Alec Chen
Blog Dev Log

Update | 2025 Feb 4

Hello!

Super Lights Out

Recently I tried out Bryan Johnson’s biological age test. Part of the test is a reaction time test, where you’re asked to download this app where you play this simple game to measure your reaction time. This app was really annoying, i.e. it’s loading for seconds between pages, and makes you go through this whole goal setting onboarding before you can just play the game. So this weekend I made a simpler version, where you just play the game, it stores all your data locally on your device, and that’s it.

You can check it out here (sorry it’s only on iOS).

Check out this comparison of the startup (after you’ve gone through the initial onboarding).

Math Academy

I’ve been doing a lot of math academy. I’m taking the linear algebra course, and I’m progressing pretty rapidly. I also recently made it to the diamond league, which is the top 1% of math academy users (yay!). This leaderboard is more of a metric of consistency than it is raw knowledge/skill, so it’s less impressive than it might sound.

I’m continuing to really enjoy it. It’s always fun to relearn things. I’m seeing/understanding a bunch of things I didn’t the first 1-2 times I learned it. There’s several things that I don’t think I ever really learned properly, but now I am.

I’ve been mildly lost with what I’m doing with my life (as is typical of this age), and doing math isn’t really a part of any grander goal that would give it a purpose. So, I’m just doing math. It’s fun. Idk. It seems like kind of a good thing to do if you don’t really know what you want to do, but you know something roughly technical lies ahead. I feel like I could never learn too much math. (That being said, it’s not like I’m even doing anything that advanced lol).

Noobular

I haven’t worked on noobular nearly as much this month. This is mainly because my wrist pain has been progressing quite a bit, and the only thing I could do to mitigate it was use my computer less. I still have things I’m trying out to help with it, but it takes time to test out different things. If anyone reading this has suggestions, I’m all ears.

Nevertheless, I made some progress. My work on noobular lately has focused on refactoring the basic primitives of the app in order to go from an internet-native textbook, to a digital private tutor. Math academy is so great mainly because it mimicks the coaching of a private tutor. If noobular is to do anything close to what math academy does, it needs to get that right.

Technically what this means: currently the app allows teachers to put in content in the following form:

But this is limited. There are some crucial features of math academy that I don’t do right now:

The key refactor here is to make the most atomic unit of content a knowledge point (this term is stolen from the goat Justin Skycak). I’m not exactly sure how Math Academy does it, but the way I’ve decided to implement things is that a knowledge point represents a piece of knowledge, and practically it just holds a set of questions. Then this can be used across all different pieces of content (modules, review, quizzes, etc.) to introduce, test, and review knowledge.

I’ve made it through most of the refactor and am almost ready to start implementing the features I’m after, but I’m coming to a point where I am rethinking my whole approach with noobular.

I’ve found it very difficult and tedious to create content. A part of this is how it’s supposed to be, but I’ve noticed I’ve lost momentum, partly because of my wrist issues, but partly because I just don’t particularly enjoy putting together content like this. I really admire math academy for rolling up their sleeves and hand-making all their content. But as a one-man team, I don’t think I can replicate math academy’s approach.

I’ve been really enjoying using math academy, and for a while I’ve been wanting to learn other sciences, but haven’t really struck the right approach (and partly I’m no better than everyone else, and struggle to follow through on random things like this). I’m thinking, I may be a better student than a teacher, so maybe I can leverage my strengths here to beta test noobular as a student.

So my plan is to reconsider using AI for content. But not just to assist me with making content, but to teach myself topics that I don’t fully know yet. My belief is that AI is fully capable of teaching the content I want to learn–say, high school chemistry/physics–but I need to convert it’s knowledge into a form that’s suited for accelerated learning, i.e. math academy style.

I’ve been really trying to stay away from the hype train, but this experiment seems really too fun and ridiculous to not try out.

Goal: use noobular to pass either the AP physics or chemistry exam (at a theoretically accelerated pace).

This will definitely take a while, since I’m planning to do a rewrite, and also implement all these new features, but I’ll keep you, my loyal and devoted readers, posted.