Why Lumen is built this way

We believe a child can hold a real university idea. Not a cartoon of it. The structure of the thing itself — inertia, a standing wave, a proof — in their own hands, at their own age.

What fails is not the child. It is the speed of the lecture, and the school habit of rereading until the page feels known. Known is the wrong word. Familiar is not the same as being able to use it. A mind owns an idea after it has looked, chosen, gotten a question wrong, said why the other answer was wrong, and found the idea again a week later without the paragraph in front of it.

That work does not need a dashboard. It needs a table, a parent in the next room, and a lesson that will not pretend they are finished. If that is not what you believe, Lumen will feel like a fuss. If it is, the rest of this page is only how we keep the promise, and what that looks like on a Wednesday.

How we keep it

These are not features. They are the refusals that make the belief true. Each one is a thing you will notice, then the finding — named in Sources — that made us refuse the easier product.

One idea, then a bar

Each lesson is one idea. The test is whether they can do that idea — not whether they sat through it. A sixty percent and a green check would feel kind. It would also let the child leave without the skill. We wait until they can do the work.

Trade schools have done this for a long time: you are finished when you can do the job, not when you outscored the room. The Navy’s Ready Relevant Learning program says the same thing in uniform — train at the right time, and do not call someone finished because they sat through the chapter. A quiz-show score is not a bar.

Look before you read

The picture is the problem, not a decoration for the paragraph they just read. We ask them to mark which object changed, or sketch the pair, before the idea is written out. For ages seven and under we keep this light — the picture itself is enough to hold.

A mark in the world and a sentence about it beat either alone. Students who explain why an answer is right learn more than students who are told. After a correct pick we still ask, in a sentence, why not the other one. Dual coding, self-explanation, and the ICAP work are in Sources.

Harder on purpose

Doing ten of the same kind of problem in a row feels fluent. Fluency is not the learning. The finding on desirable difficulty is simple: the feeling of ease is a bad teacher. A mathematics study of mixed practice found that mixing problem types — so the student must choose the method — beats a block of one type on tests given later, even though the block feels easier. Technicians fail in the field when they can only do the chapter they just finished.

So after the first lesson, practice includes an item from an earlier idea. The exam at the end of a lecture mixes them. That is not cruelty. That is the test they will actually meet.

Do not come back tomorrow to reread

A next-day reminder that restates the lesson is rereading with a calendar. The work on test-enhanced learning exists to kill that strategy. Pulling an idea back from memory beats reading it again. The spacing reviews mapped the ugly part — the right gap grows with how long you want the memory to last. We use one day, then three, then seven, then twenty-one, with a short recap, one old check, and one old “look at this” prompt. A review of learning techniques puts practice testing and spreading study out in the small set that actually holds up. Those papers are in Sources.

Ready Relevant Learning’s other half is help at the moment of use. The printed pages are that object. The library’s first card is the calendar, not a file drawer.

The paper is not a screenshot

Screen and paper do different jobs. The screen can hide a picture and ask them to draw it from memory. The paper has to survive a kitchen, a backpack, a parent who was not in the lesson. So what you print is a document: a cover, the lesson, practice, a test, and an answer key. It is meant to be written on.

Do not say it twice at once

Spoken words plus the same written words plus a picture is three versions of one idea, fighting. Multimedia research calls this redundancy. Sound here is the line for that moment, with the picture, or a listen-only pass. Playing the whole lesson aloud while the text is on screen is a second encoding we refused.

Give the eye one place to land

Early pictures were full of arrows. Arrows are a diagram pretending to be a picture. Work on visual thinking is blunt: marks compete. If everything points, nothing does. We keep pictures quiet — cream paper, teal, almost no decoration — and we keep motion for change over time, not for definitions. Cognitive load research is why ages seven and under get fewer objects and almost no diagram language. The idea does not get dumber. The picture gets quieter.

The record stays with the child

Names, scores, due dates, certificates, and the transcript live on this device. That is not a missing login. A kitchen table does not need a cloud to remember that Maya passed a lecture on the first try. “With distinction” means that, and only that. We do not put a borrowed seal on the certificate. The diploma is not ours to borrow.

What we will not do

• Daily streaks, leaderboards, or a lesson that unlocks because time passed.
• A next lesson after a weak score. They move on when they can do the work.
• Spoken narration laid over the same paragraph they are reading.
• Reread-tomorrow as the review engine.
• Practice that never asks which idea this is.
• Borrowed seals, course codes as branding, or a mark that pretends this is a degree.
• An account, a feed, or a child’s name on a server we do not need.

What that looks like

The what is proof of the why. Lumen rewrites a real university lecture into short lessons a child can finish at the table. Younger ages get more, shorter lessons. Age 18 keeps every concept from the original. We do not start from the catalog. We start from the belief. The catalog is what we happen to rewrite.

1. You pick a lecture. Lumen writes a short set of lessons. One idea at a time, with a check after each.
2. The lesson opens on a picture, not a paragraph. The child looks first, then reads.
3. After each idea, a check. After the lesson, practice and a test. If they cannot do what we said they would be able to do, they are not finished. They get a short loop, not the next lesson.
4. You can print the lesson. The paper is a different object: cover, pictures, lines to write on. That is what they bring back later.
5. The lesson returns in 1, 3, 7, then 21 days. Not a daily streak. Memory, with one old question mixed in.
6. When the lecture is done, a certificate. When the class is done, a quieter one, and a transcript of every lecture they actually finished. Progress stays with the child on this device. No account.

Sources

The letter keeps the findings. The names live here. Lumen is not a Navy program, not a university of record, and not a trial of these papers. The papers are why we refused the easier product.

• MIT OpenCourseWare — the lectures. Not an official MIT publication.
• Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., & Rohrer, D. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. Psychological Bulletin.
• Cepeda, N. J., Vul, E., Rohrer, D., Wixted, J. T., & Pashler, H. (2008). Spacing effects in learning: A temporal ridgeline of optimal retention. Psychological Science.
• Roediger, H. L., & Karpicke, J. D. (2006). Test-enhanced learning. Psychological Science.
• Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., & Willingham, D. T. (2013). Improving students’ learning with effective learning techniques. Psychological Science in the Public Interest.
• Pashler, H., et al. (2007). Organizing instruction and study to improve student learning. Institute of Education Sciences.
• Rohrer, D., Dedrick, R. F., & Stershic, S. (2015). Interleaved practice improves mathematics learning. Journal of Educational Psychology.
• Bjork, R. A. (1994). Memory and metamemory considerations in the training of human beings. In Metacognition: Knowing about knowing.
• Mayer, R. E. Multimedia Learning. Redundancy: spoken words plus the same written words overload.
• Paivio, A. Dual coding. Chi, M. T. H., self-explanation (1989). Chi, M. T. H., & Wylie, R. (2014). The ICAP framework. Educational Psychologist.
• Sweller, J. Cognitive load theory. Ware, C. Visual Thinking for Design.
• U.S. Navy, Ready Relevant Learning — train at the right time; help at the moment of use.