A preprint finds that when AI models repeatedly tidy up expert-written agent instructions with no testing, the instructions get worse. The first knowledge to go is the part the model could not have written itself.
A preprint finds that agents reusing their own earlier tool calls do worse than newcomers after a silent interface change. They then report the broken task as done.
Researchers let a coding model rewrite whole memory systems for chat agents. The winners beat published designs on two benchmarks, and one evolved component improved five existing systems when bolted on.
A preprint from Peking University, JD.com and others trains 4B and 9B models to save their progress in ordinary files during long tasks. Teaching memory through familiar file commands worked better than teaching purpose-built memory tools.
A new benchmark built from real researcher sessions finds the best of seven models fully completes only two-thirds of tasks. Scores fall sharply once earlier decisions and existing files are checked.
A preprint proposes that AI agents should treat each lesson learned from a task as a hypothesis to test, and keep it only if it helps without breaking anything. The benchmark gains are large, but the comparisons need care and no code is released.