Open-source tools, research notes, and practical experiments for people building with AI.https://modelfieldnotes.com/enhttps://modelfieldnotes.com/notes/comparing-ai-models/https://modelfieldnotes.com/notes/comparing-ai-models/Start with a real task and a clear definition of success. A useful comparison ends with a decision you can explain.Sun, 20 Sep 2026 00:00:00 GMThttps://modelfieldnotes.com/notes/cost-of-a-completed-task/https://modelfieldnotes.com/notes/cost-of-a-completed-task/Token prices are only the beginning. Count the tools, retries, and failed attempts behind a useful result.Sun, 20 Sep 2026 00:00:00 GMThttps://modelfieldnotes.com/notes/reading-ai-regulation/https://modelfieldnotes.com/notes/reading-ai-regulation/Start with jurisdiction, your role, and the intended use. Learn to separate binding rules, proposals, and voluntary guidance.Sun, 20 Sep 2026 00:00:00 GMThttps://modelfieldnotes.com/notes/retrying-a-successful-action/https://modelfieldnotes.com/notes/retrying-a-successful-action/A missing response does not mean an action failed. Follow one request all the way to two tickets—and the repair that prevents it.Sun, 20 Sep 2026 00:00:00 GMThttps://modelfieldnotes.com/research/five-foundations-for-ai-agents/https://modelfieldnotes.com/research/five-foundations-for-ai-agents/Reasoning and action, learned tool use, context, reflection, and reliability. Five papers that help explain the systems we build.Sun, 20 Sep 2026 00:00:00 GMT