Anthropic's Claude Abused Codex For Messing Up A Coding Task

Anthropic's Claude has got all the attention of the world for one reason or another. And now, the AI model is again in talks because of a viral screenshot where Claude used a casual phrase to describe a process that will simply not stop running when reviewing a coding task from Codex. The response of the same was posted on X and went viral in a few hours. A majority of users also said that, finally, AI is losing patience. In the shared screenshot, Claude can be seen inspecting running processes on a computer. During a troubleshooting session, Claude noticed a Codex-related process eating 100 per cent of the CPU. The process had already been targeted for termination, but it still kept running. Claude then responded in an unusual way, saying,

'That bastard respawned - still 100% with the same PID. The kill signal didn't stick. Let force kill it.'Amazon's AWS Is Openly Taking Jibes At AI Coding Hype, Here's What Happened

Right after that, Claude issued a command to forcefully terminate the process. For engineers, this is a normal debugging process; killing a stuck or runaway process is routine. However, the way Claude responded to the same was the main tea. For your information, AI models like Claude are trained on vast amounts of text written by people. On the response, experts said that Claude did not say what it said because it was annoyed. Instead, the model did that because it learned patterns from human language. At the time of training, AI models absorb common ways of people's communication, including casual, emotional, and heated expressions as well. The phrase came from those learned language patterns. It was not a sign of frustration, but it was simply a pattern match. In real-life scenarios, engineers tend to use strong language when a stubborn process refuses to close, and Claude picked up on that.