In this talk, we’ll explore the basics of AI-driven software agents: what they do, where they fit, and how they can occasionally create more headaches than they fix.

We’ll compare straightforward coding approaches and LLM-assisted options and discuss reliability, open-source implications, and the security angle (who would you trust to click and execute commands on your system?). By the end, you’ll have practical insights into deciding when an agent is the right call—and when it’s better to look elsewhere.