Meet the Founder

Hey. I'm Thabani.

I write software for a living. I dance. I play trumpet at church. I also have this habit of grabbing an idea and refusing to leave it alone until I've either made it real or proven it doesn't work. SymbioLearn is the one that's eaten most of my evenings.

The product story (how the Tutor works, why the name exists) lives on the About page. This is more about how I got into this, and why I keep tearing versions down.

University, a PDF, a chat box

2023. Final-year IT project. The brief in my head was: upload a PDF and talk to it. Chat with your notes.

I thought that was the product. You could use it. People could ask questions grounded in their file. I was pretty pleased with myself for a stretch.

Then I actually watched people use it, and the shine came off. It was a clever demo. I wouldn't have studied like that for a whole semester.

35 signups

I put it online around June 2024. About 35 people signed up the first day. I refreshed whatever dashboard I had like the number meant the idea had landed.

Some of them clicked around. A few tried it. Then it went quiet. I sat there waiting for them to come back, and they mostly didn't.

My first response was the developer one: add more. Of course it was. That's what I would have told anyone else too.

Notes, quizzes, the usual

So yes: notes. Quizzes. Flashcards. More ways to poke at a PDF. I spent months on things other apps already do, and a lot of those apps do them better.

I was busy. I was shipping. The product still felt like something you try once and then go back to whatever you were already using.

I didn't have a cinematic moment where everything clicked. I just kept using my own thing and thinking: wait, would I actually open this tomorrow? Usually no. Then I'd get stubborn and build another piece anyway, because I was convinced the next feature would fix the boredom. It didn't.

The bit I couldn't drop

Students can already get an explanation in ten seconds. Summaries. Quizzes. Answers. I kept adding more of that because it was buildable, and because I knew how to build it.

What I couldn't shake is more annoying: you can read a clean explanation and still not be able to say it back. Or you remember it in the session and it's gone by Thursday. Or the question is worded differently and you're stuck.

I still argue with myself about how much of that software can even fix. I don't have a tidy model I'm presenting as settled. I have a fairly stubborn feeling that if the tool never makes you do anything with the idea, it's just another tab.

That's why the current version looks so different from 2024. I'm not done. Some weeks I rewrite a piece I was sure about last month.

Talking

Voice showed up because I kept noticing the same thing in myself. If I have to explain something out loud, I find out immediately that I don't actually have it. I stall. I skip a step. I ask a question I was pretending I didn't have.

Reading another paragraph is easy to fake. Talking is not.

I didn't pick voice because voice AI is having a moment. I picked it because it's a pain to hide in. The Tutor is there to keep you in that: to ask, to wait, to push when you're being vague. I still fight with how much help is too much. Give too many answers and I've just rebuilt the original problem.

I write the code. Design, database, the boring bugs, ripping a version out when it's wrong. Most days it's me and a laptop. I'm not treating this as a sketch I forgot about. I keep coming back because I think the first version was pointed at the wrong problem, and I want to get closer.

If you want how the product is shaped now, that's on About.

The rest of the week

Dance is a real part of my life. After enough hours in a screen I need to be in my body or I get weird.

Trumpet at church is a regular part of the week. Faith matters to me. I'm not going to make this page about that.

I take on other software work too. That's how I stay sharp, and also how rent gets paid. I've started things that have nothing to do with studying. Chartysis came out of staring at trading charts and wanting a second opinion without pretending an AI knows the future. Some of those ideas stick. Plenty don't. I like making things even when they're messy.

SymbioLearn takes a lot of my thinking. Some weeks I'm sure. Some weeks I throw out a piece I liked. That's been the pattern since the university project, and I don't expect it to stop.