Is another difference that dhouston is building a huge corporation (with many employees etc), while you're building something you can run yourself? (Not a criticism - it's my aim too.)

No, I don't want Tarsnap to be something I run by myself. I want Tarsnap to be something which runs itself. I just spent two weeks at conferences and didn't log into any of the Tarsnap servers even once -- and when I spent 36 hours travelling from one to the other (Ottawa to Brisbane -- 23 hours in the air!) I was worried about whether my suitcase would make all the connections, not whether Tarsnap would break while I lacking internet connectivity.

But getting to what I think you were really trying to ask: Yes, Tarsnap is much smaller than Dropbox, and I'm happy that I haven't taken money from VCs (or any investors for that matter) who would push for faster growth. I'd rather have better product than more product; I will probably hire other people to help with Tarsnap at some point, but the question I'll have to answer is not "can this person do useful work" or even "can this person do task X better than I'm currently doing it"; rather, the question will be "can this person do task X sufficiently better to overcome the cost of my no longer understanding it".

Tarsnap is first and foremost about security. Security is about getting details right. And getting details right... well, that requires a level of understanding of how all the different pieces fit together which simply wouldn't be possible if I were hiring dozens of people and throwing them into teams to churn out new features every week.

Damn, you've caught my flair for understatement!