Yeah, I've seen this xkcd. It's on point as usual, but it doesn't show the frightening thing: it's the scientists who are going "Whoa" and believing in the results they've produced...

On the other hand, reproducing badly-designed, not-well thought of, and probably inadequately executed experiments won't be of much help to anything. In biology its not often clear what is known, what is not, and what are the open questions, and the scientists themselves have a lot of the blame for it (for religiously sticking to closed publishing models and a general lack of initiative to share their data).