What I wanted to build
Entering Hack the North, I wanted to create an idea that was AGI, but it turns out building AGI is harder than it seems. I was inspired by the OpenAI marketing video for Astra and how it could draw on a screen, but I wanted to do more and wanted to build something that could interact with the real world and bring it one dimension forward, which was making it AR.
And when doing so, I realized how much computing power was needed just to compute something properly, and that the chances of bringing something to life would be too hard to do. Even if we did accomplish it, it would just take too much time, and by the time we finished it the demo period for our project would have been over.
Why I am writing this anyway
And now this begs the question: why are you writing this if you are doing something that has already been done? Well, because it is my own blog, and also I can write about what I learned.
What I learned
What I learned right away is that the space for 3D is just not there yet, and that computing and creating is the real chokehold. As for investing into creating shapes and tools that looked realistic, there were over a million meshes needed to create an item that looks realistic, and that itself takes well over a minute, and so that is why it is so hard.
But realizing this, I think the problem that can be fixed here is creating a tool that can understand, and instead of creating something it can use a library that has certain components and take from it. Maybe that might be the play, but we will see. Thanks for reading.