OpenAI has released a model built to find security vulnerabilities, deliberately loosened the restrictions that would normally stop it writing exploit code, and restricted who can use it. The results ...

In late May, thousands of developers, system administrators, engineering managers, and open-source advocates came together ...