He read 522,607 real prompts sent to ChatGPT, Claude, Gemini and Grok
henkvaness · x · 2026-08-20
Journalist Henk van Ess analyzed 522,607 real questions users asked ChatGPT, Claude, Gemini, Grok and Copilot. His takeaway: the chatbot isn't the problem — the user is, as most people fail to use these tools effectively.
Related event: Analysis of 520K AI Prompts Puts Blame on Users, Not Chatbots(3 posts)→
More from AGI Musings
- Opinion: The End of Computer Science is Databases, Neural Networks, and Chips — _arohan_ · 2026-08-20
- AI Companies Can't Shirk Responsibility: Critique of 'Someone Else Will Handle It' Attitude — AaronBergman18 · 2026-08-20
- Automation's default trajectory pushes labor income toward zero, redistribution unavoidable — AaronBergman18 · 2026-08-20
- Are Harmful LLM Refusals Equivalent to an Orwellian State? — Hacubu · 2026-08-20
- Open-source AI singularity: usage doubles monthly, gap with closed-source shrinks — bindureddy · 2026-08-20
- Stripe partners with OpenRouter, declares 'the singularity' has begun — Charuru · 2026-08-20