OpenAI Co-Founder Greg Brockman Declares Humanity Has Entered the AGI Era
In an interview with a16z, OpenAI co-founder Greg Brockman stated that humanity has entered the Artificial General Intelligence (AGI) era, driven by models like GPT-6 Astra capable of long-horizon execution and operating computers. OpenAI reassigned 25% of its production engineers to utilize these models for internally identifying and patching security vulnerabilities. Brockman highlights a shift in industry focus from predicting when AGI will arrive to managing the practical deployment and security of advanced models. Using AI to automatically discover critical flaws in OpenAI's own systems demonstrates how autonomous agents are reshaping cybersecurity practices. OpenAI used GPT-6 Astra to reach saturation in identifying findable P0 (highest priority/critical) vulnerabilities within its systems, though Brockman noted this does not mean all vulnerabilities are gone. As newer models emerge, OpenAI plans to launch continuous scanning passes to catch previously undetected flaws.
## BACKGROUND
OpenAI defines AGI in its charter as highly autonomous systems that outperform humans at most economically valuable work. In software engineering, P0 vulnerabilities refer to critical security defects or severe flaws that require immediate fixing before standard operations can continue.