Finding Security Vulnerabilities using AI with Sami Laiho - RunAs Radio

Topics in this episode

Large language models Cybersecurity Artificial intelligence Penetration testing security

Episode notes

How are large language models changing the way security vulnerabilities are found? Richard chats with Sami Laiho about the rapidly changing landscape in security exploits. Certain LLM models like Anthropic's Mythos and Microsoft MDASH are optimized to find software vulnerabilities - and potentially fix them. And so there is an arms race of sorts, repairing old vulnerabilities before LLMs in the hands of black hats can exploit them. But what about everyone else? Sami talks about getting LLMs working for your organization to test for potential security risks and assess the impact of new vulnerabilities as they appear. Links Palo Alto Networks Claude Mythos Microsoft MDASH Zero Day Clock Security Update Guide Recorded June 19, 2026