Fair point, though I don't think that's really the problem persay.
I get discouraged when I feel overwhelming or like I'm underperforming, which was the case at harvard because of the quality of my peers and some emotional immaturity on my part. I perceived a great deal of effortless talent from our IMO guys and other classmates which got to me when I was young. This did change even during my time at college. By my senior year I was taking and enjoying math classes again with students of great calibur. I enjoyed those classes, and I didn't get discouraged.
What I meant by that comment is simply this: There is a massive intelligence differential between me and Ph.D students here. I already know from experience what math ph.D's at the ivy league are like, and it's above my paygrade. It takes me longer to learn things, usually many many passes, I wouldn't be able to do research with significantly more classes, etc.
The point of that comment wasn't that I don't feel comfortable with smart people, It was just a way of saying "I know what I am not cut out for." Just a statement of acknowledgment re: my limitations and what I need to learn effectively.
As for research interests, I really like dynamical systems and slightly less pure mathy stuff than I studied in school. Discrete things, like packing/optimzation and combinatorial problems (reading Wilson&Lint atm). I also wouldn't be planning on going to a Ph.D program immediately. I was more wondering what the best practical path to take would be if I did, after continuing to self-study? Or rather, does there exist such a path, where does it lead, etc.