I had a conversation with colleagues literally last Friday in which I mentioned a version of question #2. I've always wanted to ask that question to candidates, but felt they would struggle to think of a way to phrase an answer in an interview setting...
YC has come up with a very concise prompt, though, so I'll have to give this a try next time I'm out recruiting...
Second question is very delicate. You definitely don't want to discuss something which may end up being unattainable. Say, you're working on a new physics, and you don't have something to show in form of math theories or experiments. Chances are you'll be labelled a nut, which from a nerd community could be counterproductive (see "ought to work on a good ides which looks like a bad one"). So you want to germinate your ideas - and perhaps throw those not good enough - until you have a sort of MVP. Which defeats the idea of the second questions substantially.
If you have something to show, it's more in venue of incremental, predictable work, and less in "out there" ideas perhaps.
YC has come up with a very concise prompt, though, so I'll have to give this a try next time I'm out recruiting...