<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Model Fieldnotes — writing by Swapnil Upganlawar</title><description>Signed notes, research collections, and simulation-based guides. Automated briefings are available in the combined feed.</description><link>https://modelfieldnotes.com/</link><language>en</language><item><title>Retrying without duplicate actions</title><link>https://modelfieldnotes.com/fieldbook/duplicate-action/</link><guid isPermaLink="true">https://modelfieldnotes.com/fieldbook/duplicate-action/</guid><description>A lost response leaves an uncertain caller and a completed action. Follow the retry, inspect the duplicate, and see what an idempotency contract changes.</description><pubDate>Mon, 21 Sep 2026 00:00:00 GMT</pubDate></item><item><title>Keeping constraints through compaction</title><link>https://modelfieldnotes.com/fieldbook/forgotten-instruction/</link><guid isPermaLink="true">https://modelfieldnotes.com/fieldbook/forgotten-instruction/</guid><description>Follow a draft-only instruction through a shortened conversation. See why retained constraints help, and why permissions must still enforce the boundary.</description><pubDate>Mon, 21 Sep 2026 00:00:00 GMT</pubDate></item><item><title>Verifying completion before saying “done”</title><link>https://modelfieldnotes.com/fieldbook/premature-done/</link><guid isPermaLink="true">https://modelfieldnotes.com/fieldbook/premature-done/</guid><description>A tool can accept a request before the result exists. Separate acceptance, pending work, and completion before telling the user the task is finished.</description><pubDate>Mon, 21 Sep 2026 00:00:00 GMT</pubDate></item><item><title>When faster AI work meets the organization</title><link>https://modelfieldnotes.com/notes/ai-productivity-organizational-performance/</link><guid isPermaLink="true">https://modelfieldnotes.com/notes/ai-productivity-organizational-performance/</guid><description>How to examine handoffs, architecture, review capacity, and incentives before turning task-level AI gains into claims about business performance.</description><pubDate>Mon, 21 Sep 2026 00:00:00 GMT</pubDate></item><item><title>What changes when managers lead teams using AI?</title><link>https://modelfieldnotes.com/notes/managing-teams-with-generative-ai/</link><guid isPermaLink="true">https://modelfieldnotes.com/notes/managing-teams-with-generative-ai/</guid><description>A practical look at delegation, coaching, trust, and accountability as generative AI changes how a team gets work done.</description><pubDate>Mon, 21 Sep 2026 00:00:00 GMT</pubDate></item><item><title>What changes when AI can do the work?</title><link>https://modelfieldnotes.com/notes/what-changes-when-ai-can-do-the-work/</link><guid isPermaLink="true">https://modelfieldnotes.com/notes/what-changes-when-ai-can-do-the-work/</guid><description>My perspective on the shift from generating answers to delegating work—and what it means for products, engineering teams, and business decisions.</description><pubDate>Mon, 21 Sep 2026 00:00:00 GMT</pubDate></item><item><title>How to compare AI models for your application</title><link>https://modelfieldnotes.com/notes/comparing-ai-models/</link><guid isPermaLink="true">https://modelfieldnotes.com/notes/comparing-ai-models/</guid><description>Start with a real task and a clear definition of success. A useful comparison ends with a decision you can explain.</description><pubDate>Sun, 20 Sep 2026 00:00:00 GMT</pubDate></item><item><title>The cost of a completed task</title><link>https://modelfieldnotes.com/notes/cost-of-a-completed-task/</link><guid isPermaLink="true">https://modelfieldnotes.com/notes/cost-of-a-completed-task/</guid><description>Token prices are only the beginning. Count the tools, retries, and failed attempts behind a useful result.</description><pubDate>Sun, 20 Sep 2026 00:00:00 GMT</pubDate></item><item><title>Reading AI regulation as a builder</title><link>https://modelfieldnotes.com/notes/reading-ai-regulation/</link><guid isPermaLink="true">https://modelfieldnotes.com/notes/reading-ai-regulation/</guid><description>Start with jurisdiction, your role, and the intended use. Learn to separate binding rules, proposals, and voluntary guidance.</description><pubDate>Sun, 20 Sep 2026 00:00:00 GMT</pubDate></item><item><title>When an agent retries a successful action</title><link>https://modelfieldnotes.com/notes/retrying-a-successful-action/</link><guid isPermaLink="true">https://modelfieldnotes.com/notes/retrying-a-successful-action/</guid><description>A missing response does not mean an action failed. Follow one request all the way to two tickets—and the repair that prevents it.</description><pubDate>Sun, 20 Sep 2026 00:00:00 GMT</pubDate></item><item><title>Five foundations for understanding AI agents</title><link>https://modelfieldnotes.com/research/five-foundations-for-ai-agents/</link><guid isPermaLink="true">https://modelfieldnotes.com/research/five-foundations-for-ai-agents/</guid><description>Reasoning and action, learned tool use, context, reflection, and reliability. Five papers that help explain the systems we build.</description><pubDate>Sun, 20 Sep 2026 00:00:00 GMT</pubDate></item></channel></rss>