My project is going pretty well. I got tired of programming the framework since there's not really any immediate results, so I've been creating some cool (and useless) image filtering apps that run on the framework. I wrote a small dithering demo last night that looked pretty cool. I also want to write a cell-shading app since I've never done that before. My problem is that I can't load models yet, so I can't really have any complex scenes to test on. I'll start posting screenshots once I can load models because it looks rather bland otherwise.
On that note, I've never worked with complex models either. I wrote a quick OBJ loader once, but it was pretty useless. I want to write a loader that supports 3DS files and maybe some other format, as long as it supports animation.
Anyway, with so much to learn, I bought a couple books. Namely Focus on 3D Models to learn about models and loading them. I've also always been fascinated with Neural Nets, but my experience with AI in general is very limited, so I also bought AI Techniques for Game Programming. I wasn't going to jump into AI yet, but noaktree certainly inspired me with his lawn-mowing AI. Cool stuff!