Teppo Felin on Blindness, Rationality, and Perception

Episode Summary

Executive Summary: Russ Roberts and Tepo Feline discuss Feline’s “fallacy of obviousness”: the idea that what seems obvious depends on the observer’s questions, theories, and goals, not on reality revealing itself automatically. Using the “invisible gorilla” experiment, they argue against the view that humans are simply blind or irrational and instead emphasize perception as active, theory-driven, and central to innovation, entrepreneurship, and sound economics.

Main Topics: The invisible gorilla and inattentional blindness (Priority: 5/5): They unpack the famous selective-attention video and question the standard interpretation that humans are broadly blind to the obvious, arguing the result is better understood as attention being guided by a task or question. The fallacy of obviousness (Priority: 5/5): Feline’s core claim is that obviousness is not an objective property of reality; it is created by the questions, problems, and search images in the observer’s mind. Critique of bounded rationality and behavioral economics (Priority: 5/5): The conversation challenges Simon and Kahneman’s emphasis on blindness and bias, arguing that these frameworks can overstate human error while underplaying purposeful perception and theoretical reasoning. Theory, data, and the limits of big data (Priority: 4/5): Roberts and Feline stress that data do not 'speak' on their own; interpretation requires theory, questions, and causal judgment, especially in an era of large datasets and machine learning. Science, omniscience, and policy humility (Priority: 4/5): They criticize the tendency of scientists and policymakers to act as if they can see the world more clearly than the people living in it, noting the risks of imposing outsider judgments on local choices. Innovation, entrepreneurship, and serendipity (Priority: 5/5): Examples like Steve Jobs at Xerox PARC, Airbnb, Uber, FedEx, and the calculator replacing the slide rule illustrate how entrepreneurs see value where others see noise because they possess different theories or search images. Search images, heuristics, and human meaning-making (Priority: 4/5): Drawing on Uexküll’s Umwelten and search images, they argue that organisms and people perceive selectively based on what they are looking for, which also explains creativity, art, and discovery.

Key Arguments: The gorilla experiment shows not that humans are universally blind, but that attention is directed by a task or question; people notice what they are prepared to notice. Obviousness is relational, not objective: something becomes obvious only relative to the observer’s goals, theories, and search image. Behavioral economics often treats scientists as omniscient observers while portraying decision-makers as biased or bounded, which can create an unwarranted asymmetry. Humans are not merely error-prone calculators; they are theory-bearing agents who interpret the world and act under uncertainty. Big data cannot replace theory because correlations are meaningless until framed by informed hypotheses, causal reasoning, and relevant questions. Innovation often looks irrational ex ante because value is hidden; entrepreneurs succeed by seeing latent uses and solving problems others have not recognized. Heuristics are best understood as specific search strategies tied to goals, not as generic rules detached from context. Policy and expert intervention should be humble about local knowledge and the possibility that outsiders misread what is “optimal” for others.

Data Points: Selective attention video views: About 19 million - Russ Roberts notes the viral reach of the invisible gorilla experiment video. Video length: 1 minute 22 seconds - Roberts recommends viewers watch the clip before discussing it. Miss rate in gorilla experiment: 20-30% to 70-80% - Feline says different experimental conditions yield broad ranges of participants missing the gorilla. Book length: 741 pages - Roberts mentions Solzhenitsyn’s In the First Circle as an upcoming reading. Episode date: July 19, 2018 - Introduced at the start of the podcast. Random detail in clip: 2 letters spray painted in the background - Used as an example of an obvious feature most viewers miss when focused on counting passes. Airbnb scale: Most rooms in the world - Roberts cites Airbnb as an example of a once-dismissed idea that became dominant. Venture funding reference: $2 million - Roberts recalls an arrangement in which Xerox could invest in Apple while technology was transferred.

Pivotal Quotes: "Humans can be blind to the obvious and we're sort of oblivious to this blindness, essentially." — Tepo Feline (summarizing Kahneman): Feline describes the standard interpretation of the invisible gorilla study. "Obviousness is a function of what's in our mind, the questions that we have, the problems that we have, and how we sort of attune to the world." — Tepo Feline: Central statement of the fallacy of obviousness thesis. "The data doesn't speak. The data is silent." — Russ Roberts: Roberts argues that theory and interpretation are required to make data meaningful.

Implications: Listeners should treat attention, data, and policy judgments as theory-laden and context-dependent. The episode encourages humility about expert omniscience and highlights why innovation often depends on asking the right questions, not just collecting more information.