Is my hypothesis crazy?

Prove me wrong! Change my mind!

🧠 I just claimed that AI “thinking” isn’t an illusion, it’s a manifestation of 60 years of human mind research I fundamentally disagree with Apple’s “The Illusion of Thinking” study. When AI models reduce effort on impossible tasks, they’re mirroring human cognitive disengagement patterns documented since the 1960s. Key insight: AI token usage follows the same “inverted-U” curve as human pupil dilation during cognitive overload. This suggests bounded rationality, not system failure.

🤔 Am I anthropomorphizing AI or onto something bigger? In my paper, I propose experiments that could completely refute these claims. What do you think? Especially if you think I’m wrong!

Here’s the preprint. You can also download it. Original source: https://doi.org/10.20944/preprints202506.1675.v1