I couldn't imagine diving into open problems on your own without even undergraduate-level training. Open problems are what you work on for your PhD thesis, and

I couldn't imagine diving into open problems on your own without even undergraduate-level training. Open problems are what you work on for your PhD thesis, and before starting your thesis you take PhD-level courses to get up to date on the state of knowledge, and then a faculty member advises you on your thesis. Without knowing what you don't know, you'd end up reinventing wheels, most likely. > I have been working in high-end theoretical computer science for many year I think you may be discounting the on-the-job training you surely received. Maybe even more important would be the fact that your job helped steer you toward problems worth solving.