"I hacked this together with the worst code you've ever seen and the server is crashing every 5 minutes, but a thousand people signed up today and usage is growing exponentially."
Hmm I assumed the opposite. The mass exodus of long time users who are pissed about the API changes won't happen until there is a destination that can handle the migration without servers going down. It will be interesting to see what happens.
Focusing on tech and scalability before getting users will just result in a well-engineered ghost-town. People would rather use something slow and buggy if that's where the activity is.
It's much more of a marketing and PR challenge than a tech challenge. Having a site that works well would be nice too, sure, but finding a way to get tons of users signed up (and keeping them engaged) is far more important.
