snarxiv is an imitation of arxiv produced by algorithms similar to those behind the famous Postmodernism Server. I always thought that the essays produced by the Server made rather more sense than anyone acknowledged. The individual statements were rather bland, but considered in whole paragraphs, there was often enough content that one could try to extract an overall sense.
Such an observation could be fodder for the culture wars (now rather passe) over whether postmodern theory is nonsense or just too subtle for its critics. However, here I'm concerned with machine-generated physics ideas rather than machine-generated postmodernism. I read the arxiv daily for inspiration, and I am alarmed that I could be drawing similar inspiration from the output of snarxiv. Here, after all, we are talking about mathematical physics, a subject governed by logical deduction, not by humanistic association.
There are several possible explanations. For example, it could be that all the concepts that typically feature in hep-th abstracts are so tightly linked that even a random collection of them will remind the informed reader of the deeper oneness. Or, it could be that my intellectual development as a physicist is still so primitive that I find it stimulating to see anything at all said about exciting topics like branes and broken symmetries; but if I had the higher-level understanding possessed by the true experts, I would see the superficiality of snarxiv's combinatorial nonsense.
My idea is not to choose one explanation over another, but simply to conduct an experiment. I visited the snarxiv just now and the abstracts looked, as they always do, plausibly like the real thing. So, now we are going to examine them in turn and see just how much sense can be extracted from them. (I will add, by the way, that when I first took the test, in which one is presented with a snarxiv abstract and an arxiv abstract and asked to tell which is which, I scored almost 100% after 20 questions. I can actually read and understand the professional physics literature. But I also have a weakness for entertaining odd ideas, even if they are coughed up unconsciously by a software program.)