Good question. I've been struggling with that a bit lately as well. I mean, I enjoy what I' doing and I think we're building some cool stuff, but when I look around and think about problems like "lack of clean water" and "prevalent infectious disease", and "hunger / lack of food" and "homelessness", etc. in various parts of the world, I sometimes feel a pang of doubt and sort of a "shouldn't I be doing something more important?" moment.
So how do I stay sane? Well... a couple of ways. For one, as a very pro-capitalist Libertarian type, I have a strong belief in entrepreneurship and small business as a way of improving the standard of living for people. And I believe that democratizing access to powerful software tools (via creating Open Source products) ultimately has second-order (or 3rd-order, whatever) effects of promoting business and economic growth. I look to the day when a small coffee farm in Africa or South America may use our products (for free, even) to help build their business.
But looking beyond even that, something dawned on me recently. When Douglas Engelbart passed away recently, all of the stories about his death prompted me to go back and read more of his writings (and some of his peers) and I sort of re-discovered something I used to think about a lot, but had forgotten: The idea of "intelligence augmentation"[1] and the implication that computers should serve to make us effectively smarter, both individually and collectively, and should enable us to come up with better (and faster) answers to hard problems... things like "lack of fresh water" and "lack of food", etc.
When I first started my professional programming career, I thought about things in those terms a lot (even if I didn't know the exact phrase "Intelligence Augmentation"), but I have drifted away from that mindset over the years. But now I think that was a mistake, and I'm trying to rediscover how to focus on building technology that has the effect of "making us smarter so we can solve hard problems".
So, in the case of Fogbeam Labs specifically, we were already going down the path of developing / researching / building / promoting stuff around the Semantic Web vision, we were just coming at it from a fairly simplistic "how can we make money by helping customers with this stuff?" mindset. Now I'm pushing to adopt more of a "will promoting this kind of technology make the world better for everybody, and, if not, what will?" mindset.
[1]: http://en.wikipedia.org/wiki/Intelligence_augmentation