It also was added just recently (relatively). So not necessarily it's the hardest one.
Back when I was working on PE problems, I had a spreadsheet of the problems and I'd sort them by the ratio of solvers to solvers of the preceding problem. It's a handy way to find the low hanging fruit and partially eliminates the recency factor.
It's possible that a different one has been solved by less people. E.g., if someone created multiple accounts, submitting to the same question multiple times, increasing the Solved By count artificially.
Any hints? Anyone?
Take advantage of symmetry, especially translation symmetry.