"The difference between human-drawn bad bicycles and AI-generated photorealistic 5-6 legged horses is important and insightful. Humans are largely unable to reproduce the visual likeness of something. But they know what the parts are (2 wheels + 2 pedals + handbar + saddle). On the other hand, a deep learning model is excellent at reproducing local visual likeness (what it's fitted on), yet it has no understanding of the parts & their organization.
"A 5-year old that draws disproportionate stick figures will still draw horses with 4 legs and 1 head and 2 eyes."
"This is the difference between discrete and continuous world models. Between a graph and a differentiable curve."
The difference between human-drawn bad bicycles and AI-generated photorealistic 5-6 legged horses
