44 million short by 2030. Six in ten are holes left by teachers who quit. They are not missing — the job has become one you cannot stay in. That is where we start.
The problem
Teachers the world will be short.
44 million
teachers short by 2030
273 million
children out of school — a 7th straight year of increase
31.1M / 12.8M
of the gap: secondary / primary
They are not missing. Six in ten are holes left by teachers who quit.
UNESCO Global Report on Teachers 2024 (44M = 25.5M replacement / 18.5M new posts; secondary 31.1M, primary 12.8M) · UNESCO GEM Report 2026
Why they leave
All or nothing.
There is no five-hour-a-week teacher.
You have to be at that school.
Leave, and there is no road back.
All three exclude the same person — anyone who has another life.
Japan MEXT Teacher Shortage Survey FY2025 · UNESCO Global Report on Teachers 2024
WHAT WE ARE DOING NOW
A link to a prototype tells you nothing about what is being tested. Here is what runs today, what gets added, and what December decides.
RUNNING you can touch this today
TO BUILD before the December opening
DECEMBER DECIDES December 2026, A/B
Why start with a Japanese school?
The world's teacher shortage is not something you can reach for directly. We start where demand is clear and outcomes are measurable. Japanese children living abroad have an obvious reason to study Japanese, and progress shows. For testing whether the process works, the conditions are as clean as they get.
"Equivalent" — compared to what?
To classes taught by licensed teachers: same period, same material, comparable children. If the gap stays inside a margin fixed beforehand, we call it equivalent. Fixing and publishing that margin before we start is what makes it a non-inferiority trial — so it cannot be moved afterwards to suit us.
What if it does not work?
We publish it as it is. Whether non-experts can teach was answered twenty years ago by TaRL in India. Our bet is not there — it is on how far the cost of developing a teacher can fall. If it does not fall, this business does not stand up. That is why we measure.
Who can use the prototype?
It is open. But it is still in development, and the teachers and lessons in it are fictional examples, not results. If you try it and something is confusing, we would like to hear it.
All we can currently show is that people can be developed without a licence as the entry condition, and put in front of children. Whether that can be done cheaply is something nobody has verified. December exists to measure exactly that.
The prototype is in development; teachers and lessons shown in it are fictional examples, not results. TaRL = Teaching at the Right Level (Pratham / J-PAL).