Separate the Cost of Reading from the Cost of Operating

It may be reasonable to spend effort consuming information.

It should not require unnecessary effort to control the experience itself.

Because they are different kinds of cognitive cost. A learner may willingly scroll through a detailed explanation because reading the explanation is the task. That same learner should not have to search through several screens merely to find the button that submits an answer, moves to the next step, saves progress, or reveals a hint. AI can easily collapse these two forms of movement into a single rule such as "avoid scrolling." That removes an important distinction. A better question is: **What must remain immediately available so the user stays oriented and in control?** Those elements should be treated differently from the content whose purpose is to be explored, read, compared, or studied. This distinction allows depth without creating operational friction.

An AI-generated lesson contains a long explanation, a diagram, and two worked examples. The system attempts to eliminate all scrolling and therefore divides the lesson into six separate screens. Learners now have to repeatedly press Next and can no longer easily compare the first example with the second. The original goal was usability, but the intervention has increased interaction cost. A better design allows the teaching material to scroll naturally while keeping progress, navigation, and the primary action persistently accessible. Reading can require movement. Operating the experience should remain predictable.

More from the same record, pulled by shared topic — not written new for this page.

It may be reasonable to spend effort consuming information. It should not require unnecessary effort to control the experience itself.