In partnership with
Built by a former cannabis regulator, Policy, Decoded helps operators read the policy terrain before it shifts beneath their feet.
Today's edition is sponsored by Mintlify, the AI-native documentation platform developer companies use to turn a codebase into clean, searchable docs, with interactive API references, an in-doc AI assistant, and MCP endpoints that let AI coding tools pull straight from the source. Teams at companies like Anthropic, Perplexity, and Cursor run their docs on it. Their support keeps these briefings free and focused on what actually matters.
Two scheduling notes before we start. There is no Sunday Read-In this week, so the next editorial lands the Sunday after. And next week we're at CANNRA in Oklahoma City. If you're going, find us there, and bear with us if a few editions arrive later in the day than usual. The week after that I'm in Chicago June 14 to 16 for IgniteIt's Cannabis Capital Conference; Policy, Decoded readers take 20 percent off registration with code POLICYDECODED20.
In a formal statement on the fiscal 2027 Agriculture bill, the White House asked Congress to protect full-spectrum CBD while keeping November's restriction on intoxicating hemp, which is the administration endorsing the very split the industry keeps resisting. New York's Legislature passed the first law built to penalize inversion and protect the operators who test, label, and pay their way. Michigan's 24 percent wholesale tax, sold as road money, has brought in barely a third of what the state projected. And James Lankford and Jodey Arrington want Treasury to keep the 280E door shut, retroactive years included. Someone is deciding who stays legal and what it costs, and this week most of them work in Washington.
🏛️ The White House backs full-spectrum CBD
🗽 New York arms the OCM
💰 Republicans guard the 280E door
The beginning of wisdom is to call things by their right names.
AI agents now read your docs almost as much as humans do.
Mintlify analyzed 790 million requests across its documentation platform. The finding: AI coding agents account for 45.3% of all traffic, nearly tied with traditional browsers at 45.8%.
Two tools are driving almost all of it:
Claude Code: 25.2% of total traffic, more requests than Chrome on Windows
Cursor: 18% of total traffic
Together they account for 95.6% of all identified AI agent traffic
The rest of the field, OpenCode, Trae, ChatGPT, and NotebookLM, is showing up but nowhere close.
One caveat: OpenAI's Codex doesn't send an identifiable user-agent header, so the real agent percentage is likely even higher.
The takeaway for anyone maintaining developer docs: your documentation now serves two audiences. Structure and machine-readability matter as much as clarity for human readers.