The big surprise is that YC hasn't seen an order of magnitude increase in applicants from their first round. You'd think this impossible: Paul announced the first YC program only about a week before its deadline and they got something like 250 applications. Now they are all over the tech news and they still haven't cracked 1000? Maybe PG's new speaking approach (memorize, practice, use visuals) is an effort to get more people fired up?
Paul, regarding this new speaking style: you could make yourself into a Steve Jobs dynamo of a speaker, but I think you shouldn't. Your usual wordsmith style is probably better for attracting nerds.
Question: Paul, it sounds like you've mostly swung in the direction of encouraging people to look at their own lives for problems to solve. Whereas in "Why Smart People Have Bad Ideas" you were saying the opposite, like in the two paragraphs starting with "Reading the Wall Street Journal for a week should give anyone ideas... " Trevor says something similar in the footnotes. But it looks like YC has been placing all its bets on the first sort of ideas. There's a good reason: all the huge software successes come from the "build it for you and your friends" model because those are the problems you fully understand. But it's more common to find software companies making solid profits on products they built for others' problems. Am I overlooking YC companies building products designed for other people? Is the problem that there just aren't enough potential founders like the Octopart guys: people deeply involved in some non-computing world who also happen to be good programmers? If so, maybe the richest source of ideas and cofounders would be cruising web forums for the near-programmers (people who build things like MS Access half-solutions to their companies' problems) to listen to what they would like to build if only they knew how. Maybe YC should be recruiting there, too. Thoughts?