Well, Betteridge's Law of Headlines and all that. But don't confuse "wiki" with "wikipedia". Having reviewed scholarly journal entries published in a form where they are accessible to all, and all references are hotlinks, could only improve things. Some sort of discussion/comments associated with each article for Q&A, and forward links to all citing works would be great as well, especially works that refute the article in part or in whole.
But this is pretty much exactly what we have at the moment. Most journals will let you read papers online in this way, and provide a list of citing articles and hyperlinks to citations. Most journals accept comments, BMJ even has these as online comments. Look at PubReader for other innovations in this area. I don't think anybody has a serious complaint that academic research is organised badly, its just the cost issue that winds people up.
Academic research at the coal face is necessarily sprawling. The line in TFA that is telling is
If you’re an established researcher interested in summarizing an area of your expertise, or if you would like to write an article in collaboration with someone who is, we’d love to see you propose an article.
Which shows a fundemental misunderstanding of what research writing is. The author is asking for encylopedia or textbook articles, for which there are already plenty of outlets (these are called 'encyclopedias' or 'textbooks'). So fine if they are proposing a new encyclopedia, though god knows why we need a new one, and I'm don't see how one could or should ever become definitive. To suggest that this will replace any part of existing scientific writing is a bit misguided. Two scientists can produce reviews on the same subject given the same source material with vastly different conclusions, its important that all voices can be heard, and 'curation by a community of experts' seems like the antithesis of this.