Modernization, AI & Integration Insights
Real-world notes on legacy system modernization, AI enablement, agentic systems, and the unglamorous engineering that makes them useful.
-
GLM-5.2 High: Excellent Results on a Green-Field Project
One personal AI assistant, and a model that delivered excellent results. GLM-5.2 High exceeded expectations for green-field agentic work — and the vendor-agnostic architecture made it free to find out.
Read post → -
A fleet of agents that ships code to my repos
Five autonomous agents, one human reviewer, and the three design choices that make it actually work: vendor agnosticism, GitHub as the guardrail, and explicit limits on what each agent can do.
Read post →