Junk Food of Cognitive Labor

In the AI Story

Howard Moskowitz's optimization work in the 1980s and 1990s — documented in Moss's reporting — established that the food industry was not selling nutrition. It was selling reward signals. The consumer's body was a reward-detection system calibrated for naturally occurring food, and the industry's competitive advantage lay in producing stimuli that exceeded the system's calibration range, ensuring consumers would choose processed products over natural ones through the brute neurochemical logic of a stronger signal winning the competition for motivational control.

The technology industry has arrived at a structurally identical combination through a different optimization process. The tools were not deliberately designed to produce productive addiction the way processed food was designed to produce consumption addiction. They were designed for capability — to produce the most useful output for the user. But capability, pursued without regard to the reward-system implications, converges on the same supernormal combination: instant feedback, complete execution, continuous progress. The three features that make AI tools most useful are the three features that make them supernormal.

The cultural response to the parallel is where the analysis acquires its uncomfortable edge. When the food industry's engineering was exposed, a cultural vocabulary emerged — overeating, binge eating, food addiction, compulsive consumption. These categories did not exist when the industry began optimizing reward signals in the 1960s. They were created because the phenomenon they described was new, and they enabled the clinical and regulatory infrastructure that eventually produced the effective interventions: labeling requirements, trans-fat bans, school lunch reforms, portion regulations.

No equivalent vocabulary exists for the productive exploitation of the builder's reward system. "Productive addiction" sounds like a contradiction — how can productivity be addictive? The absence of vocabulary is itself a feature of the supernormal stimulus: the activity is so obviously valuable that the cultural immune system cannot classify it as problematic. David Kessler argued that the vocabulary gap delayed the public-health response to processed-food addiction by at least a decade. The technology industry currently benefits from the same gap.

Origin

The sugar-salt-fat framework was established in the food-science and public-health literatures across the 1990s and 2000s, synthesized for general audiences in Moss's 2013 book Salt Sugar Fat: How the Food Giants Hooked Us and in David Kessler's 2009 The End of Overeating.

The extension to cognitive labor — specifically the mapping of feedback speed to sugar, progress continuity to salt, and execution completeness to fat — is developed in the present volume as an application of Barrett's supernormal-stimulus framework to AI-augmented work. The structural parallel had been noted informally in the productivity and technology-criticism literatures but had not previously been specified with this mechanistic precision.

Key Ideas

Multi-dimensional exploitation. The most addictive stimuli combine multiple supernormal features simultaneously, each defeating a different regulatory mechanism.

Sugar/salt/fat maps to feedback/continuity/completeness. The three features that make AI tools useful are the three that make them supernormal along every regulatory dimension.

Capability optimization converges on exploitation. Tools designed for maximum usefulness, without regard to regulatory implications, produce the same combination as tools designed deliberately for addiction.

Vocabulary gap protects the stimulus. The absence of cultural categories for productive addiction allows the pattern to operate without social sanction.

Environmental intervention is the template. The effective response to supernormal food stimuli was environmental modification, not individual education — the same template applies to productive stimuli.