Why would I want to do that? Remember, I'm just trying to make timelock encryption practical, which means we want a good estimate on what's the fastest implementation of the underlying chain kernel that exists now and will in the future. I really don't care who collects the bounty; if it's the same small group of people over and over again, so what? I do care that the best implementation possible - probably a highly clocked custom ASIC with exotic cooling - isn't that much faster than the best implementation possible without too much capital investment - probably a high-end overclocked FPGA with exotic cooling built by some kid in university as a term project.