3 ms·

the author is rationalizing their choice. nonsense. > Access to journals is prohibitively expensive and therefore practically unavailable to independent resear

the author is rationalizing their choice. nonsense. > Access to journals is prohibitively expensive and therefore practically unavailable to independent researchers ... and > Apart from the financial walls that academic publishers encourage, there are frequently de facto barriers to participation as a contributor, ... the Cornell Arxiv approach to free publishing is already spreading into other fields like Bio Arxiv. The substack culture isn't actually free. > Finally, the review process itself may be subject to intellectual protectionism and even unintentional gatekeeping that prevents research from reaching a broader audience that can help the ideas grow ... Without review process the volume of papers that get produced and published freely online will be so much that people will feel the pain. This is already happening in ML community because everyone dumps a slightly different but not so diff paper on Arxiv.

newswasboring 6y ago
> Without review process the volume of papers that get produced and published freely online will be so much that people will feel the pain. This is already happening in ML community because everyone dumps a slightly different but not so diff paper on Arxiv. I wish someone would make a technological solution to this. One way is an aggregator of scores which weights exponentially big for known subject experts. This is basically review but all experts are allowed to weigh in, not just the chosen reviewers which are either random or sometimes even biased. Another far fetched dream solution would be to come up algorithm to review the paper, or at least find if its similar to some other paper. The way papers are currently written it would have to be some sort of natural language processing AI, something like GPT-3 maybe. Of course what I describe is basically youtube for science and there can be a problem because of that. These days creators cater their videos to the algorithm rather than the audience. Since there isn't really a way that the algorithm can capture actual audience demand nuances, videos are being made predominantly for lowest common denominator of a genre. This can manifest very poorly in this papers system and then it would matter so much more.

newswasboring 6y ago

> Without review process the volume of papers that get produced and published freely online will be so much that people will feel the pain. This is already happening in ML community because everyone dumps a slightly different but not so diff paper on Arxiv. I wish someone would make a technological solution to this. One way is an aggregator of scores which weights exponentially big for known subject experts. This is basically review but all experts are allowed to weigh in, not just the chosen reviewers which are either random or sometimes even biased. Another far fetched dream solution would be to come up algorithm to review the paper, or at least find if its similar to some other paper. The way papers are currently written it would have to be some sort of natural language processing AI, something like GPT-3 maybe. Of course what I describe is basically youtube for science and there can be a problem because of that. These days creators cater their videos to the algorithm rather than the audience. Since there isn't really a way that the algorithm can capture actual audience demand nuances, videos are being made predominantly for lowest common denominator of a genre. This can manifest very poorly in this papers system and then it would matter so much more.