The problem is that there is a lot of incentive to publish peer reviewed articles and very little incentive to publish articles confirming another paper. In my experience, the reproducibility of papers is tested in the worst possible way: during the process of more novel research.
If we are to continue this exponential growth in technological and scientific progress we need to set up a far more efficient way of reproducing the papers, independently of ongoing novel research. This seems like a great way, especially considering the benefit it will give to startups and other companies that can allow new R&D to rely on papers that have been tested outside of the peer review/slow meta-review process in science.
I think this kind of thing is probably a well-intentioned pipe dream. You can't predict which work is going to be important, so you can't predict, a priori, which work to verify. So you either have to verify all of it, or you have to guess, or...you can just let people do what they do right now: verify the work they care about.
You can't predict which work would be important in general, yes, but when it comes to identifying important protocols in organic chemistry and biology, there is a lot more information on what is important and what is not. Most of the time, in o. chem, protocols are incremental improvements over previous protocols and it's very easy to figure out which of them are most often used.
That's not to say that these sorts of things aren't useful exercises, just that it's pretty wishful thinking to argue that this sort of peer-replication is going to make any sort of dent in the system. Scientists are a naturally skeptical and conservative group of people, and one of the first lessons that gets drilled into you in graduate school is that you have to take individual papers with huge grains of salt.