404 Media⁠48%

Companies Are Making Claude and Codex Talk Like Cavemen to Stop AI’s Soaring Costs ⁠84%

By Joseph Cox⁠67%

6/30/2026, 1:33:16 PM

BS Summary: This article contains 14 faulty reasoning types, including Confirmation Bias, Hasty Generalization, and Post Hoc (False Cause), with Appeal to Authority as the most egregious example at 59.3% saturation with 99 hits. Analysis detected 514 faulty-reasoning hits from 167 analyzed words, generating a BS Score of 75% and a BS Rank of ⁠84% (3,694 of 21,886 articles). This article is worse (more manipulative) than 83.10% of the article peer group.

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. 
Article reasoning-pattern comparisonThis article: 50.3%Joseph Cox: 2.3%404 Media: 3.5%Confirmation Bias50.3%This article: 0.0%Joseph Cox: 0.9%404 Media: 0.9%Anchoring Bias0.0%This article: 10.2%Joseph Cox: 5.7%404 Media: 3.8%Availability Heuristic10.2%This article: 13.8%Joseph Cox: 0.8%404 Media: 1.6%Representativeness Heuristic13.8%This article: 0.0%Joseph Cox: 0.4%404 Media: 0.5%Hindsight Bias0.0%This article: 0.0%Joseph Cox: 1.6%404 Media: 2.8%Overconfidence Bias0.0%This article: 16.8%Joseph Cox: 13.2%404 Media: 5.6%Framing Effect16.8%This article: 0.0%Joseph Cox: 1.0%404 Media: 0.4%Loss Aversion0.0%This article: 0.0%Joseph Cox: 0.5%404 Media: 0.6%Status Quo Bias0.0%This article: 0.0%Joseph Cox: 0.1%404 Media: 0.2%Sunk Cost Effect0.0%This article: 0.0%Joseph Cox: 1.0%404 Media: 2.8%Optimism Bias0.0%This article: 0.0%Joseph Cox: 3.5%404 Media: 1.6%Pessimism Bias0.0%This article: 18.0%Joseph Cox: 20.0%404 Media: 11.5%Negativity Bias18.0%This article: 0.0%Joseph Cox: 1.0%404 Media: 0.9%Self-Serving Bias0.0%This article: 0.0%Joseph Cox: 0.4%404 Media: 0.7%Fundamental Attribution Error0.0%This article: 0.0%Joseph Cox: 0.0%404 Media: 0.1%Actor-Observer Bias0.0%This article: 0.0%Joseph Cox: 0.8%404 Media: 0.8%In-Group Bias0.0%This article: 0.0%Joseph Cox: 0.0%404 Media: 0.3%Out-Group Homogeneity Bias0.0%This article: 24.0%Joseph Cox: 0.9%404 Media: 1.0%Halo Effect24.0%This article: 0.0%Joseph Cox: 0.0%404 Media: 0.2%Horn Effect0.0%This article: 0.0%Joseph Cox: 0.0%404 Media: 0.0%Dunning-Kruger Effect0.0%This article: 0.0%Joseph Cox: 1.7%404 Media: 1.0%Recency Bias0.0%This article: 0.0%Joseph Cox: 0.7%404 Media: 0.3%Primacy Effect0.0%This article: 0.0%Joseph Cox: 0.2%404 Media: 0.0%Blind-Spot Bias0.0%This article: 0.0%Joseph Cox: 0.5%404 Media: 0.3%Ad Hominem0.0%This article: 0.0%Joseph Cox: 0.1%404 Media: 0.3%Straw Man0.0%This article: 59.3%Joseph Cox: 4.0%404 Media: 3.3%Appeal to Authority59.3%This article: 16.8%Joseph Cox: 1.4%404 Media: 1.3%False Dilemma16.8%This article: 0.0%Joseph Cox: 0.9%404 Media: 1.2%Slippery Slope0.0%This article: 0.0%Joseph Cox: 0.0%404 Media: 0.2%Circular Reasoning0.0%This article: 35.9%Joseph Cox: 9.5%404 Media: 7.8%Hasty Generalization35.9%This article: 0.0%Joseph Cox: 0.7%404 Media: 0.3%Red Herring0.0%This article: 9.0%Joseph Cox: 1.1%404 Media: 0.7%Bandwagon9.0%This article: 8.4%Joseph Cox: 6.3%404 Media: 4.8%Appeal to Emotion8.4%This article: 0.0%Joseph Cox: 0.2%404 Media: 0.7%Begging the Question0.0%This article: 28.1%Joseph Cox: 1.8%404 Media: 2.0%Post Hoc (False Cause)28.1%This article: 0.0%Joseph Cox: 0.0%404 Media: 0.1%Tu Quoque0.0%This article: 0.0%Joseph Cox: 1.7%404 Media: 0.5%Burden of Proof0.0%This article: 0.0%Joseph Cox: 0.3%404 Media: 0.4%Appeal to Nature0.0%This article: 0.0%Joseph Cox: 0.0%404 Media: 0.2%Composition/Division0.0%This article: 0.0%Joseph Cox: 2.2%404 Media: 2.9%Anecdotal0.0%This article: 0.0%Joseph Cox: 0.0%404 Media: 0.1%No True Scotsman0.0%This article: 0.0%Joseph Cox: 4.6%404 Media: 2.1%Ambiguity (Equivocation)0.0%This article: 0.0%Joseph Cox: 0.0%404 Media: 0.0%Gambler’s Fallacy0.0%This article: 0.0%Joseph Cox: 0.0%404 Media: 0.1%Middle Ground0.0%This article: 0.0%Joseph Cox: 0.0%404 Media: 0.1%Personal Incredulity0.0%This article: 0.0%Joseph Cox: 0.0%404 Media: 0.1%Special Pleading0.0%This article: 0.0%Joseph Cox: 0.0%404 Media: 0.1%Genetic Fallacy0.0%This article: 9.0%Joseph Cox: 4.8%404 Media: 2.6%Unattributed Quote9.0%This article: 0.0%Joseph Cox: 2.7%404 Media: 1.6%Quote-first Misdirection0.0%This article: 8.4%Joseph Cox: 9.8%404 Media: 11.5%Biased Writer Voice8.4%This article: 0.0%Joseph Cox: 2.2%404 Media: 1.5%Indoctrination0.0%This article: 0.0%Joseph Cox: 2.5%404 Media: 0.6%Politically Left Leaning Bias0.0%This article: 0.0%Joseph Cox: 0.2%404 Media: 0.0%Politically Right Leaning Bias0.0%This article: 0.0%Joseph Cox: 7.7%404 Media: 2.9%Attempt to Sell a Product or S…0.0%

167 words analyzed.

Speakers

2speakers50%attributed speech83writer words
Voice mapSelect a segment to jump to its words
Writer's voice • 14 words • 0.0% coverage404 Media • 30 words • 0.0% coverageWriter's voice • 23 words • 0.0% coverageWriter's voice • 14 words • 100.0% coverageWriter's voice • 17 words • 0.0% coverage404 Media • 37 words • 0.0% coverageWriter's voice • 15 words • 100.0% coverageOpenAI • 17 words • 0.0% coverage
Selected voice

404 Media

100%flagged-word coverage
67 attributed words80% of attributed speech100% writer coverage
0%10.0%20.0%Unattributed Quote-18.1 ptsWriter: 18.1%404 Media: 0.0%0.0%Biased Writer Voice-16.9 ptsWriter: 16.9%404 Media: 0.0%0.0%

Attribution is sentence-level. Pattern percentages are calculated only from words assigned to that voice.

Loading…
Loading…
Loading…
Loading…

Analysis

Hover over highlighted words in the article to view the associated bias or fallacy analysis.