Philosophy · 2026-09-24 · 1:50

When the agent does the work, the scarce input is the open question

Name: When the agent does the work, the scarce input is the open question
Uploaded: 2026-09-24
Duration: 1 min 50 s
Description: For most of working history the scarce resource was execution: hands, hours, headcount. As execution gets cheaper the bottleneck moves to the question, and that part stays yours. Worth practising.

A 2023 study sorted every question 165 undergraduates wrote in two question-asking tasks from the Torrance tests: open or closed, possibility thinking, thinking across time. The clearest result, holding across all four sets of analyses, is that people who scored higher on originality and fluency asked more open-ended questions, the kind without one correct answer.

In favour

It gives measurable shape to something usually left vague, and it lands where work is heading: as agents take over execution, the human contribution moves upstream to framing.
A closed question gets a correct answer. An open one gets a direction nobody specified in advance, which is exactly what a system optimised to answer cannot supply for itself.

Worth watching

Fluency is partly a count of questions, so more questions mechanically means more open ones; the interesting link is the one with originality.
Divergent-thinking tests are a proxy for creativity, not creativity, and several effects showed up in one form of the test and not the other. A careful small study, not a law.

Our takeFor most of working history the scarce resource was execution: hands, hours, headcount. As execution gets cheaper the bottleneck moves to the question, and that part stays yours. Worth practising.

Source: Acar, Berthiaume & Johnson, «What kind of questions do creative people ask?», Journal of Creativity 33 (2023)

← All Focus posts

AI-narrated · AI-generated content

Transcript

When the agent does the work, what becomes the scarce input? A small study from 2023 points to the questions people ask. It sorted every question that 165 undergraduates wrote in two question-asking tasks. Open or closed. Possibility thinking. Thinking across time. The clearest result held across all four sets of analyses. People who scored higher on originality and fluency asked more open-ended questions, the kind without one correct answer. It gives measurable shape to something usually left vague. And it lands where work is heading. As agents take over execution, the human contribution moves upstream, to framing. A closed question gets a correct answer. An open one gets a direction nobody specified in advance. Exactly what a system optimised to answer cannot supply for itself. Two things to keep straight. Fluency is partly a count of questions, so more questions mechanically means more open ones. The interesting link is the one with originality. And divergent-thinking tests are a proxy for creativity, not creativity. Several effects showed up in one form of the test and not the other. A careful small study, not a law. For most of working history, the scarce resource was execution. Hands, hours, headcount. As execution gets cheaper, the bottleneck moves to the question. And that part stays yours. Worth practising.