Meta AI pairs action agents with dedicated memory agents
DeepLearning.AI reports that Meta AI is pairing primary action agents with dedicated memory agents to address context rot, in which long contexts cause AI agents to forget early mistakes.
The memory agents keep structured notes on tools and past errors, then inject short reminders only at crucial moments. DeepLearning.AI says the approach raised Claude Sonnet 4.5 benchmarks from 37.6% to 45.9%. Its technical analysis describes the method.