That is a type of question we think about a lot. But we know that our guesses are likely to be pretty bogus. We approach scaling YC the way you would approach

That is a type of question we think about a lot. But we know that our guesses are likely to be pretty bogus. We approach scaling YC the way you would approach scaling software. And the way scaling works is that you can never be sure what the bottlenecks are, or how you're going to get around them, till you hit them. So it's interesting to speculate about, but not useful except in a mind-opening way.