All writing
DevRelAICareerDeveloper Tools

The Developer Advocates Are Dead, Long Live Developer Advocates

March 4, 2026

In late 2023, I was trying to figure out what to do with my life.

I'd just left Houdini Swap — a crypto exchange I co-founded that processed over $1.6 billion in transaction volume. Before that, I was a full stack engineer at FitRankings, building software and running product demos for enterprise clients like H-E-B and government contracts with the US Air Force and Space Force.

I was a decent engineer. Self-taught, started coding at 27. I could build things. But I was honest with myself — I was never going to be a principal engineer at Google. What I was really good at was building things AND explaining them. Taking something technical and making a room full of non-technical people understand why it mattered.

So I started looking at developer relations roles. Developer advocacy. The job where you sit at the intersection of engineering and people.

And every single person I talked to said the same thing.

"Don't do it. DevRel is dead."

The Funeral

They weren't making it up.

In July 2024, the DevRel industry had its official funeral. Multiple eulogies were delivered in the same month.

Swyx — one of the most respected voices in developer tooling — wrote "DevRel's Death as Zero Interest Rate Phenomenon." His thesis: what died wasn't developer advocacy itself. What died was the ZIRP version. The version that grew fat during the era of zero percent interest rates and infinite venture capital.

He cataloged the symptoms:

  • Two blog posts a month counted as acceptable output
  • "Day in the life at " recruiting videos counted as DevRel content
  • Traveling to 6+ conferences a year, each burning 1.5 weeks of productivity
  • People with zero engineering experience getting hired as "developer advocates"
  • Teams that deliberately chose unmeasurable metrics, then acted surprised when the CFO cut the budget

His verdict: "ZIRP DevRel is dead, and it's thankfully finally okay to say it out loud."

Keith Casey went even harder. His piece, "Developer Relations: A Painful Reckoning," argued that DevRel died because DevRel killed itself. Five specific failures:

  1. Marketing figured out how to talk to developers. DevRel was no longer the only team that could do it.
  2. DevRel chose terrible metrics — at one point literally claiming the role was "unmeasurable."
  3. DevRel pissed off their own engineering teams — engineers were told "no" to conferences while watching DevRel staff post from exotic locations.
  4. Most DevRel people weren't actually developers. Strong marketers, never shipped code.
  5. The longer you're in DevRel, the more your technical skills atrophy.

Lee Briggs wrote "The Death of Developer Relations" in December 2024 after attending major tech conferences and realizing he could count on one hand the number of DevRel people he ran into.

The layoffs were real. 262,000 tech workers fired in 2023 alone. DevRel was first on the chopping block:

The 2024 State of Developer Relations report found that "proving business impact" jumped to the #2 challenge DevRel teams face, up from #7 in 2023. When budget season came, there was nothing to defend.

I saw all of this in real time. I took the hint. I didn't pursue DevRel.

I went a different direction entirely.

What I Did Instead

I went all-in on AI.

I enrolled in Gauntlet AI — one of the most intensive AI engineering programs in existence. 80-100 hours a week for 10-12 weeks. Not a bootcamp. A gauntlet. I came out the other side as an AI engineer.

I started building things. MCP servers. Open-source tools. Claude integrations. A project called ThinkGate that automatically routes prompts to the right Claude thinking mode.

Then I started teaching. First at Alpha School — kids and high schoolers. Then college students in Puerto Rico. Then I launched an AI meetup in Austin called AIxEDU. I built ITeachYouAI because I realized the same thing everyone in this space eventually realizes: the tools exist, but an estimated 84% of the planet has never used AI seriously. Someone has to bridge that gap.

My meetup went from 0 to 150+ members in four months. Every month it gets bigger. The people showing up aren't engineers. They're marketers, small business owners, educators, creatives. People who never thought they could build software, trying to figure out where AI fits.

I'm building open-source tools in public. Shipping MCP servers and Claude integrations on GitHub. Doing 1:1 consulting with people who want to understand how AI fits into their work. Standing in front of a room once a month showing non-technical people what's possible.

One day I looked up and realized something that hit me like a truck.

I was doing DevRel. I had been doing it the entire time. I just didn't have the title.

The Resurrection

While I was busy not pursuing DevRel, something happened.

DevRel came back. And it came back bigger than ever.

Swyx — the same guy who wrote the funeral — published a follow-up in October 2025: "DevRel is -Unbelievably- Back." Google searches for "developer relations" hit all-time highs. His updated thesis: "For the first time in 5 years, people are strongly believing that Enterprise sales is helped by having great bottom-up adoption."

Then he tweeted something that crystallized everything:

"Anthropic has had a Head of Developer Relations role open for the last 4 months, despite offering $385-460k salary."

That's not a dead job title. That's a job title so important that even $460K can't fill it fast enough.

Here's who's hiring right now, February 2026:

  • Anthropic: Founding DevRel Lead, DevRel for MCP, DevRel for Claude Developer Platform — $385K-$460K
  • OpenAI: Developer Advocate, $220K-$345K
  • Google Cloud: AI Developer Advocate (hired one this week)
  • JetBrains: AI Developer Advocacy Team Lead (just promoted one, hiring more)
  • DeepLearning.AI, NVIDIA, OpenRouter: All hiring DevRel

But the signal that matters most: Lee Robinson left Vercel to join Cursor as VP of Developer Experience.

Lee Robinson. The most recognized DevRel person in frontend development. The guy who basically IS Next.js content. He left the hottest web framework company on the planet for an AI coding tool.

Cursor hit $1 billion in annual revenue in under 24 months. SaaStr called it the fastest value creation in B2B history. $29.3 billion valuation. 1 million daily active users. And what they needed wasn't more engineers. They needed someone to help people understand the tool.

OpenAI grew their go-to-market team from ~50 to 700+. "Forward-Deployed Engineer" postings rose 800% in 2025. Replit hit ~$253 million ARR and their CEO said: "We don't care about professional coders anymore."

The AI coding tools market is $7.37 billion today, projected to reach ~$24 billion by 2030.

Every one of these companies needs the same thing: people who can build AND explain. Engineers who sell. Technical people who inspire.

Developer advocates.

The Gap Nobody Can Close

Here's the data that makes the case undeniable.

The tools are everywhere. The understanding isn't.

85% of developers say they use AI tools regularly (JetBrains 2025, 24,534 respondents). But only 29% trust the accuracy of those tools (Stack Overflow 2025, down from 40% the year before). The gap between "I've tried Copilot" and "I trust it enough to integrate it into how I work" is massive.

Overall positive sentiment toward AI tools fell from 70%+ to 60% (Stack Overflow 2025). A randomized controlled trial from METR studied 16 experienced open-source developers across 246 tasks and found they actually took 19% longer when using AI tools. But — and this is the insane part — they believed they were 20% faster. They couldn't even tell they were slower.

Zoom out further:

This is an education crisis disguised as a technology problem.

You can't solve it with documentation. You can't solve it with a chatbot. You need a human being who can build things, explain things, and make people feel like this is for them. Not scary. Not replacing them. For them.

That's what developer advocates do.

The Plot Twist

Here's the part that makes me laugh.

The old critique of DevRel — the one that was supposedly the nail in the coffin — was that DevRel people "aren't real developers." They can't ship production code. They're just good at talking.

In the ZIRP era, that was a real problem. You needed technical depth to be credible, and a lot of DevRel people didn't have it.

But in the AI era? It's a superpower.

When AI writes the code, the value shifts. The developer's role is moving from "writer of syntax" to "architect of intent." The value isn't in typing code — it's in knowing what to build, why to build it, and being able to communicate that clearly to other humans.

I'm technical enough to build an MCP server. I'll never be a staff engineer at a FAANG. And I think that's exactly the right profile for this moment.

Because the gap isn't in the code. The code is getting commoditized by the day. GitHub Copilot writes 46% of its active users' code. Copilot is now a bigger business than GitHub was when Microsoft acquired it.

The gap is in the humans.

The millions of people who need someone to show them what's possible. Not through a sales deck. Not through a webinar. Through building something live and saying "you can do this too."

The Trench Generals

Stanford launched CS146S — "The Modern Software Developer" — a 10-week course on prompt engineering, agent architecture, MCP, and agentic development. Replit's CEO said they're done focusing on professional coders. The entire industry is screaming the same message: the future isn't code-first. It's intent-first.

And the guides through that transition will be developer advocates.

Not the old kind. Not the conference-hopping, metrics-dodging, never-shipped-code kind. The new kind. The ones who build real things, teach real people, and aren't afraid to say "this part sucks" when it does.

The trench generals of the AI era. The frontline leaders responsible for onboarding millions — eventually billions — of people onto AI tools. Not through marketing. Through showing them.

Two years ago, everyone told me DevRel was dead. I listened. I went a different direction. I became an AI engineer, started a community, built tools, taught people.

And now I'm doing DevRel anyway — because it turns out the world needs it more than ever.

The developer advocates are dead.

Long live the developer advocates.