SELF-IMPROVEMENT

A new August 2026 paper argues that coding agents improve faster when they compare attempts across tasks and lineages, not just retry one failed trajectory.

Agents that critique their own output, learn from mistakes, and get better over time - the three patterns that actually ship, from simple reflection loops to tree search and meta agents.

Lilian Weng argues self-improving AI won't start with models rewriting their weights - it starts with the harness. Here's what that means for developers building agents.

AI agents that reflect on failures, accumulate skills, and get better with every session. Reflection patterns, memory architectures, skill extraction, and working code examples for building agents that actually learn.