Matt, all of the following are probably true. I graduate:
(a) two years from last month.
(b) way too late, because so many interesting companies are doing so many interesting things right now.
(c) way too soon, because answering the Big Questions is probably going to need the kind of deep thinking that tends to be enabled by academia and/or monasteries, not the Real World.
The Big Questions I have in mind are hard to even express clearly right now, much less answer. I think of the family of Big Questions as all following the pattern “how do you find a GPL2-like sweet spot which in one swoop provides more rights for users, solves coordination problems for developers, and leaves room for economic incentive for companies.” If you can find that sweet spot (or something related; this is a gross oversimplification in some ways) you’ve done some really interesting thinking and solved a lot of problems for people.
GPL itself is a fairly good example of a legal document which creates this sweet spot around traditional client software. We need to create the same sweet spot around user data, to encourage users to give up their data to the googles of the world while giving them a high degree of confidence that their rights will be respected. And we need to create the same sweet spot around web services and APIs- Bungee may be doing very interesting things, but if their legal relationship to the folks they serve was standardized and gave incentives to every party in the transaction I’m certain even more interesting possibilities would open up. And of course I still quixotically insist that such a position must be possible in trademark :)
Because copyright is not involved, solutions for these problem spaces probably look more like contract than license, but they may include other aspects of the law as well. For example, the solution to the personal data problem may borrow from common-law property rules like bailment, or (in the worst case) from regulatory law like that which created the SEC. More generally, the complex legal redirection that is copyleft is probably fairly straightforward compared to some of what would be necessary to solve some of these. That is why I think tackling some of these are probably multi-year projects- more than I can cram into just the next two years, probably. Knowing myself, I will probably be making a stab at them anyway ;)