The Rep You Skipped
Cognitive Systems AI Systems

The Rep You Skipped

Your memory is a library with a librarian. Only the librarian can get out of practice.

Ibrahim AbuAlhaol, PhD, P.Eng., SMIEEE

AI Technical Lead

Published: September 14, 2026 | Reading Time: ~6 min read

Three causes of bad memory are commonly attributed to Ibn Sina: illness, distraction, and laziness. The attribution travels a lot further than anyone can source it, so treat it as inherited folk wisdom rather than a quotation. It is still worth a look, because two of those causes have nothing to do with the device in your hand, and the third one that device made much worse.

Illness is a doctor's business. Distraction is about how you handle information in the moment it arrives, and I wrote about that in The Memory Workbench.

That leaves laziness. Most people skip past it, because it sounds like an insult. It is not. It describes something quite specific, and it is the one thing your tools quietly take away from you.

A library and a librarian

Ibn Sina (980 to 1037) was a Muslim polymath who wrote in Arabic. A thousand years ago he described memory as two separate things, and the split has held up remarkably well.

Picture a library.

The first part is the shelves. Everything you have ever learned is filed there. Ibn Sina called this al-hafiza. The shelves are passive. They hold books and do nothing else.

The second part is the librarian. When you need something, the librarian walks the aisles and brings it back. Ibn Sina called this al-dhakira, and he was careful to say it is not a place where anything is kept. It is an act. It is the walk.

Walk that aisle often and the path stays worn and easy to follow. Stop walking it and nothing on the shelves changes at all. Every book sits exactly where you left it. The only thing that fades is the trail between them.

That is the whole idea. Everything else follows from it.

The two parts of memory Ibn Sina kept separate The shelves hold everything you have learned and never do any work. The librarian walks the aisles and brings a book back when you need it. Because only the librarian does anything, only the librarian can get out of practice. Two parts, only one of them works AL-HAFIZA The shelves Everything you have learned is filed here Passive. Does nothing. AL-DHAKIRA The librarian Walks the aisles and brings the book back Active. Can get rusty. fetches from Only one of these two can get out of practice.
Figure 1. Ibn Sina separated the part of memory that keeps things from the part that goes and gets them. Sources: Stanford Encyclopedia of Philosophy, "Arabic and Islamic Psychology and Philosophy of Mind"; Encyclopaedia Iranica, "Avicenna vi. Psychology."
Shelves do not get out of shape. Librarians do.

Why laziness is the one that bites

Read the three causes again with that picture in mind and they stop being a flat list. Illness damages the building. Distraction means the book never got filed at all. Laziness is something else: the librarian stops walking.

Shelves cannot weaken, because they never do anything in the first place. The librarian is the only part of the system that works, which makes it the only part that can get slow, and the only part that can get faster.

Science arrived at the same place much later. In 2006 Henry Roediger and Jeffrey Karpicke showed that testing yourself on something makes you remember it far longer than reading it again, even though reading again feels more productive while you do it. What helps is not putting information in. It is pulling it back out. Robert and Elizabeth Bjork gave the principle a name, a desirable difficulty: the effort of retrieving is the thing that works, so removing the effort removes the benefit.

Ibn Sina ran no experiments. He simply noticed which part of memory does the walking.

The four hundred millisecond problem

Here is where your tools come in.

You are reading something and cannot remember what a word means. For most of human history, what happened next was this. You paused. You reached for it. You either got it or you did not, and then you looked it up. That reach takes two or three seconds, and that reach is the exercise.

An AI answers in about four hundred milliseconds. The answer arrives before the reach has even started.

That is the entire problem, and it is worth saying precisely, because the usual worry about AI and thinking is too vague to act on. Nothing is wrong with the answer. The answer is often better than what you would have remembered. The trouble is when it arrives. It lands in the exact moment your librarian would have stood up, so the librarian stays seated. Do that forty times a day and you have not lost a single book. You have stopped sending anyone to fetch them.

You will not notice for a long time, because the shelves are fine. Everything is still there. What fades is your ability to reach it on your own, and you tend to find that out at the worst possible moment: in an interview, in front of a customer, in a conversation where pulling out your phone would look strange.

Three ways to get an answer and what each does to your librarian Remembering on your own sends your librarian down the aisle, which counts as practice. Asking a model gets the same answer while your librarian stays seated. Having the model ask you sends your librarian down the aisle under harder conditions, which counts as practice and more of it. Same answer, three different effects on your librarian YOU REMEMBER IT YOURSELF a question your librarian the answer Slow, sometimes wrong, and it counts as practice. YOU ASK THE MACHINE a question the machine the answer your librarian, seated Fast, usually right, and no practice happened. THE MACHINE ASKS YOU its question your librarian your answer Harder than you would pick, and it counts double. Amber is you. Blue is the machine. Only the direction of the question changes.
Figure 2. The middle row is what almost everyone does by default, and it is the only one of the three that leaves your librarian sitting down. Sources: Roediger and Karpicke (2006); Bjork and Bjork (2011).

Let it ask instead

The same tool is also the best memory trainer ever made, and almost nobody sets it up that way.

Quizzing yourself with flashcards has been around forever, and most people give up on it, because making the cards is boring and they go out of date. An AI fixes both problems. It can read what you read this week, ask you about it in words you have not seen before, tell whether your answer was genuinely right or only close, and come back later to the ones you missed.

It takes one instruction and one habit.

You are in quiz mode.

When I ask about something I have already read, do
not answer it. Ask the question back to me, wait for
my attempt, then tell me what I got wrong.

Answer directly only if I say "just tell me".

The habit is the harder half. When you feel the pull to ask, wait. Say your answer out loud first, even a bad one, and then ask. The bad answer is the exercise. Getting it wrong still counts: guessing and missing, then seeing the right answer, beats reading the right answer twice.

The old idea that held up

Ibn Sina thought the brain worked through chambers and humours, and that part is simply wrong. The idea underneath it is not. He separated the part of memory that keeps things from the part that goes and gets them, and he put all the effort on the second one. A thousand years of research later, that is still where the effort lives.

Which leaves AI with two settings that look identical from the outside. Both hand you a correct answer in about a second. Only one of them leaves you able to answer without it next year. The difference comes down to who asked the question.

What to do this week

  1. Before you look up something you have read before, say your answer out loud first. Even a wrong one. That guess is the whole exercise.
  2. Put a quiz mode into whatever assistant you use, with a phrase that switches it on, so choosing it costs you no thought.
  3. Pick five things you should know cold for your work, and have the AI test you on them once a week instead of explaining them to you.
  4. Every so often, answer a real question with the laptop shut. If that keeps getting harder, you now know which part of you is out of practice.

Related Articles

References & Extended Literature

  1. Stanford Encyclopedia of Philosophy. "Arabic and Islamic Psychology and Philosophy of Mind." Stanford University. https://plato.stanford.edu/entries/arabic-islamic-mind/
  2. Encyclopaedia Iranica. "Avicenna vi. Psychology." https://www.iranicaonline.org/articles/avicenna-vi/
  3. Black, D. L. "Rational Imagination: Avicenna on the Cogitative Power." University of Toronto. http://individual.utoronto.ca/dlblack/articles/Aviccogitart.pdf
  4. Roediger, H. L., III, and Karpicke, J. D. (2006). "Test-Enhanced Learning: Taking Memory Tests Improves Long-Term Retention." Psychological Science, 17(3), 249 to 255. https://doi.org/10.1111/j.1467-9280.2006.01693.x
  5. Bjork, E. L., and Bjork, R. A. (2011). "Making Things Hard on Yourself, But in a Good Way: Creating Desirable Difficulties to Enhance Learning." In Psychology and the Real World. Worth Publishers. https://bjorklab.psych.ucla.edu/wp-content/uploads/sites/13/2016/04/EBjork_RBjork_2011.pdf
  6. Internet Encyclopedia of Philosophy. "Avicenna (Ibn Sina)." https://iep.utm.edu/avicenna-ibn-sina/