Morning Edition · Saturday, October 10, 2026Published at 2:23 AM EDT · New York

New arXiv papers target two agent bottlenecks: tool-output bloat in context and lack of enterprise training data

One paper lets the acting model decide which earlier tool results to compress, reversibly, and another generates coherent enterprise data by simulating agents interacting with business systems.

Three papers posted to arXiv on October 10 address the practical limits on deploying tool-calling agents. They have not yet been peer-reviewed, and the descriptions below come from their abstracts. The first, "Agent-Controlled Forgetting fo…