KSI pushes the Pareto frontier of self-improvement agents 🚀
A simple thought experiment: What if we stop trying to cram intelligence into agents, and instead externalize it into a shared knowledge base?View quoted post on X
Introducing Knowledge-centric Self-Improvement (KSI).
TL;DR: We show that a remarkably simple knowledge curation mechanism pushes the Pareto frontier of self-improvement agents, outperforming SOTA agent-centric approaches (e.g., harness or prompt optimization) across reasoning, coding, and terminal benchmarks at a fraction of the cost.👇