Agents After Dark, by Prefactor
Conversations on AI agents in production
Agents After Dark, by Prefactor. Conversations with the people running AI agents and the Model Context Protocol in production, on evaluation, security, and what separates a demo from a live system.
-
Ajay Kumar on running MCP in production inside a regulated enterprise
Head of MCP Services, InfoTrack
InfoTrack sells authoritative legal and property data, and it built a dedicated Model Context Protocol business unit rather than treating MCP as a side experiment. I…
-
Aaron Vanston on building trustworthy AI when a mistake can be life and death
Co-Founder and CTO, Buildpass
Construction is an unforgiving place to put AI. The data is scattered across drawings, site diaries, forms, and worker credentials, and a wrong answer can mean a def…
-
Matthew Blode on shipping 90% AI-written code at 60 million users
Senior AI Engineer, Linktree
Matthew Blode writes more than 90 percent of his code with AI, using Claude Code, OpenAI Codex, and Cursor, yet writes less code than he ever has. He walks Matt Doug…
-
Andrew Bird on what an AI agent actually buys
Head of AI, Affinda
Affinda sells document processing, and Andrew Bird, its first Head of AI, spent 18 months rebuilding the product on top of large language models rather than bolting …