OpenAI's Astra Cybersecurity AI Model Nears Launch, Utilizes 'Recurrent Depth' Technique

OpenAI is expected to 'soon' launch Astra, its most advanced cybersecurity AI model, designed to compete with Anthropic's Mythos. The model is capable of autonomously finding and exploiting unknown cybersecurity flaws, with its most advanced features initially limited to a select group of testers due to the need for strong safeguards. Reports indicate that Astra uses a technique called 'recurrent depth,' which improves cost and performance but makes the AI's reasoning harder to monitor.