← heapsort-ai

internal feedback

1 items

RESEARCHarXiv CS.AI·4/20/2026

The World Leaks the Future: Harness Evolution for Future Prediction Agents

This research addresses the challenge of future prediction using LLM agents, where evidence evolves and useful supervision arrives only after an event is resolved. It introduces "internal feedback" derived from revisiting predictions over time and proposes "Milkyway", a self-evolving agent system that updates a persistent state to enhance prediction accuracy.

27