AI development

AI factories and AI-native engineering: architectures, agent development, and how production AI features actually get shipped and maintained.

AI features look easy in a demo and hard in production, where they need data, evaluation, cost control, and governance. This category covers both the foundations and the fast-moving news. On the foundations side, it explains what an AI factory is, how retrieval-augmented generation works for product teams, how to test LLM outputs, how to judge whether your company is ready, and what an AI governance policy should contain. On the news side, it follows the agent stack as it changes, from MCP servers and decision models to agent harnesses and launches like OpenAI's Dots, with vendor claims labeled as claims. If you are starting an AI initiative, begin with the readiness assessment and the evaluation guide. If you are building agents, start with the MCP explainer and the decision models guide. The AI development, AI factory, and AI agents pages explain how we take AI features to production.