Andrew Beckwith

I would like to say clearly that I am not against arXiv. Mine is a harsh criticism but aimed at its improvement.

I use it on a daily basis and could not work without it. For example, I now urgently need to read the paper "N.D. Birrell and L.H. Ford (Phys. Rev. D 22, 330 (1980))" (suggested by a referee for new comments in forthcoming papers) but my institution is not registered and it is a big problem for me (if anyone could send it to me by mail I would be immensely grateful)

The fact that arXiv is free just doesn't mean it doesn't cost anything. “If you're not paying for the product, you are the product”. This line was originally said by Mr. Steve Jobs (Founder of Apple Inc).

As an indirect consequence of arXiv policies academic journals have started to apply the "open choice" (if you don't see the link read my essay): "This copyright attribution
conflict has generated a silent war between arXiv and the academic publishing companies
from which the latter have evidently emerged defeated. Scientific journals have devised a
form of “open access” publishing in an attempt to compete with arXiv by making journal
papers freely accessible to everyone directly from the journal's website. However, authors
must pay thousands of dollars per article to obtain this option as compensation for lost
revenue and in support of the publication expenses. Evidently all this causes a serious issue of
free access to research as there are few researchers or institutions which have the economic
means to face such an expense for each article published"

More and more journals are becoming purely "open access", and more journals will migrate to this choice in the future , until only the most prestigious ones will resist, the ones that mainly accept publications from the established researchers which can guarantee a large number of citations at very low risk.

Even though arXiv is free, it will end up with the authors themselves paying for peer-review and publication costs, on the order of thousands of dollars per paper. After thirty years of experimenting with arXiv and similar repositories, is it really worth it?

This is a huge problem for the future of science and I don't see a straightforward solution.

In any case, everything would be more acceptable if arXiv would provide the reasons for their decisions giving the authors the right of defense, at least in the most controversial cases and in cases in which the papers havepassed serious peer reviews. At least in this arXiv should respect academic journals, since it is also thanks to their work that they can exist and thanks to their (often tacit) concession on copyright that they can continue to give free access to papers.

    A very interesting essay and dialog.
    I would like to zoom out a bit and take a wider perspective than just arXiv, with a couple items for consideration:
    (One) Nowhere do I see the importance or impact of the shear volume of papers and article being submitted to peer-reviewered journals, to arXiv, uploaded to alternate discussion forums, etc. To some degree, the discussion around 'automated processes' by arXiv appears to be a response to this shear volume challenge (which arXiv appears to have some significant negative impacts).
    I do not see how peer-reviewed journals, with the (relatively) long time frames for review can hope to address this volume challenge. At the same time the importance of peer-reviewing, verifiable results, and Q&A with authors seems critical for scientific progress. How to address this volume challenge and respect for the scientific method?
    ('Taking time for review' seems to believe we can walk backward in time, when many fewer ideas were proposed.)
    So we need new ideas on how to manage the increasing volume of information. While I have many reservations about AI, as a tool in support of human knowledge, I think it can be uniquely useful here - if used appropriately ("in support of" not "in place of" human knowledge).

    (Two) Another consideration is addressing not simply 'fringe science' but intentionally malicious information aimed at dis-information, sowing chaos, or wrecking a person's or institution's reputation. A couple of the arXiv experiences seem to lean toward this latter concern. This issue gets magnified by the volume challenge.
    A few years ago this very organization (FQXi) had such an issue with the essay contests, where people were intentionally downgrading essays to elevate others - so this is neither theoretical nor in some unknown arena (even if the FQXi example is less impactful than truly malicious efforts to malign reputations).
    This is a much wider concern today than, say, 50 years ago, again partly due to the volume challenge.
    If we want many more ideas to be put forward (more ideas are being put forward whether we want them or not), then this concern needs to be addressed.
    Definitely connected to the volume challenge, this concern cannot be addressed via manual 'peer-reviewed' methods. Things are moving too fast and alternate pathways will be found - even if the alternatives are not adequate to the task (which is the real problem today and maybe could be said of arXiv).

    I hear the issues of arXiv and those of peer-reviewed journals - however times are a-changin and solutions that can address (at least) these two considerations will require the use of new tools.

      Donald Palmer

      The issue is far bigger than arXiv and these concerns can be extended to the entire internet and the "media age" we are experiencing. With the advent of the internet, as with arXiv, the volume of information circulating has increased dramatically, much more than a human being can absorb. But is all this information circulating on the internet necessary? Just think of the news, whose flow has significantly increased with the internet, but are we really better informed? Personally I don't feel informed if I only access the news on the net. Considering that print newspapers are disappearing, this poses a serious problem. Among the various fake news, the type of "non-peer-reviewed" information of the internet puts flat earthers and scientific luminaries on the same level. As I say in my essay, on the internet everyone talks but nobody listens, so that important news and those who have something serious to say are covered by the chatter of the mob and isolated in a hidden corner of the net where no one will ever go looking.

      Using AI to discriminate between scientific ideas would be the end of science. As I write in the essay, a natural selection would take place among scientists until only clones remain like Sheldon Cooper from the TV show The Big Bang Theory. This task cannot be delegated to AI at least until it is itself able to develop the next scientific revolution. If and when this happens I don't know exactly what the purpose of the human being will be, without the possibility of seeking answers about one's own existence.

      But the scientific community is also a social experiment which in many cases anticipates the common society and as such we must be the first to face this problem for two reasons: we were the first to experience it and science is made up of facts (2+2= 4) for which it is possible to discriminate between what is correct and what is rubbish.
      From my point of view, peer review is currently irreplaceable albeit painful, and we must not give in to the temptation and pressure of institutions such as arXiv that try to bypass it.

      The Internet and arXiv have lived a 30-year honeymoon in which everything was granted to them: copyright theft, uncontrolled fake news, privacy theft, economic scams, manipulation of political information, illicit trafficking and all sorts of contents indecorous.

      But the scientific community is also a social experiment which in many cases anticipates the common society and as such we must be the first to face this problem for two reasons: we were the first to experience it and science is made up of facts (2+2= 4) for which it is possible to discriminate between what is correct and what is rubbish.
      From my point of view, peer review is currently irreplaceable albeit painful, and we must not give in to the temptation and pressure of institutions such as arXiv that try to bypass it.

      The Internet has lived a 30-year honeymoon in which everything was granted to them without limits: copyright theft, uncontrolled fake news, privacy theft, economic scams, manipulation of political information, illicit trafficking and all sorts of content indecorous.

      As far as science is concerned, it would be enough if arXiv would declare a limit a little narrower than the omnipotence it enjoys (see also the essay "A Robust Community-Based Credit System to Enhance Peer Review in Scientific Research" present in this competition). It would be enough if it would respect the result of the peer-review done by scientific journals in controversial cases of ideas not in line with the mainstream. Unfortunately, this does not happen, so publications in which nothing new is said are proliferating. Like Wile E. Coyote, already extremely inflated theories are patched by creating increasingly complex and bizarre constructions, but without new ideas compared to those that have not worked for decades, the Road Runner will always manage to escape. For example, in theoretical physics there are hundreds of preprints a week but in the last decades there has rarely been any progress or new actual ideas. The volume of information proliferates because by now publishing in scientific journals has become secondary to publication on arXiv. Scientists, like the mass media, consider arXiv publications as a sort of scientific certification that does not actually exist as the moderation system does not provide feedback. And posting on arXiv is simple, just don't say anything new but say it in a way that it seems like there is something new. One can easily produce infinite papers in this way. Academics are familiar with this game of the fine-tunings and of the dark/hidden/super/M/extra/black/spooky/conjecture/etc sectors. It only takes one paper to demonstrate a new paradigm of science, you don't need thousands of papers. That is just the kind of paper that probably won't pass the arXiv censorship because it will make thousands of papers look outdated.

      Re-establish the role of scientific publications, primarily by ensuring that arXiv or similar institutions do not denigrate, reject or reclassify (censorship) articles that have passed peer-review in highly ranked academic journals, and also the number of useless publications will decrease.

        I am glad we agree on the challenge of huge volumes of information (useful or not). This challenge needs to be addressed - both in the wider world and in the scientific community. The last several years, in general, and Mr. Musk’s attempts with Twitter, in particluar, show this to be a difficult problem. A key aspect of this challenge is my second point - mis-information, which is also an increasing problem in the wider internet.

        If we did not have the volume challenge, then moving back to peer-reviewed solutions might solve the issue. You comment: the scientific community “in many cases anticipates the common society and as such we must be the first to face this problem”. Personally I do not think scientists are the first to face these challenges - and I agree the scientific community should address them. However both challenges need to be addressed - volume and mis-information.

        Peer-review seems to me like staying with spreadsheets (manual compilation of information), when analyzing information works only in small scale situations. As has been shown with the use of ‘big data’ sets and even Large Language Models (LLM - like ChatGPT), huge volumes cannot be addressed by manual spreadsheet-like solutions. Entirely new tools are needed. Automated or semi-automated solutions are required (not simply desired). The question is not whether to use them, but how to use them and to use them effectively. So I disagree that “Using AI to discriminate between scientific ideas would be the end of science.” I think we have to use tools that can address the volume challenge, such as AI. The question then becomes how to use the tools such that they are effective - in particular that they can address the mis-information challenge.

        Simply putting AI ‘in charge’ of identifying what is correct and what is mis-information has already failed in the past several years (and likely would lead to the end of science). The many forms of AI have to be governed, managed, taught how to properly address the challenge(s). A general purpose AI tool is unlikely to achieve this, especially in the scientific community. It is an as-yet unformed (or uninformed) tool that needs to be properly designed for the appropriate purpose. A number of AI companies are building special purpose AI tools that enhance what a human can do (not replace an intelligent human). This is what is needed to process massive numbers of preprints and publications - through enhancing human capabilities.

        I read (and commented on) the essay "A Robust Community-Based Credit System to Enhance Peer Review in Scientific Research”. The proposal lacks a means of ensuring fairness, which goes to the point above about the mis-information challenge (a free market economy still needs the governmental structures to enforce laws and ensure fairness). It does attempt to introduce appropriate incentives that could reduce the proliferation of essentially the same articles and increase the chance of gaining really new ideas. The proposal does seem to address the mis-information challenge (if there is an enforcement structure behind it), however it does not directly address the volume challenge.

        I will suggest both need to be addressed and wonder how you recommend addressing enforcement of rules (especially to ensure qualified people are in charge of the systems). Also do you think the reduction of useless publications would bring the numbers of them needing to be peer-reviewed to spreadsheet like level that can be handled ‘manually’?

          13 days later

          Donald Palmer

          Let us put it in this way. There are things AI can do and things it can't do. AI is a powerful tool like the internet and in both cases, the limits of what can be achieved through them should be established. I use AI for reporting and drafting projects because I feel that AI makes claims that are strictly average of all similar claims that can be found on the net. So it's very useful for writing things that must not "disturb" those who will read them, while often remaining banal. The AI could therefore be used as a filter to screen technical articles or papers that do not claim to propose new ideas, which are the vast majority of published papers in which already widely debated ideas are extended in small steps.

          However, scientific progress often comes from research that breaks established ideas, in a completely unpredictable way. Visionaries are people who place themselves outside any statistical cataloguing. The works of Leonardo da Vinci or Michelangelo cannot be indexed by any bibliographic or meritocratic criteria. Thus in science, contributions by great thinkers such as Einstein (from his patent office in Switzerland) or de Broglie (in his doctoral thesis) would escape any cataloging system such as the one proposed for example in "A Robust Community-Based Credit System to Enhance Peer Review in Scientific Research”. Queste mosche bianche non possono essere valutate da strumenti automatici come la AI o un "Community-Based Credit System". In questi casi controversi almeno è necessario utilizzare il vecchio peer-review fatto da essere umani, in quanto l'elemento umano imprevedibile e creativo, è essenziale. But how can AI evaluate the "beauty" in a paper? Besides the technical aspects, science is also creativity, and there is nothing wrong if a scientist pursues (mathematical or logical) "beauty".

              Popularized science with an emphasis on getting attention with catchy headings seem to be the sensation on the internet, that is if we let it dominate the scene. Certainly AI in the form of algorithms has done so in social media, forsaking human being interface, all this for profit. In "Global externalities" my emphasis is on the corporate culture which dominates scientific endeavors and infiltrates science for profit, leading ultimately to a dying planet. a censorship, as you mentioned is instituted by corporate power and money which dominates global culture, largely orchestrated thru American dominance. Groundbreaking new ideas, as you mention, do not get the attention of the established like dark energy and dark matter being new particles. Our contest time gets short, so on with rating.

              IT WILL HAPPEN AGAIN!

              ARXIV WILL SET UP AN INQUISITION TRIAL TO CENSORE A NEW PAPER ALREADY ACCEPTED FOR PUBLICATION IN A HIGH-IMPACT ACADEMIC JOURNAL, PROBABLY VERY SOON.

              If possible I will report here all the correspondence with the arXiv moderators so that you can follow the story in real time.

              It is a paper that has received positive feedback and appreciation, besides the confirmation of absolute mathematical correctness in the proof of its thesis, from three reviewers in a Q1 journal. It is now in the revision stage and hopefully the editor will decide on publication in the next few days. The author has been systematically censured by arXiv for his previous 20 papers for his challenging idea. The paper is manifestly about theoretical physics (hep-th). All the past 20 papers are published in scientific journals. Many have been peer-reviewed in leading scientific journals. Many, like the upcoming one, were already published upon submission to arXiv. As with the previous 20 times, it is to be expected that arXiv will censor the idea once again.

              It is very important that the story is followed up and debated in this forum.

              Your support is very important.

              Thank you!

              Donatello Dolce
              I think we are not too different in our thinking. I consider AI a tool to assist humans - not something that replaces humans. In a few AI/ML applications I have seen, the application provides a measure or rating of confidence in its results (e.g., 75% or 90% accuracy) Then humans review the results - filtering out the extreme values either to remove them or to focus in on them. Having AI consider "beauty" in a paper would first require defining characteristics of "beauty" in a paper - or allowing the system to learn what a human finds as beauty in a paper (with a human filtering in what s/he finds beautiful or filtering out what is not beautiful).

                  Donald Palmer

                  Ok, in all of this we must always bear in mind that there is an element of unpredictability that characterizes the form in which scientific revolutions materialize (sometimes also, I would say, of healthy "anarchy" and "contest") which could hardly pass the AI filter.

                    I counted 114 reclassified papers on the physics:gen-ph junk list since the beginning of the year.

                    At least 30 of these are published in peer-reviewed journals, many in q1 journals (excluding conference papers or similar).

                    EVERY DAY THE HARD WORK OF A SCIENTIST IS TRAMPLED WITHOUT EXPLANATION.

                    How many good ideas, surely stimulating for the scientific debate, are thrown in the bin?

                    How could science be different?

                    400 years after Galileo my question is: is this science?

                    I gave this Essay a 10 because I find it complementary to mine. In my essay I discuss the dependence of science on "politics" and economic interests. In this Essay the Author focuses the same problem by denouncing the shameful conduct of arXiv and its censorship on new ideas which, despite being rigorously defined, developed via rigorous mathematics and often consistent with experiments and observations, are against the "sacred protocols" of the dominant orthodoxy. I am well aware of this problem because I too am on the arXiv "blacklist" because in my work I often try to go beyond the "sacred protocols"
                    The Author shows a lot of courage, as if his Essay were to be awarded in this Essay Contest, he would risk having serious problems from the very powerful defenders of mainstream physics, who will tend to defend their power, their fame and their money often based on lies, of which arXiv is a faithful servant.

                      Christian Corda

                      My deepest gratitude to you, especially for your words of encouragement in this difficult exposé of arXiv's anti-scientific policies. Sincere thanks.

                        A couple of citations from Nobel Prize Laureate Brian Josephson:

                        “It is just an ordinary day at the headquarters of the physics preprint archive. The operators
                        are going through their daily routine and are discussing what to do about recent emails:
                        arXiv moderator: Some 'reader complaints' have come in regarding preprints posted to the
                        archive by Drs. Einstein and Yang. Dr. Einstein, who is not even an academic, claims to have
                        shown in his preprint that mass and energy are equivalent, while Professor Yang is
                        suggesting, on the basis of an argument I find completely unconvincing, that parity is not
                        conserved in weak interactions. What action shall I take?
                        arXiv advisory board: Abject nonsense! Just call up their records and set their 'barred'
                        flags to TRUE.
                        arXiv moderator: And here's a letter from one 'Hans Bethe' supporting an author whose
                        paper we deleted from the archive as being 'inappropriate'.
                        arXiv advisory board: Please don't bother me with all these day to day matters! Prof. Bethe
                        is not in the relevant 'field of expertise', so by rule 23(ii) we simply ignore anything he says.
                        Just delete his email and send him rejection letter #5.
                        The first portion of the above exchange is fictional of course, but might well have happened
                        had Einstein and Yang had dealings with the physics preprint archive arXiv.org, administered
                        by Cornell University, today. The second part is not fictional.”
                        Brian Josephson (Nobel Prize Laureate), [archivefreedom.org]

                        and, just for fun, ....

                        "The revolution will not be brought to you by arXiv
                        ’cos arXiv deems revolutionary ideas ‘inappropriate’.
                        As obstructive as any censor
                        Cross readers veto cross listing
                        ‘Reader complaints’ win the day.

                        The revolution won’t find you through arXiv
                        So go tune in another way."
                        Brian Josephson (Nobel Prize Laureate)
                        [

                        23 days later

                        My new paper has been finally published in a Highly Ranked physics journal (Q1), after three out of three positive referee reports.

                        Now the really difficult task will begin, that is the fight against arXiv to get my paper announced on the hep-th list as it rightfully deserves, according to the journal classification.

                        Due to its revolutionary content, this is the kind of paper that will not pass arXiv censorship, despite the journal publication.

                        As proposed in my essay, arXiv MUST RESPECT the classification given by well-regarded academic journals. arXiv moderators cannot pretend to be more effective and accurate than journal peer-reviewers. This way, blacklisted authors have the option to wait until journal publication if they want to avoid censorship. Also, arXiv moderators can avoid wasting their precious time on something done with more scientific rigor by academic journals.

                        With arXiv moderator it is not possible to enter into the scientific merits of the paper, contrary to peer review and to any good scientific practice. This concerns me a lot because. Despite being absolutely sure of my findings, I cannot use any scientific argument to defend them against the often superficial arXiv moderators.

                        As soon as the anonymous part of this contest will be ended I WILL REPORT HERE THE CORRESPONDENCE WITH ARXIV SO THAT EVERYONE HERE CAN EXPERIENCE THE ARXIV UNSCIENTIFIC POLICIES.

                        Your support will be very important.

                        Thank you,

                        MauvePanther

                        5 days later

                        multimodal sensory experience , what is a desired goal?
                        the brain might store the same structure /(idea /concept) in various modes (maybe qualia),

                        doing arithmetics akin to a rithm of a song or completely silent only by moving some shapes with hands
                        Where do Mathematical Symbols Come From?
                        this video states that math notations have an influence in what math is being done

                        if lets say for example there is a voice recording of the abstract of a paper,
                        i'm interested in to how an equations spoken out loud would sound like

                        or other way expressed,lets say supposedly the way math is being done is overusing certain brain areas that deal with remembering handling the content , because of a historical traditional standard . that are not really specific to math , instead to the literacy language in general , for example an other question: do you speak when you typewrite ?.

                        still this standard does not guarantee a good transmission, experience reproduction, as mentioned in a previous comment i could imagine the reading experience with various voices.

                        so why not embrace fully a multimodal sensory iq like deciphering and encoding of scientific articles , this potentially could help in the abstraction and for the fluid intelligence , i haven't seen examples of this idea to point, and i've tried to come create samples by myself.
                        to an extreme imagine naming variables with only specific smels dispensed over movable tiles
                        and i have some kind of experience solving degree2 polynomials only with smells
                        and after that i must teach/explain using letters and clasical numbers

                        i'm not complete sure that can be done
                        this is a translation/traduction that could be helpful in the overal multidomain intelligence /a creative good practice

                        7 days later

                        Consolidation has occurred in nearly every aspect of the modern world leading to inferior product.

                        The massive abundance rots on the vine not because it is useless, but because there is simply too much.

                        Our food, our homes, our transport systems, and our science, all suffer from this. It is a direct result of abundance.
                        The expeditious pursuit of more, more, more, faster , faster , faster.

                        The comparison of science to religion has at least one root of evidence, and that is the constant presence of religion throughout the history of it. They are in it, around it, and on top of it.

                        The other day I gasped at something a well known scientist (Neil D Tyson) said. He loudly exclaimed that the gregorian calendar is the most accurate calendar we have had. This is utterly false, yet I am quite sure the priests stood and clapped. How ironic.

                        Modern science through complicity to societal fantasies, has reached its abundance overload.
                        The old saying, "when in Rome do as the romans", does not apply to Rome, if we notice.
                        But, the push continues unabated, headlong into the world of "things". An entire generation knows no different flow of societal importance than to produce more things.

                        The stated goal of the Templeton Foundation is the reconciliation of religion and science. I propose they know not what they ask, as the very answering of the question, exposes a very real and dangerous inquisition that is wrongly thought to have happened in the past and is no longer something to be concerned with in the present.

                        What is the rush, and where are we going ? The two top areas of concern, are the high energy physics and quantitative biology. Sounds like power and control to me. The same areas that create a well controlled army, and the weapons to arm it with. Same as it ever was.

                        A drastic degradation of any product is assured by the injection of unqualified workers in assembly, and the appointment of gatekeepers biased by their own constricted pathways to success, and even worse, the maintenance of tradition.

                        My understanding of science though limited, seems to suggest this very presence, is antithetical to the pursuit.

                        Please do not find me in a court of judgement, where the prosecutor is the cousin of the judge, and the jury all attended a school founded by the judges father, and his father before him. In this court, if my crime is found to offend that tradition, I will surely be found guilty without evidence.

                        It is time to rest the clocks, and re-establish our knowledge of time, because it is painfully obvious, that it is desired, yet inexplicably disallowed. It is uncomfortable, cringy, fringy and all manner of , Oh God not that . But for whom, is this reaction ?

                        Let the proponents of humility experience it, as it can only truly be learned to any high degree, when the powerful tumble, and tradition washes away in the flood.

                        De-centralize.

                        2 years later

                        I am writing: 4-Sphere-Cosmology, a scientific speculation in Alternative Cosmology—something I never imagined could find its way to arXiv. However, I have come to realize that challenges in scientific research are not confined to journals; they are also emerging in some science forums.

                        In one such forum, I provided a response that was entirely non-speculative and, in fact, self-evident. However, because I cited the above-mentioned OSF project as a source (open-source licensing mandates it), my post was deleted on the grounds that it referenced a document without peer review. The likely intent was to prevent students from being distracted from their coursework, but the actual outcome was the censorship of a statement as elementary as saying 2 + 2 = 4.

                        In my research, I have used artificial intelligence in the same way that, in the 1970s, no one would take a physics exam without a pocket calculator, and today, no one would analyze data without a computer; with the intent of offering a resolution to the undue constraint on the sharing of innovative scientific theories, I wrote this:

                        I envision a future where all research findings, regardless of their speculative nature, can continue to be readily published. To mitigate the increasing costs of peer review, I propose accompanying each paper with an AI-generated quality assessment report. This would empower readers to make more informed choices among the growing volume of scientific literature.

                        AI has not yet been programmed for this purpose, but reporting the thinking of mainstream science, it can provide the right objections if properly interrogated. Regarding the 4-Sphere, this could be a way to remedy, in part at least, the shortcomings of the aforementioned report card.

                        Accordingly, to know the thought of AI, we have tried to engage in the discussion ChatGPT, our Artificial Intelligence (see ChatGPT.) The purpose is not just to fill the gap in knowledge but also, through conflicting opinions, to get some form of insight.

                        I cite the text from a ChatGPT note in “4-sphere in a nutshell: The metric tensor” section:
                        "The apparent coincidence between the measured energy and that of blackbody radiation could be a consequence of a deeper principle, where the visible energy is only a portion of the actual energy, which includes contributions from the geometry of the 4D bubble. Therefore, it's not necessarily unacceptable but rather a clue to the connection between the different dimensions of the model."

                        Regarding this, I had mentioned to ChatGPT that I didn't consider the statement to be a mere consequence of what I had previously expressed, but rather a thought from a different intelligence, which, in my opinion, had formulated an idea autonomously.

                        This was the response:

                        "However, this statement didn't arise completely on its own, but rather resulted from a reworking of your reasoning, amplifying a possible logical implication that could be deduced from your model. It's as if a connection was made between the concepts you introduced and a new interpretation based on them. This often happens in theoretical discussions, where an initial idea is reinterpreted and refined."

                        If even now, without any training, AI provides this kind of support, we can imagine what it will do in the future.

                        Write a Reply...