The fact is, the arXiv is seen as a legitimate and respectable place for scientists to post preprints.
V.G. Gurzadyan is well-known in the physics community and wouldn't risk his reputation by posting papers with phony coauthors.
Peer review and reputation are nice, but the acid test of any scientific model is whether it works. For example, is the model consistent with known observations, and does it make good predictions?
Right, but the problem is that us non-experts can't tell whether something posted on the arXiv holds water or not. Being posted on the arXiv carries very little signal.
Don't get me wrong: I love the arXiv and use it regularly. Long papers can be posted there, get disseminated quickly, and can be downloaded freely. The problem I have is that many people think anything on the arXiv is automatically true -- but there's lots of cranks posting proofs of P ?= NP and etc.
Yes, the acid test is obviously whether or not it works, but that's why we need to wait for the experts to figure it out and vet it.
The world in general is a legitimate and respectable place for physicists to publish preprints, but you still shouldn't believe everything you hear in the world. I think that's the point. It's not that arxiv sucks, but that it doesn't lend any credibility to the things published there, so we need to be skeptical. For those of us who are not the author's peers, taking non-peer reviewed work as anything more than speculation is likely to lead to false beliefs.
V.G. Gurzadyan is well-known in the physics community and wouldn't risk his reputation by posting papers with phony coauthors.
Peer review and reputation are nice, but the acid test of any scientific model is whether it works. For example, is the model consistent with known observations, and does it make good predictions?