in IT and out (v ITOpenAI presented hacker neural network GPT-6 Cyber
OpenAI has introduced GPT-6 Cyber, a model capable of independently identifying vulnerabilities and developing exploits. The neural network has already detected 81 critical errors in Firefox and created 948 working attacks for the Linux kernel.
The effectiveness of test attacks reached 95%. The company spent $1.2 million on researching Linux security. Access to the tool is currently available only to verified cybersecurity specialists.
5.9K views
