In today’s issue:

- OpenHands CodeAct 2.1 achieves state-of-the-art on SWE-Bench Verified.

- How companies like Perplexity and Superhuman are building advanced AI agents.

- Summary of the memory types needed to build effective AI agents.

- CrewAI closes $18M funding to accelerate their growth efforts.

- Automatic agentic workflow generation, personalized web agents, LLM-based agent benchmarks, and much more.