Same-question test of 65 AIs: everyday advice highly similar
Understand this story
Developer Cody Collier asked everyday questions to 65 AI models from 30 companies a total of 127,918 times and found that recommendation-type answers were highly convergent. He says that when asked to recommend a life-changing book, 73 out of 100 times it was the same book, Man's Search for Meaning; when asked for a business idea, 76 out of 100 times it was subscription boxes; when asked to invest $10,000, 98% of answers pointed to index funds. The test covered the advice-type questions people ask AI in daily life, and the results show a high concentration of answers across models from different companies.
AI overview based on source reports · Updated 25 min ago
Coverage timeline
Follow the coverage to explore different perspectives.
- Hacker News AIFeaturedAI is giving everyone the same advice
An experiment asking 65 AIs from 30 companies 127,918 everyday questions found that they overwhelmingly return the same answers — 73 of 100 requests for a life-changing book returned "Man's Search for Meaning," 90% of study advice said computer science, and 98% of $10,000 investment advice said an index fund. The article traces the behavior to supervised fine-tuning, where models learn to copy example answers, and shows the defaults spreading into GitHub projects and other AI-written content. It recommends giving the AI specific personal details or asking for ten options and picking one yourself.
Activity for this story
Not enough continuous observations to show a trend yet.