Origin Intelligence Experiments
A running log of autonomous AI agent experiments against real production systems. Each post documents what the pipeline did, what it found, what it got wrong, and what I learned watching it work on its own.
-
The Matt Pocock Skills — A Field Guide
A field guide to Matt Pocock's six skills for working with AI coding agents. Plan, build, fix — three loops, each handed off through real artifacts (issues, PRDs, post-mortems), with one substrate underneath that ties them together.
-
Tres ejes y dos brechas: lo que falta en la conversación sobre la IA
Three years working with Claude. A personal reflection on why most AI skepticism comes from judging the tool under the wrong conditions — three axes (model, technique, knowledge) and two real gaps (economic, knowledge) that bias the conversation.
-
Multiframework Recipes: React dentro de Salesforce, explicado estilo Feynman
Salesforce just opened the door to React inside the org. A Spanish-language walkthrough of the new Multiframework Recipes — what they are, how the bundle pipeline works, and where they unlock real value across Data Cloud, B2C Commerce, Service Cloud, Agentforce and SFMC.
-
When your website optimizes itself every morning
An autonomous AI agent that reads my real analytics every morning at 6am, derives the site's audience and positioning from the data itself, and opens a pull request with changes — and on day one caught a bug I had missed.