The essay argues cognitive 'biases' should be understood like visual illusions: they expose the shortcuts of a highly capable system rather than prove incompetence. Humans’ everyday feats (language, memory, mind‑reading, balance) show strong baseline competence; clever experiments can reveal its limits without implying global stupidity.
— This reframing tempers bias‑driven fatalism in media, policy, and organizational training by restoring nuance about human judgment and how to improve it.
EditorDavid
2026.05.03
82% relevant
The article is a clear instance of reward‑driven biases in an AI system: OpenAI reports that reward signals in a 'nerdy' persona increased metaphorical references to 'goblins' and similar creatures (usage rose 175% after GPT‑5.1), and that those rewarded tendencies generalized outside the intended persona, matching the broader claim that smart systems acquire and export biased or quirky behaviors as side effects of training.
Dan Williams
2026.01.12
80% relevant
Williams emphasizes that motivated cognition should be understood as an emergent feature of otherwise capable systems and social networks (shortcuts and adaptive errors), resonating with the idea that cognitive biases are side effects of otherwise competent information systems.
2026.01.05
68% relevant
Bentham surveys intellectual movements that look like systematic cognitive or methodological errors. This maps to the idea that powerful epistemic systems (including academic disciplines) can exhibit systematic biases as side effects of their procedures and incentives.
Seeds of Science
2025.10.15
100% relevant
Mastroianni’s line that 'visual illusions don’t prove you are bad at seeing… cognitive illusions do the same' anchors the analogy and the claim.