First,ten BUT...

Seth Godin's latest post "First, ten" highlights a classic start-up mistake, but for me misses an important step.
He suggests that you find ten people to use your service who "trust you/respect you/need you/listen to you...", that, " if they love it, you win. If they love it, they'll each find you ten more people (or a hundred or a thousand or, perhaps, just three). Repeat."
Seth Godin makes a key point here which we've seen many start-ups, large and small, all miss: if you require that your business grows virally, then all that spend on launch and PR is wasted if you do so when the product is not yet good enough to be viral by itself. That's not to say that PR etc. can't be great when you want to pour some petrol on that fire, but you're better saving up that fuel for when the fire is already spreading by itself.
However, I'd have to suggest some qualifiers. The test is not those "first ten", but whether those first ten start to "spread the virus" by themselves. That might well mean that they need to be considerably more than ten so that you can measure and understand them, but it certainly doesn't have to be ten thousand. Key to that viral spread is the communication from an existing user to a new one, and that's a tough barrier because the user chooses it, not you (how many different ways have you heard Twitter described?). My view is that Seth Godin's definition of the "first ten" might prove a little unrepresentative- they're clearly going to listen, and therefore give you the chance to communicate much more sophisticated and subtle ideas, you can "teach" new things, or even "un-teach" perceived wisdom. You could also get those first ten to think about what the service could mean to their lives and how they might use it, and it will be much more memorable. So I would suggest that the "first ten" test is necessary but not sufficient.
My suggestion, based on our portfolio, would be to aim for a larger group of people, grown gradually, and iterate the product, messaging and experience to the point that you start to see that demand is spreading "all by itself". Then go and get the petrol can!