In early 2025, Hacker News hosted a thread with a disarmingly simple title: "What even is DevRel in 2025?" It was posted by someone who works in developer relations and was genuinely asking. They'd noticed that every company hiring for DevRel seemed to want something entirely different, and that the role had become impossible to define.

The responses were a post-mortem on a decade of developer marketing investment.

Conference talks. Technical blog posts. Stack Overflow presence. Discord server moderation. Meetup sponsorships. The canonical DevRel toolkit, assembled between roughly 2015 and 2022, was described in the thread as producing output that "looks busy" without anyone being able to explain what it actually moved. One commenter put it plainly: "Most DevRel programs are built around outputs rather than outcomes."

The problem is structural. And the timing is particularly bad for brands that spent the last five years building that kind of presence.

Where the Audience Went

Developers used to find answers in a predictable sequence: search Google, land on Stack Overflow, browse documentation, check GitHub issues, maybe ask on a forum. This sequence shaped where DevRel programs built presence. Stack Overflow badges. Google-optimised technical tutorials. Active GitHub repos with responsive maintainers.

Then developers started asking ChatGPT instead.

This is not a marginal shift. The 2025 Stack Overflow Developer Survey confirmed what most developers could observe from their own behaviour: AI tools have become the primary first-stop for technical questions. The forums, the documentation sites, the carefully SEO'd tutorials — they still exist and still get traffic. But they're no longer where the answer-seeking starts.

DevRel programs optimised for a discovery path that has been rerouted. The blog post that ranked #1 for a technical query in 2022 now competes with an AI response that synthesises information from that post and a hundred others. The Stack Overflow answer that earned reputation and brand visibility now feeds a language model that doesn't attribute sources.

THE TIMING PROBLEM

Companies that built developer marketing infrastructure between 2018 and 2023 — content machines, Stack Overflow presence, conference speaker pipelines — are watching the ROI on that investment erode in real time. The channels didn't disappear. They just stopped being where the audience starts its journey.

Show HN vs. Tell HN

Hacker News has two post formats that reveal everything about how developer communities evaluate new things: Show HN and Tell HN.

Show HN is for things you've built. You put up a link, describe what it does, and the community tries it. Tell HN is for information the community should know. The asymmetry matters: Show HN posts get the benefit of the doubt. The community's default is to engage with something that demonstrably exists and can be evaluated. Tell HN posts face instant skepticism. Telling Hacker News something requires credibility you've already established through demonstration.

Most developer marketing is structured as Tell HN. We believe in developer productivity. Our platform empowers engineering teams. We're committed to the open source ecosystem. Claims, without demonstration. The developer community's response to claims without demonstration is the same whether it's on Hacker News, a conference stage, or a sponsored blog post: polite indifference at best, active dismissal at worst.

# What developers flag immediately in brand content

❌ "We empower developers to ship faster."
❌ "Our platform is developer-first."
❌ "We're committed to the developer ecosystem."
❌ "Loved by 10,000+ developers."

# What earns engagement instead

✓ A working demo they can run in 30 seconds
✓ Benchmarks with methodology attached
✓ "We found X bug in our own system. Here's how."
✓ A CLI tool that solves one specific problem
✓ An RFC posted for community comment before building

The Authenticity Test Is Technical, Not Tonal

Gaming communities test brand authenticity through cultural signals — language, timing, references. Developer communities test it differently: they test technical accuracy.

A developer reading brand copy will notice, immediately, whether the writer understands the technical context. This is not about jargon. It's about precision. A blog post that describes a system architecture incorrectly, uses a term slightly wrong, or oversimplifies in a way that reveals misunderstanding — these are the tells. They signal that the brand hired someone to write developer content without giving them the knowledge to write it correctly.

The community is not hostile to brands on principle. They're hostile to brands that perform technical understanding without having it. The tell is almost always in the details: a description of how something works that doesn't quite match how it actually works.

Technical credibility cannot be purchased. It can only be demonstrated. And the community will check.

What the DevRel Role Actually Became

The Hacker News thread surfaced a secondary problem: the DevRel role itself has become incoherent. Companies hire for it with wildly different mandates — some want content creators, some want community managers, some want sales engineers, some want evangelists. The person in the role often can't explain their own job to a skeptical CFO because the job was defined around activities rather than outcomes.

This is a symptom of the same underlying problem. DevRel was built as a discipline for a world where presence in developer spaces translated into brand recognition, which translated into product adoption. The middle step — presence into recognition — worked when developers discovered things through the channels DevRel occupied. Now that developers start their journey differently, the conversion chain breaks.

The brands that have maintained developer trust through this transition share a characteristic: they never relied on the channel doing the work. They built things the community found useful and let the community do the distribution. Open source libraries used by thousands of projects. Internal tools that became standards. Documentation so good it spread by recommendation. These don't depend on any particular discovery channel because developers who benefit from them tell other developers directly.

What Works Now

Developer communities in 2026 are not unreachable. They're just unreachable by the 2019 approach.

01 Ship something useful, not something promotional. The currency in developer communities is tools, libraries, and knowledge. A CLI that solves a real problem earns more trust than a year of conference talks. "Help First" programs that create genuine value without requiring conversion outperform campaigns that gate content behind a signup.
02 Technical specificity over values claims. "We use Rust for our core processing pipeline because of X, and here's the tradeoff we made with Y" is trusted. "We're committed to performance and reliability" is not. Specificity is its own form of credibility — vague claims are easy to make, specific ones aren't.
03 Be findable where AI looks, not just where people search. If developers are starting with ChatGPT, the question isn't "what does our SEO rank?" — it's "does our documentation, codebase, and public writing contain the accurate technical information that AI systems will surface?" Different optimisation target entirely.
04 One deep community beats ten shallow ones. Developer communities are highly networked within domains. A brand that becomes genuinely useful to the Rust community, the Kubernetes community, or the ML infrastructure community will be discussed by members across the ecosystem. Scattered thin presence in many places achieves nothing in any of them.

The DevRel playbook being dead is not a reason to stop trying to reach developer audiences. It's a reason to stop trying to reach them the way everyone else was. The community hasn't become less influential — developers still recommend tools, spread knowledge, and make purchasing decisions for their organisations. The reach is there. The access routes just changed.