I took OS apart ... org. Don't confuse it, it's not an operating system.
It's also, but you can use it on any Linux machine. Other than Nixies. Darwin, Mac.
dev, D-E-V. And you can check it out. It's free.
You can just try it out right now. One of the most beautiful aspects about Nix and Flocks is that because of how it manages its container stack, like it's being contained not with containers, is that it can run side by side with anything on any machine. So you can try it out, spin it down, and everything works just the same way.
But yeah, so I definitely give a shout-out to try out the community that opens their stuff as best as possible. So Nix, Flocks. Yeah.
Excellent. Hey, Ron, thank you, man. I appreciate you.
Hey, y'all. We're live. We're in Salt Lake City.
I think we still got more to come for today, day one. Stay tuned. You're watching Techstrong TV.