AI is everywhere, but how much of it is real, and how much is just lipstick on a prompt?

In this episode of Security Done Different, Jav & Yak cut through the BS around AI in cybersecurity. There’s a difference between genuine innovation and a glorified wrapper around ChatGPT. And right now, the hype is drowning out the people actually building meaningful tech.

They talk about the evolution of AI at CloudGuard, from early automation and orchestration to where things are heading with agentic AI. Think less “magic black box”, more modular intelligence that mirrors how real analysts work: triage, enrich, remediate, repeat. We also dig into small vs large language models, the illusion of intelligence in GenAI, and why prompt engineering is both powerful and deeply limited.

Javid lays out a blunt truth: slapping “AI-powered” on your landing page doesn’t make it so. The real work is hard, expensive, and often invisible, but it’s where the real impact happens.

If you’re tired of AI vaporware and want to understand what good actually looks like, this one’s for you. Expect raw insights and real-world context.

Hype fades, capability doesn’t.