Bruce Schneier's blog points to a new site from Matthew Green that attempts to estimate how much cryptanalysis each public-key assumption has attracted. The project's abstract describes the goal plainly: estimate how much public cryptanalysis each assumption has received, using the literature itself as the basis.
Rather than surveying researchers or running experiments, the approach treats published papers as a record of effort. By counting or weighing the literature, the site builds a proxy for the number of cryptanalyst-hours invested in each algorithm. Schneier's brief write-up presents the effort as a useful way to think about why some algorithms are seen as more battle-tested.
The significance, at least as presented, is that confidence in public-key systems often rests on how long they have survived scrutiny. A quantified estimate of that scrutiny makes the comparison more concrete. Since this is a single source, there are no conflicting accounts to weigh; the details beyond the abstract and Schneier's framing are not described in the original post.