<?xml version="1.0" encoding="utf-8"?><feed xmlns="http://www.w3.org/2005/Atom"><title>Ju Lin's AI Weblog: Mcp</title><subtitle>An independent research notebook on AI engineering, agents, models and the systems around them.</subtitle><id>https://julin.ai/atom/tags/mcp/index.xml</id><link rel="self" type="application/atom+xml" href="https://julin.ai/atom/tags/mcp/index.xml"/><link rel="alternate" type="text/html" href="https://julin.ai/tags/mcp/"/><author><name>Ju Lin</name></author><updated>2026-08-27T00:00:00+12:00</updated><entry><title>Where MCP Is Headed Next</title><id>https://julin.ai/2026/08/27/mcp-roadmap/</id><link rel="alternate" type="text/html" href="https://julin.ai/2026/08/27/mcp-roadmap/"/><published>2026-08-27T00:00:00+12:00</published><updated>2026-08-27T00:00:00+12:00</updated><category term="field-notes"/><category term="mcp"/><category term="agents"/><content type="html">&lt;p&gt;&lt;a href="https://blog.modelcontextprotocol.io/posts/mcp-roadmap/"&gt;The MCP maintainers published an updated roadmap&lt;/a&gt;.&lt;/p&gt;
&lt;p&gt;Since March, MCP went stateless (no more session handshakes), turned Tasks into an official extension for long-running work, and shipped enterprise auth pieces like issuer validation and Client ID Metadata Documents.&lt;/p&gt;
&lt;p&gt;Next up: one HTTP-native transport instead of separate stdio and HTTP paths, real agent identity (workload federation, token exchange), and progressive tool discovery for large catalogs. No firm dates yet.&lt;/p&gt;
&lt;p&gt;MCP is leaving its single-session, request/response origins behind for something built for long-running, multi-agent systems.&lt;/p&gt;</content></entry></feed>