I agreed with much of this article, but not that line. A lot more people have heard of YCombinator since a year ago, so all else equal, you'd expect applications to rise over that period. So comparing the number of apps with the corresponding value from a year ago isn't a fair comparison.

Indeed, looking at http://siteanalytics.compete.com/ycombinator.com/?metric=uv we see that the number of unique visitors to ycombinator.com is up 3x from a year ago. But applications are only up 40%. To be fair, we probably can't expect applications to grow 3x since many of those visitors are international, or otherwise unlikely to apply to YC. But to make the statement pg wants to make, I think we need a better metric than raw growth in applications.