Inspiration

I'm really interested in world models ever since working on a crowd simulation tool and running into certain limits. So I began thinking about things we can and do model beyond text and I was curious what it would mean to model's one's learning.

What it does

It essentailly learns about your learning style through various (infinite games). In turn you learn in various and each time get a more personalized engaging lesson. There's various componets but it personalizes learning games to you so that you can better understand concepts of your choosing.

How we built it

My techstack was largely Godot. A lot of it was testing theories to see if something was possible in the Godot game engine.

Challenges we ran into

One of the hardest parts was having to pull off a remote code execution on my system. That and hotloading the edited and generateed games into place without exiting the game.

Accomplishments that we're proud of

As i get older and more experienced itt will always nice to acknowledge you work in spaces with less and less tutorials to guide you. Eventhough it's very difficult it's cool to know you're doing something new.

What we learned

Remote code execution and hot loading are very difficult to deal with and precise care is necessary.

What's next for

I think there's more that can be done both within and beyond my compute. I would love to see larger research labs take up this efffort. Personally I want to develop to create more long horizon games

Built With

  • claude
  • gdscript
Share this project:

Updates