PG was asked a question at the YC event that essentially boiled down to: "What are the best startups that will make use of Instant Personalization?" His answer

PG was asked a question at the YC event that essentially boiled down to: "What are the best startups that will make use of Instant Personalization?" His answer was that it was impossible to predict but whatever it was, it would be something that at first seemed wrong, ie something that wasn't how Facebook intended the API to be used, and something that they wouldn't be sure they should even allow. This isn't that sort of case: a previous version of Matt's app that used the API in the exact same way was thoroughly reviewed and approved. Nevertheless, one can easily imagine that the truly disruptive Facebook Instant Personalization Apps may only even have a chance of surviving the FB approval axe with the support of an organization like YC. This, in my opinion, is among the biggest reasons why the YC/Facebook deal will be so beneficial for YC companies, and why Jason Calacanis' criticism was mis-placed: the real risk is not that Facebook will steal your disruptive idea that is obviously good, the risk is that Facebook will just shut off your weird idea which isn't obviously good at all and only later turns out to be quite disruptive.