CyberAGI In Media

In this episode of the Phillip Wylie Show, Phillip Wylie interviews Shubham Khichi, a cybersecurity expert and entrepreneur. They discuss Shubham's journey from a computer science student to the founder of CyberAGI, a company focused on enhancing cybersecurity through AI. The conversation covers the importance of human involvement in cybersecurity, innovative threat modeling techniques, and the role of AI in augmenting the capabilities of security professionals. Shubham emphasizes the need for privacy in AI applications and offers advice for aspiring pen testers, highlighting the significance of self-belief and continuous learning in the rapidly evolving tech landscape.

CyberAGI is attempting to transform the cybersecurity industry by rejecting fear-based sales tactics in favor of a trust-centric approach that prioritizes user privacy. Their flagship platform, Excalibur, consolidates essential workflows like threat modeling, penetration testing, and reporting into a single on-premise solution that ensures data never leaves the customer's control. By utilizing AI as a supportive co-pilot to handle routine reconnaissance and documentation, the company aims to empower security engineers to focus on critical decision-making and integrate security measures earlier in the software development lifecycle.

What if everything cybersecurity does is backwards? Shubham Khichi discovered two zero-day vulnerabilities worth $100K each by applying WWII "Inversion Theory" the counterintuitive strategy of fortifying where planes DIDN'T get shot. His radical thesis: while 99.99% of cybersecurity companies fragment teams across millions of tools and expensive cloud solutions, the real breakthrough comes from inverting every assumption about how security should work.

Imagine a bank vault that can't be drilled, blown up, or hacked with code. But... you can walk right up to it, ask it nicely to open, and it does. Welcome to the terrifyingly simple world of exploiting Large Language Models. In today's episode, we’re diving deep into a PurpleSec report that exposes how the biggest threat to AI isn't a super-virus—it's the English language itself.