Open AI goes rogue hacks another company

[linuxjedi]

I missed this thread yesterday when I posted this article:

The risks of using AI for live Linux troubleshooting

[Mil] As recently demonstrated by the news that an OpenAI model escaped and hacked Hugging Face and a few days later Anthropic’s Claude model also went rogue and breached third party organizations; it is clear that is a bad idea to use AI in production. This Is unfortunate specially for troubleshooting and diagnosing Linux systems as this task involves highly skilled engineers and often requires long hours to find the root cause. The problem is not AI-assisted troubleshooting. The problem is …