I’ve been fascinated with ancient Egypt since I was a kid. I used to build little model pyramids and I genuinely believed that if you put a piece of fruit in the position of the King’s Chamber it wouldn’t go off, and that water placed in that position would somehow make plants grow faster. I was maybe eight years old and already conducting pseudoscientific experiments with oranges and geometry!
As I got older the fascination shifted from pyramids to hieroglyphs. There’s something amazing about a writing system that survived for three thousand years, that we lost the ability to read for another fifteen hundred, and that a French guy cracked in the 1820s using a stone that a soldier tripped over in Egypt. The whole story is pretty cool. Plus who didn’t love the Stargate movies!
So last year I joined the Egypt Exploration Society in London, who run hieroglyph courses as evening Zoom classes. I missed the introductory course and jumped straight into the intermediate level. That was a mistake! The course is currently on Egyptian funerary texts and it requires a significant grasp of grammar, and grammar wasn’t something I had studied in any real depth. I’m now waiting to do the next round of courses beginning with the beginner level in the hopes of catching up.
But here’s the thing that struck me. The course is taught by brilliant Egyptologists who clearly know their subject inside out. The problem isn’t the teacher or the material. The problem is that the course moves at one speed, teaches in one style, and assumes everyone in the class is absorbing information the same way. I’m someone who consumes information in bite-sized chunks, who learns by doing rather than listening, who needs to see patterns and connections before individual facts make sense. Sitting through a two-hour Zoom lecture on Middle Egyptian verb forms is not how my brain wants to learn this.
And it made me wonder how many other people have had the same experience with traditional courses and just assumed the problem was them.
The idea that wouldn’t go away
At this point in my life I’m eagerly learning many new things and wanting to brush up on topics I haven’t studied for years. I studied quantum and statistical physics at university as well as astrophysics, and while I keep up with the news in those areas, I wanted to see whether I could at my age still understand the topics and bring myself up to speed. But I don’t have the time to wade through complex textbooks or take structured online classes. I want to learn at my own pace, in my own style, in a way that maps onto how I’m actually absorbing information right now.
And it occurred to me that there isn’t really a learning system out there that can do this. There are apps like Duolingo and Udacity that are great at what they do, but they still provide a specific training experience designed for a broad audience. They have a particular approach and if it suits you then great, but if it doesn’t then you’re just another student who bounced off the platform.
What I wanted was something that could understand my starting point, figure out how I learn best, know what I already know about a topic, and then build a completely bespoke course around me. Not a one-size-fits-all curriculum with some adaptive difficulty bolted on, but a genuinely personalised learning experience that changes its approach in real time based on how I’m responding.
So I built one. I’ve been building a lot lately!
Lantarn and the Adaptive Learning Protocol
The system I’m building is called Lantarn (although I migt change the name to be honest). At its core is something I’ve called the Adaptive Learning Protocol, which is a set of rules and structures that any AI teaching system could use, regardless of what it’s teaching.
When you first sit down with Lantarn, it doesn’t throw content at you. It starts by understanding you. It asks about your education history, how you like to learn, what your objectives are. Then it presents information in various ways, sometimes as objective facts and statistics it asks you to remember, sometimes coaching information through an interesting story-like narrative, until it understands how you best absorb and retain information. It builds what I call a cognitive profile, which is essentially a map of how your brain works when it’s learning.
Then it takes the subject matter and organises it into a knowledge graph, which is a structured map of concepts and how they relate to each other. For Middle Egyptian hieroglyphs, this graph has 25 nodes covering everything from phonograms to logograms to determinatives, organised in a sequence that makes pedagogical sense.
Here’s the bit I’m most excited about. The knowledge graph isn’t something I had to build by hand from my own expertise. I was able to feed my entire collection of study materials into the system, all the PDFs, research papers, notes, and course material I’ve accumulated from Gardiner’s Egyptian Grammar, Allen’s Middle Egyptian textbook, and the EES course itself, and the system was able to parse and organise all of it into a coherent training plan. If I knew all of this material well enough to structure a course myself, I wouldn’t need the system in the first place!
As you learn, Lantarn uses a modified spaced repetition algorithm (based on SM-2, the same system Anki uses) to schedule reviews at the optimal time for retention. But unlike a simple flashcard app, it’s also monitoring your understanding and engagement in real time. If your scores start dropping or it detects that a topic is becoming more complex than your current approach can handle, it adjusts its teaching methodology on the fly. It might switch from abstract explanation to concrete examples, or from text-heavy content to visual diagrams, or from rapid-fire questions to a slower narrative exploration.
Why this matters beyond my hieroglyphs hobby
The protocol is topic-agnostic. That means you can swap out the hieroglyphs knowledge graph for music theory, business finance, programming languages, medical terminology, or literally anything else that can be structured as a set of interconnected concepts. The learning engine doesn’t care what it’s teaching. It cares about how you’re learning.
I built it this way deliberately, partly because as a polymathic learner I want one system that works across everything I’m interested in, but also because I think education itself needs a tool like this.
Here’s what I genuinely believe. The moment you walk in the front door of a university, many of its courses are already out of date, especially in areas like technology and artificial intelligence. The learning style that most students go through, rote memorisation, homework, exams, is a system designed for a different era. It cannot suit the learning style of every student, and I think that’s why so many children not only hate school but drop out and fail. It’s not because they’re not capable. It’s because the system wasn’t built for them.
We now have access to AI that has been trained on the entire corpus of human knowledge. It should be relatively straightforward for an AI-based tool to understand a student’s starting point, their learning style, what they already know, and the best ways to teach them, and then to integrate that with their actual objectives, which nowadays aren’t necessarily a degree. It could just be a desire to learn a particular topic for a particular reason, or for the pure joy of learning.
The classroom isn’t dead, but it needs to change
I wouldn’t say traditional education is broken, but it’s certainly not optimal. I don’t think the idea of a classroom is irrelevant, especially for socialisation, which I think is actually suffering in the modern world. We need students to socialise more and interact more with people! But I think the learning aspects need to be much more individual, much more unique, and much more tailored to each student.
Lantarn could work alongside traditional teaching methods, as a companion that helps a student study at home in a way that’s tailored to them. But it could also work independently for someone like me who simply wants to learn Egyptian funerary texts or brush up on quantum mechanics without enrolling in a formal course. The source material can be provided by the student themselves, textbooks, online material, PDFs, whatever they have, and the system integrates all of it into a training plan so that the student gets a comprehensive look at a topic that isn’t skewed by a particular author or teacher’s perspective.
What I’m building towards
I want Lantarn to be a self-contained platform where knowledge topics can be freely available and downloadable, or custom knowledge bases can be put together by anyone who wants to learn something specific and unique. A retired teacher who wants to learn astrophysics. A teenager who’s obsessed with ancient languages. A business owner who needs to understand accounting but learns better through stories than spreadsheets. All of them should be able to sit down with a system that understands how they think and teaches them accordingly.
The goal is simple, really. Make learning enjoyable again. Make it personal. Make it something that adapts to you instead of demanding that you adapt to it.
I’m going to keep building this. The hieroglyphs domain is my test case, and I’m using myself as the first guinea pig, which means I’m simultaneously trying to learn Middle Egyptian and debug the system that’s teaching me, which is exactly as chaotic as it sounds!
If this interests you
If you’re an educator, a subject matter expert, or someone who has always wanted to learn something but couldn’t find the right course, I’d love to hear from you. I’m particularly interested in people who have domain expertise and would like to see their knowledge turned into an adaptive learning module.
And if you’re curious about hieroglyphs, I also have an app for that. Try Hierolyte on the App Store.