
James explains why coding agents are already taking over parts of his workflow, how vector databases became a core building block for modern AI systems, and why retrieval still matters even in a world obsessed with bigger models. They also get into the real mechanics behind RAG, hallucinations, MCP, long term memory for agents, and the challenges of building production grade AI systems that can search, reason, and scale reliably. If you want a practical conversation about where agent infrastructure is going and what engineers should actually pay attention to, this episode is worth watching. ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ ⏰ TIMESTAMPS: 00:00 – AI is already taking parts of engineering work 01:03 – James Luan’s background and first AI moments 07:07 – Why Zilliz was built and how vector databases fit in 16:58 – Long term memory, agent search, and reasoning workflows 21:37 – MCP, tooling limits, and real world production issues 31:02 – Are coding agents already replacing parts of engineering? 35:52 – AI for travel planning, presentations, and parallel work 38:57 – NotebookLM, Gamma, and James’s favorite AI tools 39:45 – Where to find James and Zilliz ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ Sign up for free ➡️ https://www.jotform.com/ ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ Follow us on: Twitter ➡️ https://x.com/aiagentspodcast Instagram ➡️ https://www.instagram.com/aiagentspodcast TikTok ➡️ https://www.tiktok.com/@aiagentspodcast ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬