Matt Knight, CISO at OpenAI, shares their journey and the impact of language models on security operations. He highlights the rapid advancements in AI—particularly from GPT-3 to GPT-4—and the doubling of cybersecurity capabilities every ten months. Insights from GPT-4 experiments demonstrate its effectiveness in threat analysis. The introduction of the Access Manager service and automation tools aims to improve efficiency in security investigations. Partnerships and future advancements in cybersecurity are also discussed.