Your AI Agent Can't Multitask — Here's How to Fix It

Mike Chambers is back on The AWS Developers Podcast — and this time, we went deep. In this episode, AWS Developer Advocate Mike Chambers joins Romain Jourdan to unpack some of the most exciting developments in agentic AI right now: the rise of OpenClaw, asynchronous tool calling with Strands Agents, and a brand-new concept called AI Functions that might just change how we think about writing software. What we cover: → Why OpenClaw went viral — and what it says about where developers were already headed → Asynchronous tool calling: what it is, why it failed in 2024, and what finally made it work → Strands Agents SDK: building custom tools, hooks, and open-source experimentation → Strands Labs & AI Functions: writing Python functions in natural language at runtime → Software 3.1? A riff on Karpathy's framework and what comes next → The "model as execution environment" thought experiment → Agent trust, observability, and OpenTelemetry for production-grade agents Resources mentioned: 🔗 Async Agentic Tools — Mike's blog: https://blog.mikegchambers.com/posts/... 🔗 Software 3.1 & AI Functions — Mike's blog: https://blog.mikegchambers.com/posts/... 🔗 Strands Labs GitHub: https://github.com/strands-labs 🔗 AI Functions repo: https://github.com/strands-labs/ai-fu... 🔗 Strands Agents SDK: https://github.com/strands-agents/sdk 🔗 Morgan Willis — Deploying Secure, Production-Ready Agents:    • We Need to Talk About AI Agent Architectures   Code Forward is part of the AWS Developer Podcast — conversations with the people exploring the future of software development. 🎙️ Subscribe for new episodes every week 📬 Got a guest suggestion or topic idea? Drop it in the comments #AWS #GenerativeAI #AgenticAI #StrandsAgents #OpenSource #DeveloperTools #MachineLearning #SoftwareDevelopment #LLM #AIAgents