The Hallway Track
8 tracked signals on tooling.
MCP 2.0 drops stateful sessions, collapsing tool calls to a single HTTP request
“This is so much cleaner from both a client- and server-side implementation perspective. It's also a better fit for building scalable web applications, since now you don't need to maintain server-side state to keep track of those session IDs, or worry about routing the same session to the same backend machine.”
Browser agents fail due to poor infrastructure around models, not model intelligence; better page representation makes cheap models fast and reliable.
“models are pretty smart, but it's the infra around them that sucks”
MCP standardizes agent-tool connections like USB-C standardizes device charging
“MCP is a standard plug that any compliant model can use. You write the server once, you connect it to everything.”
smevals is a new open-source eval framework for comparing AI model and prompt configurations
“I've been trying to figure out an approach I like for evals for several years now. smevals is my third iteration on the idea and it feels right to me.”
llm-anthropic 0.25.1 adds Claude Opus 4.8 and fast mode support
Simon Willison released llm-typesafe 0.1a0, a plugin adding TypeSafe AI's Jev model to the LLM tool.
“I built this new plugin for LLM to add support for TypeSafe AI's new Jev model.”
llm-openrouter 0.7 adds Responses API support and three new server-side tools
uv 0.12.0 changes default project structure to src-layout with new build backend
“I've so far avoided using src layout in my own projects just out of inertia. I think it's time I switched.”