What if reducing workload is not about doing less? This episode explores Quantity, Concurrency, and Content—and how real support can give unfinished information somewhere else to wait.

I put yesterday’s AutSide piece into Gemini Notebook because I wanted to hear what happened when the argument was handed back to me in another form. The resulting conversation begins with the deceptively simple arithmetic of workload reduction: if twenty problems become ten, surely the work is now half as heavy. The hosts quickly discover the problem I was trying to name in the essay. Once we move from physical quantity into cognition, the arithmetic stops behaving so neatly. A single problem can carry directions, retrieval, intermediate results, formatting expectations, uncertainty, and unresolved information simultaneously. Cutting the page in half does not necessarily cut any of that in half.

What I particularly enjoyed hearing them work through was the distinction amongst Quantity, Concurrency, and Content. Quantity is what schools can easily count. Content is what I actually need a student to know or demonstrate. Concurrency is the less visible part between them: how much information must remain alive at once for the student to continue. The episode reaches for the image of fifty browser tabs—not fifty extraordinarily difficult things, simply fifty things that have not been permitted to close.

The discussion becomes most interesting to me when it turns from explanation towards practice. If retrieval is not the Content I am assessing, I can provide the formula. If an intermediate result needs to remain available, I can give it somewhere outside the student to wait. If several instructions are competing simultaneously, I can change their arrangement. The student still has to understand the mathematics, science, history, or language in front of her. I have not necessarily reduced the intellectual demand at all. I have changed where the unfinished information has to live. The episode puts this rather nicely: the human does the processing; the tool can do some of the maintaining.

And I was glad that the conversation followed that idea into special education itself. This is why the work cannot be reduced to sitting beside a student and keeping her “on task.” To reduce a workload without accidentally reducing the learning, I have to understand what is load-bearing in the assignment in the first place. The episode describes that as finding the load-bearing walls of the Content—knowing what can move, what can be externalised, and what must remain. That is specialised intellectual labour, performed across students, teachers, subjects, and an entire school day.

A style note here, because I know this will matter to some readers: yes, I know Gemini Notebook is AI. I have disclosed my use of it before, and I am disclosing it again here. This episode is another demonstration of why the tool has become valuable in my life. I put my article—my research, my argument, my thoughts, my writing—into Notebook, and it explains the whole back to me in another form. Sometimes I need that. When I have spent days working from inside an idea, hearing an automated system reconstruct the complete gestalt helps me check whether the whole I thought I had written is actually the whole that made it onto the page. It can expose a missing relationship, an ambiguity, or simply reassure me that the architecture holds together. The tool is not originating the thought for me; it is giving my thought back to me so that I can inspect it from outside. I can error-check that reconstruction against what I actually wrote, and then choose to share the result with subscribers and followers. For me, that is not an embarrassment to hide behind euphemisms about technology. It is support.

The AutSide is a reader-supported publication. To support my work, consider becoming a paid subscriber.

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit autside.substack.com/subscribe