Companies slash AI costs 65% by making Claude and Codex talk like cavemen

Companies Are Making Claude and Codex Talk Like Cavemen to Stop AI's Soaring Costs

A senior OpenAI employee has contributed code to the project, simply called 'caveman.' Companies are deliberately making their AI tools speak like cavemen in an attempt to stop burning through AI tokens and curb their massive expenditure on AI, 404 Media has found. The tool turns the usually verbose outpost of LLMs like Claude Code, Codex, or Gemini into a much more to the point answer. Think less "you're right to push back, I was wrong," and more "Hulk smash." Use of the caveman plugin is in direct response to the skyrocketing and unpredictable cost of AI. As 404 Media previously reported, companies are scrambling to stop spending so much on AI, with consulting giant Accenture finding much of the "soaring token spend" is thanks to people using AI to convert PDFs to presentations. People using caveman include developers at OpenAI, Nvidia, and GitHub, according to the tool's creator. A senior OpenAI employee has even contributed code to the project, adding support for OpenAI's Codex tool.

Companies Hope To Save On AI By Talking Like Cavemen

After training the likes of Claude to mimic intelligent human speech patterns, it's becoming more affordable to dumb these computers down The world of AI is ratcheting up costs all over the place. Consumer technology, computer parts and gaming are experiencing the hardware crunch, but if it makes you feel any better the AI firms aren't in the clear either. In recent weeks, many of the most popular AI firms have altered their subscription model from a flat rate to a per use one. It reflects the prohibitively expensive upkeep of these services and forcing many major companies who embraced AI to become more conservative. Some companies believe they've found a new solution to the rising costs: Make it talk like Tarzan. According to 404 Media, a new plugin is making the rounds at companies looking to reduce their token usage. In their report, staff at multiple companies, including OpenAI itself, have turned to 'Caveman,' which instructs the likes of Claude to talk less like SNL's Master Thespian and more like SNL's Caveman Lawyer. "I made Caveman back in early April because I was using Claude Code heavily and noticed a lot of my token spend was going to unnecessary prose," Caveman creator Julius Brussee tells 404, "pleasantries, hedging, transitions, and chatty language that does not really matter." With the plugin, responses that are often humanized, familiar and fishing for conversation by design are reduced to curt barks and grunts. Less friendly advice, more marching orders. Brussee attests that by reducing the IQ of AI, his token usage has reduced down some 65 percent. So much about AI feels like a metaphor wrecking havoc, but you can't help but laugh that after years of campaigning on the human-like intelligence of these chat bots, bean counting is forcing a frontal lobotomy. Much of the AI world and quest for investments has leaned heavily on pop cultural understandings of machine learning. They've gone to absurdist lengths to anthropomorphize language learning models. Now that the bottom line actually matters, things have rapidly changed in the space. Maintaining these AI services is prohibitively expensive, and none of the major firms have found a way to be naturally profitable. Earlier this year OpenAI shuttered Sora, their video generator bleeding a million a day, complicating an otherwise lucrative deal with Disney.