LLMs vs. Agents: Why Brains Alone Aren’t Enough

Description

This hands-on workshop, led by Nanda Kumar, moves beyond the hype by first demonstrating the surprising limitations of raw LLMs—including hallucinations, math errors, and memory loss. The session then introduces the Agentic AI framework, showing how it solves these core issues by adding crucial memory, tools, and reasoning capabilities. Attendees will run live, side-by-side LLM vs. Agent comparisons in Google Colab, concluding with an interactive challenge focused on building usable, real-world AI systems.