Pattern Templates

In the AI Story

Goldberg's research on expertise, developed in dialogue with studies of chess masters, elite physicians, and expert performers across domains, converges on a single neurological claim: expertise consists not primarily in the memorization of facts or the learning of procedures but in the deposition of pattern templates that allow the expert to perceive situations the way novices cannot perceive them. The expert and novice do not analyze the same situation more or less quickly. They perceive different situations, because the expert's pattern templates restructure the perception itself.

The templates are built through the specific effortful processing that the novelty-routinization gradient describes. Each episode of struggle with a novel problem — the prefrontal engagement, the working memory strain, the failed approaches, the eventual resolution — deposits a partial template. After enough episodes, the template is complete. The next encounter with a structurally similar problem does not require effortful analysis because the pattern is recognized.

The templates are invisible. They cannot be observed in the expert's behavior except by inference from the speed and accuracy of the expert's judgments. The expert often cannot articulate how she recognized the pattern — the recognition happened below the threshold of conscious awareness, produced by neural systems that do not generate verbal reports. This invisibility is one reason expertise has been systematically undervalued in AI discussions: the visible output of expert work is the solution, and the AI that produces the same solution appears to have done the same work. The pattern templates that produced the expert's rapid recognition are not visible in the output and so are assumed, incorrectly, not to be necessary.

For the AI-augmented workflow, pattern templates are the critical variable that explains why the same tool produces radically different quality in different hands. A senior engineer with decades of deposited templates directs the tool with precision — her prompts embed the architectural knowledge her templates have accumulated, her evaluation of outputs draws on the pattern library that distinguishes elegant from merely functional solutions. A junior developer without the templates directs the same tool with adequate but undifferentiated results. The tool is identical. The conducting is categorically different.

Origin

The concept draws on decades of expertise research, particularly the chess studies of De Groot and Simon that demonstrated experts remember board positions by structure rather than piece location. Goldberg integrated these findings with his clinical neuropsychology into the framework articulated most fully in The Wisdom Paradox (2005).

Key Ideas

Expertise as perception. The expert does not analyze faster — she perceives differently, through the restructuring that pattern templates produce.

Templates are deposited through effort. Each struggle with a novel problem contributes to the template that will eventually recognize problems of that class.

The library accumulates. Decades of varied engagement produces a vast template library that is the neurological substance of what Goldberg calls wisdom.

Templates are invisible. They cannot be observed except through the speed and accuracy of expert judgment, which is why their contribution is systematically undervalued.

Templates direct tools. The quality of AI-augmented work depends heavily on the pattern templates the human directs the tool with.