Brockman: OpenAI enters AGI era with enhanced security measures

On September 15, 2026, Greg Brockman, co-founder of OpenAI, stated in an interview with a16z that humanity has entered the era of artificial general intelligence (AGI). AGI, defined as a system capable of performing complex cognitive tasks across various domains and surpassing human performance in most economically valuable work, is now becoming a reality. Brockman noted that models like GPT-6 Astra have achieved long-term autonomous execution and computer operation capabilities, marking the threshold for an AGI era. Brockman revealed that OpenAI has reallocated 25% of its production engineers to focus on system defense and security architecture upgrades. The team used Astra to identify and fix multiple critical vulnerabilities. While new issues were initially discovered, the process has now reached saturation, with Astra identifying most critical vulnerabilities. Brockman emphasized that this does not mean the system is completely secure, and new rounds of audits will be conducted as new models emerge, with enhanced cybersecurity measures deployed to uncover previously undetected flaws.