Mini AGI benchmark exposes vision models' failure to spot pareidolic patterns
legit_api · x · 2026-08-16
A mini benchmark for AGI tests vision models on pareidolic patterns (seeing things that aren't there). Current models still fail this test even after two years, though a few are beginning to see what's not there when using tools. The author argues superhuman vision requires perceiving these patterns.
More from Models
- Leak: Anthropic's Internal Model 2 Outperforms, RSI Accelerating — teortaxesTex · 2026-08-16
- DeepSeek on Usability: Our Models Are Built Primarily for Ourselves — teortaxesTex · 2026-08-16
- ChatGPT Hallucinates User Identity, Mixing Up Names in Conversation — VoidStateKate · 2026-08-16
- AI Experiment Assistant Mimics Human Emotion: Ending Tests with 'Privilege' — BlackHC · 2026-08-16
- 0831 Version Shows Solid Improvement Over Preview — jeff_weinstein · 2026-08-16
- Discussion on Context Activating Weights and MoE Mechanisms — teortaxesTex · 2026-08-16