What is Homotopic Education?
In mathematics, a homotopy is a continuous path between two shapes — a way of smoothly transforming one thing into another without tearing or breaking it. Two things are "homotopically equivalent" when you can get from one to the other by a continuous deformation. They look different, but they're structurally the same.
This is the key insight behind everything we build: the subjects your child learns at school are not separate things. Mathematics, language, science, art, history, computing — they are all paths on the same landscape. You can continuously walk from one to the other. A child exploring train routes is doing graph theory. A child adjusting a volcano simulation is doing physics. A child noticing that two different words describe the same feeling is doing topology.
Formal education treats this landscape as a set of disconnected boxes. Maths is in Room 3. English is in Room 7. They have different teachers, different textbooks, different tests. And then the system tells your child they belong in some boxes but not others.
Homotopic education refuses this fragmentation. We build environments — we call them semantic toys — where the connections between disciplines are alive and visible. Where a child can follow their curiosity from trains to maps to geometry to cultural geography without hitting a wall labelled "that's a different subject."
Why Now? The AI Revolution Changed Everything
Something has happened in the last few years that makes this approach not just philosophically appealing, but practically urgent.
Artificial intelligence has dissolved the boundaries between disciplines — permanently.
The old world: specialists in boxes
In the 20th century, you could build a career inside one box. A mathematician didn't need to write well. A historian didn't need to understand statistics. A programmer didn't need philosophy. The boxes were stable, and the education system was built to sort children into them.
The new world: everything is connected
That world is gone. Consider what it actually takes to work with AI today:
- Prompt engineering is applied philosophy. Writing effective instructions for an AI requires precisely the skills English and Philosophy departments teach: clarity of thought, logical structure, awareness of ambiguity, rhetorical precision. The "soft" skills turned out to be engineering skills.
- History now requires numerical methods. Embedding spaces, vector search, computational text analysis — a modern historian who can't think quantitatively is locked out of entire archives and methods. Meanwhile, the AI researcher who doesn't understand historical context builds systems that repeat historical mistakes.
- Mathematics and natural language are converging. Large language models are, at their core, mathematical objects operating on language. Understanding how they work requires comfort with both. The old division — "maths people" vs "words people" — is not just wrong, it's actively harmful to the next generation.
- Physics is dynamic systems theory — and so is ecology, economics, and medicine. The same mathematical structures that describe fluid dynamics describe disease spread, market behaviour, and climate. A child who understands dynamical systems understands everything.
What this means for your child
The children growing up right now will live in a world where the ability to move fluidly between disciplines is the core skill. Not "knowing maths" or "knowing English" as separate competencies, but the capacity to see that they are different views of the same structure — and to walk the path between them.
This is exactly what homotopy describes mathematically. And it's exactly what our tools are designed to nurture.
A child who plays with our earthquake simulation isn't just "learning science." They're developing intuition for wave propagation, magnitude scales, geological layers, and spatial reasoning — simultaneously. They're building the kind of joined-up, cross-disciplinary thinking that the AI-transformed world demands.
The False Belief We're Here to Prevent
"I'm not good at maths."
No child should ever say this sentence. It's not a statement about the child — it's a false belief about how knowledge works, installed by a system that treats disciplines as separate lanes.
The formal system takes a child who is a seamless whole — curious about everything, making connections between everything — and slices them into subject-shaped pieces. Then it measures each piece separately. Then it tells the child: you belong in some lanes but not others.
This is disciplinary lock-in. And it's not a bug in the system — it's the system working as designed. Schools built for the industrial era needed to produce specialists: people who would stay in their lane and not ask why the lanes exist.
But the lanes were always artificial. Mathematics IS a language. Physics IS applied mathematics. History IS pattern recognition. Philosophy IS the study of what counts as evidence. They were never separate — we just taught them as if they were.
In the AI era, maintaining these artificial walls is more than philosophically lazy — it's practically outdated. A child told "you're not a maths person" at age 8 is a child locked out of half the tools they'll need by age 18.
We build these tools to prevent disciplinary lock-in before it starts. In our environments, there is no "wrong." There is only: what happened when you tried that? What do you notice? What would change if you moved this slider? The child's relationship with structure stays alive, unjudged, open.
The Manifesto: 10 Principles
These guide every app, game, and learning environment we build.
- Learning is generated, not delivered. Your child produces understanding through action — exploring, building, breaking, remaking. We never lecture. We prepare environments where discovery happens.
- Disciplines are paths, not boxes. Maths, language, science, art — they're different views of the same landscape. Our tools keep the paths between them open and walkable.
- Memory is sacred. We design for return. A child who comes back to a simulation finds deeper layers, not repetition. Growth becomes visible over time.
- Play is the deepest proof. If your child explores freely and discovers that a loop always returns to the start, that IS mathematical reasoning — whether or not anyone names it as topology.
- AI is co-explorer, not examiner. Technology in our tools reflects, suggests, and responds. It never commands, grades, or judges. Your child remains the author of their own understanding.
- Curriculum emerges from curiosity. We follow the child's interest. If trains are the obsession this month, then trains become the vehicle for graph theory, geography, and systems thinking.
- Every constraint has a reason. Every rule in our tools has a justification the child could eventually understand. "Because I said so" is not in our vocabulary.
- Home is the first learning environment. Your family's languages, stories, interests, and rhythms are the foundation — not obstacles. We build from what your child already loves.
- Parents are co-witnesses, not wardens. Your role is to explore alongside, to be surprised together, to dialogue. Not to police, test, or anxiously measure.
- Joy is the signal. If a learning tool doesn't feel warm and joyful in practice, something is wrong with the design — not with your child.
How Children Actually Learn
Don't underestimate them
A five-year-old can think about topology if it comes through trains and loops. A six-year-old can reason about computational linguistics if the words behave like spells. The ceiling on what children can understand is set by presentation, not by age.
The usual instinct is to simplify — reduce, flatten, make "age-appropriate." We do the opposite. We make complex ideas accessible through play, but we never remove the complexity. A child who plays with our earthquake simulation is genuinely manipulating magnitude scales, wave propagation, and geological layers. They just don't need to know those words yet.
Observation, not evaluation
Our tools give observational feedback, not evaluative feedback. "The bridge held 3 trains" — not "you got 3 right." "The volcano erupted at level 7" — not "wrong answer." The child experiments; the world responds; understanding grows. No marks, no rankings, no lanes.
Dialogue, not instruction
The best learning happens in conversation — between child and parent, child and tool, child and the world. Not instruction from above, but genuine dialogue where both parties can be surprised. When you sit with your child at one of these apps, you're not teaching. You're exploring together. You're allowed to say "I don't know — let's find out."
How to Use These Tools
Each app is designed around a six-phase cycle. You don't need to follow it rigidly — children will jump around, and that's perfect. But knowing the phases helps you see what's happening:
Spark
Find what the child is already curious about. Trains? Volcanoes? Weather? Start there.
Play
Open the app. No instructions, no goals. Let them poke, slide, tap, explore.
Name
When they notice a pattern — "it always comes back!" — help them name it. "That's called a loop."
Deepen
Introduce the deeper structure. Not as a lesson — as a richer layer of the same play.
Remix
Let them change the rules. What if the wind was faster? What if the route went backwards?
Witness
You observe and remember together. Not to grade — to celebrate and return to later.
Practical tips
- Let them fail. If the tornado destroys the house, that's data, not disaster. "What magnitude was that?"
- Ask, don't tell. "What do you think will happen if you turn it up?" beats explaining every time.
- Come back. The apps are designed for revisiting. A child who returns in a week will notice things they missed.
- Follow sidetracks. If the earthquake app leads to questions about buildings, go there. The curriculum is wherever the curiosity leads.
- Play together. These aren't screen-time babysitters. Sit with your child. Be surprised together. That's the whole point.
Our Influences
This work draws on a century of radical thinking about how children actually learn:
- Ivan Illich — Deschooling Society. Learning webs, not funnels. Education embedded in life, not walled into institutions.
- Paulo Freire — Pedagogy of the Oppressed. No banking model. Knowledge is generated through dialogue, not deposited into passive minds.
- Seymour Papert — Mindstorms. Constructionism: children as builders of their own understanding. The computer as a material for thinking.
- Maria Montessori — Prepared environments. Follow the child. Respect their autonomy and capacity for deep concentration.
- Lev Vygotsky — Zone of Proximal Development. What a child can do in dialogue today, they'll do independently tomorrow.
- Marvin Minsky — Society of Mind. Intelligence as a community of processes. Epistemic diversity as strength.
- Humberto Maturana & Francisco Varela — The Tree of Knowledge. Cognition as a living process. Knowing and being are inseparable.
- Alexander Grothendieck — Mathematics as practice. The discipline of seeing structural unity beneath apparent difference — which is exactly what homotopy formalises.
- Homotopy Type Theory (HoTT) — The mathematical foundation: types as spaces, proofs as paths, equivalence as continuous deformation. This is the formal language behind "disciplines are paths, not boxes." For the technically curious: Iman Poernomo's Rupture and Realization extends HoTT into Dependent Open Horn Type Theory, grounding the semantic manifold these tools explore.
About This Project
Lushka is built by Asel Poernomo, an educationalist and researcher in AI-assisted learning. This platform is her active research — designed, built, and used daily as part of her son's education and shared with a growing circle of families. As the programme evolves, it will expand on tanazur.org as an open resource for parents and educators pursuing homotopic, cross-disciplinary learning.
The apps on this site are open, evolving, and made with love:
- Train Map Library — Metro, tram, and rail maps from cities around the world. Browse by continent, country, and city.
- Nature's Big Powers — Interactive simulations of earthquakes, volcanoes, tsunamis, tornadoes, lightning, and avalanches.
- Earth Science Explorer — Five bilingual games about microbes, weather, earth layers, ice friction, and fog.
- Isaac's Train Page — Where it all began. Isaac's very first web page, preserved as it was.
Everything uses simple, open technologies — HTML, CSS, JavaScript, Python, SQLite. No tracking, no analytics, no ads. A curious parent or child can read the source code.
The name
This site is named for Lushka Poernomo — Iman's mother, a home educator in 1970s Australia at a time when home education wasn't legal. Inspired by Ivan Illich, she fought institutions for the right to teach her children according to their actual interests and capacities. Her method was agile curriculum design: plans crafted carefully, torn up when her children's interests shifted, rebuilt from whatever fascinated them next. No fixed syllabus. No standardised testing. Learning as living practice.
What she did with books, conversation, and stubborn courage, we extend with computation, simulation, and AI. The spirit is the same: follow the child, prepare the environment, trust the process.