𝗧𝗵𝗲 𝗗𝗮𝘆 𝗔𝗜 𝗔𝗿𝗴𝘂𝗲𝗱 𝗪𝗶𝘁𝗵 𝗠𝗗𝗡 𝗔𝗻𝗱 𝗟𝗼𝘀𝘁
AI coding assistants change how you write software. They write tests and review pull requests. But they fail at one thing: knowing what browsers support today.
Mozilla tested this recently. They used Claude Code to ask about new Firefox features. The AI claimed Firefox did not support the Web Serial API. It even said Mozilla had no plans to add it.
The AI was wrong. Firefox 151 already had Web Serial support.
This happens because the web moves faster than AI training data. Browsers ship new APIs and CSS features every few weeks. An AI trained months ago will guess when it lacks current data. It sounds confident, but it is often wrong.
Mozilla fixed this with the MDN MCP Server.
MCP stands for Model Context Protocol. It lets AI models talk to external data sources. Instead of relying on memory, the AI checks live documentation.
The workflow changes from this: Developer → AI Memory → Answer
To this: Developer → AI → MDN → Answer
The AI now uses structured data from MDN and Browser Compatibility Data. It does not guess. It checks.
The results were impressive: • Compatibility info became accurate. • It identified new features correctly. • It made fewer wrong assumptions. • Responses were twice as fast.
Queries are faster because the AI no longer searches the web or parses messy HTML. It gets direct, machine-ready data.
You can use this with many tools: • Claude Code and Claude Desktop • Cursor • VS Code • Zed
To add the MDN MCP Server to Claude Code, run: claude mcp add --transport http mdn https://mcp.mdn.mozilla.net/
The lesson is simple. AI does not need bigger models. It needs better access to facts. The future of AI is not replacing documentation. It is AI reading the documentation with you.
Source: https://dev.to/josephciullo/the-day-ai-argued-with-mdn-and-lost-mm7
Optional learning community: https://t.me/GyaanSetuAi