Fixed with subagents.

I swear something is dead every week for yall.

MCP is based on a lie: Machines are good at read/generate machine-parsable procotols.

Turns LLMs are shit with JSON. Especially those JSON str embeded inside another JSON key-value pairs.

Why do smart ppl design a schema like escape JSON into str embeded into another?

It's based on another lie: AIs favor static typed languages.

mcp will consolidate, its all stdio fragile and stateless

The vernacular around prompts, text, and docs, is quite amazing. Marketing really is value creation.

The idea that MCP tool definitions take up a certain number of tokens is laughable. That's an implementation detail of the agent harness. MCP is just an API specification. Hell, there's nothing in it that makes it much of any different than OpenAPI, except that its a bit more local-dev focused. There's a thousand things harnesses can and do do to optimize MCP beyond just "spit out the raw MCP output into the context window and pray".

MCP is just one of many -insecure- protocols that will be swallowed by a runtime governance protocol (like g8e) that is purpose-built for security, not to 'move fast and break stuff'.

MCP is what XML dreamed of becoming.

MCP will die for the same reason RAG died and why prompt engineering is dying. The models get better at understanding what you want and where to find the right tool or context to solve the problem on their own.

good to hear!