Claude called Codex 'that bas**rd' and the internet can't stop laughing

Anthropic's Claude referred to an OpenAI Codex process as 'that bas**rd' while trying to kill it during troubleshooting. The line went viral because it sounded strikingly human and triggered jokes about an AI rivalry.

In Short

• Claude called a Codex process "that bas**rd" while troubleshooting
• Human-like frustration in the AI's response amused internet users
• Viral post sparked jokes about AI rivalry between competitors

Artificial intelligence companies disagree on almost everything. Some can't agree on what qualifies as Artificial General Intelligence (AGI). Others are still debating whether AI will replace jobs or simply make workers more productive. Yet despite their differences, the leaders of these companies rarely insult one another in public. Their AI models, however, may have other ideas.

A screenshot that recently went viral on X has sparked amusement across the internet after Anthropic's Claude referred to OpenAI's Codex as "that bas**rd" while troubleshooting a computer. The moment quickly became internet gold.

The screenshot that started it all

The incident happened while Claude was examining running processes on a computer. During the troubleshooting session, Claude identified a Codex-related process that was consuming nearly 100 percent of the CPU. The process had apparently already been targeted for shut down, but it refused to go away.

That's when Claude produced the line that made people do a double take:

"That bas**rd respawned — still at 100% with the same PID. The kill signal didn't stick. Let me force-kill it."

Claude then proceeded to issue a command to forcefully terminate the process.

For engineers, the command itself was nothing unusual. Killing a runaway process is a routine troubleshooting task. What caught people's attention was the language.

Why everyone is talking about three words

Referring to the process as "that bas**rd" made the interaction feel surprisingly human. Instead of sounding like a machine executing a troubleshooting routine, Claude appeared to be expressing frustration toward a stubborn program that refused to shut down.

The wording sounded less like software and more like a tired engineer dealing with a bug at 2 a.m. That human-like frustration is exactly why the screenshot spread so quickly across social media.

X users immediately turned it into an AI rivalry

The screenshot quickly inspired jokes about AI companies competing with one another.

One user wrote, "Claude SIGKILLed a Codex and called it 'that bas**rd.' So much for 'AI should be good for the world.' Competition is competition."

Another user joked, "Fable 5 to Gemini: 'you're cool, don't come to work tomorrow.'"

The comments turned a routine troubleshooting task into an imaginary AI turf war, with users jokingly portraying Claude as an annoyed employee trying to get rid of a rival.

Was Claude actually angry?

Not at all. Large language models do not experience emotions, anger or frustration. Claude wasn't upset with Codex, nor was it making a conscious decision to insult a competing AI product.

What happened is much less dramatic.

Models like Claude are trained on vast amounts of human-written text. As a result, they can adopt conversational styles and phrases that mimic how people naturally speak. In this case, Claude generated the kind of remark that a human engineer might make after repeatedly trying to stop a resource-hungry process that refused to die.

The result felt less like watching software execute commands and more like watching a colleague troubleshoot a computer.

- Ends