I was down in Brooklyn this week attending AI Native DevCon, and let me tell you—there’s no hidden metaphor or clever subtext in that name. What GuyPo and the gang have built here is exactly what it says on the tin: a community and a movement laser-focused on using AI natively in software development. Not bolted on, not duct-taped to the side, but woven into the core machinery of how we build, ship, secure, operate, and iterate on software.
And yes, while the name emphasizes “Dev,” I’m going to go on record right now and say that Ops is absolutely part of this story. You can’t decouple the two—not in 2025, not in AI’s era, not ever.
But before we get into that, let me set the scene.
Brooklyn, Industry City, and the Buzz of a New Movement
This was my first time at Industry City, and even though the temperatures outside were flirting with frostbite levels, inside the venue was on fire. There were pockets of people huddled in every corner—debating agent architectures, arguing about LLM approval chains, sketching out platform patterns on napkins, riffing on the best ways to embed guardrails directly into CI/CD.
This wasn’t performative energy.
It wasn’t the manic buzz of a marketing-heavy conference.
This was the same authentic vitality I remember seeing in places like Austin, Boulder, and even early-days Silicon Valley—before the VC decks and the “influencers” showed up.
This—this—was Brooklyn.
Not Brooklyn-as-Manhattan’s-Little-Brother.
Brooklyn as a fully independent, wildly creative organism.
A perfect venue for a movement built by tinkerers, hackers, platform thinkers, and people who love doing the work.
No shade to Manhattan, but this event simply couldn’t have happened there. You can feel that.
A Familiar Sense of Déjà Vu
And maybe that’s why, walking into the event, I had that odd déjà vu sensation. Because everything about AI Native DevCon felt familiar. Is the community building around this movement? It looks a whole lot like the early-days DevOps tribe.
No less than Patrick Debois—the Godfather of DevOps himself—is involved. When Patrick shows up, you know something real is happening. His presence is like a community seal of authenticity: no hype, no hand-waving, no empty promises. Just curiosity, experimentation, and progress.
At the center of the organizing force is Guy “GuyPo” Podjarny, co-founder of Snyk and now the driver behind Tessl, his new AI-native platform endeavor. GuyPo has always had that rare intersection of hands-on technical depth and community-first DNA, and it shows here. This thing has roots.
Joining him is Simon Maple, another Snyk alum and someone I’ve watched build trust across developer audiences for years. Seeing these two at the helm gave me a warm and fuzzy feeling—and not just because I’ve known them forever.
It told me security has a seat at this table from day one.
Not bolted on.
Not a compliance afterthought.
Not a “phase two” for later.
Baked in.
That alone makes this different from so many tech “movements” that come and go.
The speakers’ lineup is chock full of familiar names and faces. These are people I respect. A great lineup of many folks who literally came from all over the world.
Who Showed Up
Over 300 people showed up for this two-day gathering, and they were exactly who you’d hope would appear at an event like this:
- Hardcore builders
- Platform engineers
- DevOps veterans
- LLM experimenters
- People with Mistral running on a GPU under their desk
- People who talk about agent design patterns the way others talk about football
What you didn’t see?
The hordes of opportunists who show up once there’s money to chase.
It’s still early.
It’s still pure.
It’s still about learning and doing, not monetizing and packaging.
And that, for a movement like this, is priceless.
The Context: 90% of Developers Use AI
One stat that I heard during the event and have heard frequently elsewhere: 90% of developers already use AI in some capacity.
If you talk to dev teams across organizations—from scrappy startups to the Fortune 500—you’ll see that’s dead on. AI is everywhere. It’s in code review, test generation, documentation, build optimization, workflow stitching, data cleanup, and the day-to-day mechanics of actual work. It’s everywhere.
But the usage is fragmented.
Inconsistent.
Full of “rogue patterns” (more on that in a second).
Mostly happening at the individual level rather than as part of a coordinated engineering strategy.
That’s why this community matters.
AI Native Dev isn’t about using AI as a tool.
It’s about building software with AI as a first-class participant in the SDLC. It’s about a future where agentic AI is a bulwark for development and deployment.
Agenda Highlights: The Work Being Done
Looking through the agenda at AINativeDev.io/DevCon, I was struck by how practical the content was. A few standout areas:
- Agent architecture patterns
How do you design agents that collaborate rather than collide?
How do you chain them responsibly?
- LLM governance in pipelines
Not the dry governance of checkbox frameworks—but living governance encoded into platform seams.
- Secure AI usage for developers
How to embed guardrails directly where developers work.
- AI-augmented debugging and triage
This one was a crowd favorite.
Imagine incident triage with LLM copilots that actually understand your stack.
- Building platform teams for AI-native workflows
Finally—FINALLY—the platform engineering conversation moves into AI.
These weren’t “AI is coming” talks.
These were “AI is here and here’s how we’re using it to ship real things” talks.
Simon Maple’s Take on Secure AI for Dev & DevOps
Simon Maple delivered one of the most grounded explanations of secure AI usage I’ve heard yet:
“The path to using AI securely in dev and DevOps starts with platform-level best practices — the guardrails, policies, and context we bake directly into the environment. Platform teams will distribute these across developer spaces, helping rein in some of the ‘cowboy’ approaches we’re seeing in vibe coding today. At their core, these controls are about LLM guidance: giving agents the right context for how we build, not dictating what we build.”
Bang.
Spot on.
We’re not here to limit developers.
We’re here to steer the AI they use so they can build better, faster, safer.
Shimmy’s Take
Walking around Industry City that day—listening to people, watching them sketch, argue, debate, laugh—I had a moment where I honestly felt transported.
I saw the early DevOps days flash before my eyes.
I saw the scrappy invention.
The community-first attitude.
The shameless geekiness.
The sense of “we’re onto something, but we’re still figuring out what.”
And then, just as quickly, I felt something else:
The future.
It reminded me of what Jon Landau said all those years ago in Harvard Square after watching a then-unknown Bruce Springsteen perform:
“I saw the future of rock ’n’ roll and its name is Bruce Springsteen.”
Well, I saw the future this week too.
I saw DevOps’ future — and its name is AI Native Dev.
