Researchers used Claude to hack OpenAI

Has anyone else noticed how much Claude and OpenAI hate each other. Try running local models On the same machine and look at how they lock each other down and trip over each other’s trap. Could be explained by chaos, but I remain suspicious