CyberWire Daily

Omer Akgul, PhD, researcher at RSA Conference studying LLM evaluation. He discusses measuring LLM consistency and using those metrics in cybersecurity. Short takes: what consistency means for model reliability. How automated metrics compare to human judgment. Practical calibration steps before deploying LLMs in critical workflows.

Ethan Cook, a writer and researcher at N2K who offers an outsider’s take on cybersecurity, reflects on the season’s conversations. He discusses fear and opportunity in hiring, why entry-level roles are misread, and the case for skills-based hiring. They debate trade versus profession pathways, diversity’s role in problem solving, and practical fixes like state programs and community college routes.

Ben Yelin, Program Director for Public Policy & External Affairs at UMD’s Center for Cyber Health and Hazard Strategies, explains administrative subpoenas and legal limits on surveillance. The conversation hits global zero‑day alerts, telecom intrusions, zero‑click LLM risks, ransomware outages, and troubling Chrome extension leaks. Short, sharp, and topical.

Simone Petrella, CEO of CyberVista and former DoD threat intelligence analyst, built a company to bridge the cybersecurity skills gap. She discusses career pivots from counterterrorism to cyber, hiring for analytical potential over tech pedigree, empowering teams to own work, and the mantra of faking it until you make it.

Piotr Wojtyla, Head of Threat Intelligence at Abnormal AI, explains InboxPrime AI, an AI-powered phishing kit. He covers how it uses Gmail, templates, and AI to craft polished, believable emails. Discussion includes its low one-time price, scalability features, and the shift toward autonomous, industrialized phishing tools.

Mike Carr, Field CTO at Xona, a specialist in event logistics and cybersecurity for large-scale events. He discusses securing temporary networks at the 2026 Winter Olympics. Shortcomings from vendor shortcuts and undocumented hotspots create gaps. He outlines what attackers often go after and why real-world constraints force risky compromises.

Tony Scott, former U.S. federal CIO and current CEO of Intrusion, shares his perspective as a seasoned technology and cybersecurity leader. He discusses rapid two-factor adoption after major breaches. He talks about continuous monitoring, modern infrastructure, and rebuilding institutional expertise. He raises concerns about staffing cuts and outsourcing in government cybersecurity.

Zach Edwards, senior threat researcher at Silent Push who tracks internet-scale scans and bulletproof hosting. He explains what bulletproof hosting is and how peers and ASNs let bad actors persist. He talks about using peering and sanctions to disrupt criminals. He also covers how defenders can spot and block malicious hosting.

Aaron Isaksen, AI research and engineering lead at Palo Alto Networks focused on applying AI to cybersecurity. He discusses why engineering excellence must come before ethical debates. He explores adversarial AI reshaping defenses. He warns about risks when human checks are removed and how AI changes code quality and lifecycle practices.

Kim Jones, season host and cybersecurity leader with decades in intelligence and security, reflects on the season's arc. She revisits talent challenges, hiring driven by fear and urgency, and why developing pipelines beats poaching. Conversations touch on diversity boosting critical thinking, the limits of certifications, and whether cybersecurity is a trade or a profession.