A colleague recently shared an interesting article that points to a big wave of change for developer documentation. InfoQ’s cloud senior editor Steef-Jan Wiggers describes a sudden deluge of AI-generated contributions flooding major open-source projects. The culprit is vibe coding, a software development approach where the user relies heavily on AI to generate, refine, and debug code based on conversational, natural language prompts, often without meticulously reviewing or even fully understanding the underlying code.
In particular, Wiggers reported a metric that should give pause to every documentation leader. In the months after ChatGPT launched in November 2022, something unexpected happened at Tailwind CSS. Downloads of their framework continued to climb, but traffic to their official documentation plummeted 40% and revenue dropped 80%.
What’s happening is this: developers are increasingly getting technical support from their AI tools instead of turning to the formal documentation. Those extensive documentation suites that your team spent weeks, months, and years creating and disseminating are no longer reaching your intended audience of software developers. This transformation raises questions for documentation teams:
- Who are we writing for? If a human reads only the high-level summary, and the AI reads the detailed API spec, what does a good doc suite look like?
- How are audience needs changing? When will we see the first document titled How to Prompt AI Bots to Write Code Using My API or The Vibe Coding Prompt Cookbook?
- How do documentation practitioners demonstrate added value? If doc traffic is down 40% because our customers are successful using an AI intermediary, how do we prove our team’s ROI to the C-suite?
AI coders need different docs
Let’s be realistic, humans aren’t going anywhere any time soon. The programmers who used to write code are being redeployed from the engine room to the bridge. They’re steering the (AI) ship. Those coders are now the architects, the orchestrators. Instead of writing code directly, they write the prompts that direct the AI tools, and those tools write the code. Similar transformations are happening to technical writing, and will spread from there to the rest of industry.
This tectonic shift mandates a radical change in focus for documentation teams. We must deprioritize authoring low-level programming docs (like code examples) and pivot to supporting the human architect at the bridge. High-level documentation about software design, implementation choices, and system-level rationale is not just vital, it’s the new center of gravity.
Why? Because AI handles code generation, but the crucial, high-value decisions—the choices about which system, platform, or API to use, and how to prompt the AI to implement it correctly—are still 100% human-driven. This change isn’t confined to open source; it represents a comprehensive, industry-wide re-alignment of technical documentation priorities. Failure to re-evaluate your documentation strategy in light of this new reality is a mistake, and could erode your product’s position in the marketplace.
Revisit documentation priorities now
Identifying the highest impact technical communication is right in our wheelhouse. Since our founding in 1990, Expert Support has completed thousands of documentation projects for tech companies of all sizes, from startups to Fortune 50s. These assignments are often with software development groups who are creating very complex software, and need complete suites of effective developer documentation.
Over the years, one factor remains constant: technological change. Yet, the fundamental challenges of helping target readers find and understand the crucial information they need to be successful hasn’t changed at all.
If your team is seeing a drop in doc traffic, don’t panic, just start pivoting. At Expert Support, we’re actively working with clients to transform how they think about documentation, emerging best practices, and new workflows for this AI-empowered world. I’d love to have a conversation about what new trends you’re seeing with your customers, both internal and external, and compare notes.
If you’d like to discuss, drop a meeting on my calendar here.
Paul Gustafson is the CEO of Expert Support. A long-time Silicon Valley veteran, Paul has helped hundreds of clients optimize their technical communication strategies. Expert Support is a leading provider of technical writing and documentation services, helping companies ensure their developer content is impactful and ready for the next wave of technology.