The Global Learning XPRIZE (For Education) Announcement is Still Making My week — MEET JANUS

Ajibola Lawal
May 17 · 6 min read

I know you are thinking: this guy is still shouting?

Yes, yes, I am.

But wait first.

You might be worried for me. Like, Jibola what if expectation doesn’t meet reality, what will now happen?

Nothing will happen, if we are to be truly honest. It simply just means that (to use a Christian reference) the veil in the temple is torn. Pandora’s box has been opened. The possibility of creating an autonomous Literacy and Numeracy learning engine is here.

IF kitkitschool and onebillion and the other finalists are unable to deliver on the Global Learning XPRIZE it’s also okay. The next iterations of people trying their hands at it will. What do I mean? Do you remember a time when it was deemed impossible (and a possible health risk) to complete running a mile in under 4 minutes? (How will you remember? Were you born then?) And then a Roger Bannister did it, and then all of a sudden it was so possible that 1,400 people have broken that barrier since then.

Roger Bannister, record breaker (Cr: The Independent)

But I digress.

So why did I bring you here today? Remember yesterday I told you my mind was buzzing with possibilities. Well, I want your mind to buzz along with mine, a little. Let’s dream a little together shall we?

So before we go to dreamland, a few things

  1. I am purposely not going to use too many technical terms, so that my thinking is accessible to everyone. And even when I do, I will break down what the term means to you.
  2. Assumptions going in: The apps are already able to do what the Global Learning XPRIZE set out to do. That is, an app that a child can interact with unsupervised, to learn basic literacy and numeracy skills without the aid of a Teacher.
  3. For the purpose of this conversation, the new application — the one where we improve on the existing Global Learning XPRIZE winning apps — will be called JANUS (This bears no relation to Tony Stark’s JARVIS, but they might be kissing cousins) so that I don’t have to keep saying: Global Learning XPRIZE winning apps and we can delineate our modifications a little bit from the original.
  4. I have purposely NOT reviewed the apps before writing yesterday, and this morning as well. Because I don’t want to be disappointed (the skeptic in me) and because I suspect some of my suggestions may already have been implemented (if they haven’t wtf are they doing?)
  5. No matter how fantastic and magical the things we imagine below, THEY ALREADY EXIST IN ONE WAY OR ANOTHER.
  6. I will give the different pieces of the puzzle first, and then I will tie in, how they all dovetail into each other, at the end.

So let’s do this!

Step 1: Imagine layering adaptive learning algorithms (don’t stress too hard, I will explain this in a second) over the current XPRIZE learning apps.

Adaptive Learning Algorithm — If you’re an 80s baby, you probably remember Dungeons and Dragons and how you and someone else could read the same book, and you guys would end up with radically different outcomes. If you’re a 90s baby, you probably played Pokémon or Legend of Zelda. You and a friend could start out at the same time, but you got different paths based on the choices you made at the beginning of the game. I̶f̶ ̶y̶o̶u̶’̶r̶e̶ ̶a̶ ̶0̶0̶s̶ ̶b̶a̶b̶y̶,̶ ̶I̶ ̶d̶o̶n̶’̶t̶ ̶k̶n̶o̶w̶ ̶f̶o̶r̶ ̶y̶o̶u̶.̶ In essence, this is what an Adaptive Learning Algorithm for Education does, it presents a way to customize the learning path of any child according to how the child interacts with the app. 3 + 4 is too hard? No problem, let’s start with 1 + 1! Do you grasp things faster than most? No problem, we will accelerate your path!

Step 2: Now imagine with me, that as JANUS is constantly customizing the learning path of each child, it is gathering information as it goes. So Tejiri likes to learn with shapes, but Mohammed prefers colors and Chiamaka likes learning with shapes and colors. With this information, imagine that JANUS then classifies all the children using the JANUS platform by how similar their learning paths are.

Step 3. Now imagine that, as a follow-on to Step 2, JANUS is incentivised (imagine it being given a digital cookie) every time it is able to accurately predict a child’s learning path, before having evidence (that gets validated) (This is already possible btw). Not only this, that JANUS is given a virtual cookie when it is able to bring a child to pass a learning outcome, with the shortest number of steps.

Step 4. Now imagine that using Steps 1–3, JANUS can predict with a low margin of error, by simply interacting with a child (note how this dynamic has reversed) when a child is an outlier — whether they have learning difficulties or they have genius level IQ. Imagine being able to spot a genius child running around and rolling tyre in Ugep and make moves to accelerate their scholarship. Not only that, it would also be able to predict a predilection for Math, Art, or even Languages and so forth.

So how do they tie in together? So right now, we have apps that are supposed to teach children the basics of Literacy and Numeracy without supervision, right? Good. Regular education has a defined path for EVERYBODY, and these apps probably do as well.

As a departure from this, JANUS would create customised learning paths for every child it interacts with. But it doesn’t stop there. It would be a waste of an interaction for JANUS to simply learn about one child and stop there, and learn totally afresh with every child it interacts with. This is why Steps 2 and 3 are important. JANUS would in essence be a learning child itself. Constantly learning.

So while baby JANUS is teaching literacy and numeracy:

  1. It is learning about each child
  2. Then it is learning about how children learn, in general.
  3. Then it is learning about different types of children, and how they learn in particular.
  4. It is learning about the most efficient ways to teach children

Such that on May 17 2019, baby JANUS knows nothing about children. But by May 19, 2020 JANUS learns through experience that children from around the Lake Chad area like to learn with numbers more than literacy and thus it can even use numeracy lessons to teach literacy. Thus JANUS is able to offer learning paths and create shortcuts that are able to faster ascertain a child’s learning style, from the very first interaction.

I suspect that, if these steps are layered over the current platforms that won the Global Learning XPRIZE for Education, we may be able to drop the current 1 hour a day for 15 months to 9 months? Or maybe even 6. I mean, already with good Teachers and concerted effort, you can get the average child reading in 3–4 months. If JANUS is able to drop that time down to 6 months. You can solve Nigeria’s problem of 13m out of school children in 2 years!


Listen, I am going to let this marinate in my head for a second first.


That said, the ideas we have imagined here together , are opensource. I don’t care if someone picks it up and makes a billion dollar startup out of it (Let me not lie, I care. It will pepper my soul). But as long as by the end of this decade, every child born in Nigeria has the basics of literacy and numeracy covered (as a first step). If this serves as inspiration for that? I will honestly die a happy man.

P.S. I once wrote a separate treatise about how ideas are cheaper by the dozen, when Dayonmon Bhenk stole my idea and made a product out of it. But that’s a story for another day.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade