Avoidance of Doubt

Issue 30 · December 2020

Join the conversation on Twitter @Briefer Yet.

Did a colleague forward the The Brief to you? Sign up via the button below to get your very own copy.

The New Year Is (Finally) Here

As the year comes to an end, traditionally we celebrate — however we celebrate — and this year should be no exception. If you don’t celebrate, we hope you will take some pleasure in the fact that we are celebrating and intend to pass some of our good cheer and best wishes along to you. As an economist would say, happiness is a non-rival good, and everybody can share it with no diminishment of their own. So be well and look forward to the new year, which we hope will be happy and healthy. And prosperous, too: that still counts.
 
After the year we have had, celebration may seem a bit unwarranted. We are concerned that 2020 is going to put all the dystopian novelists out of business. And as bleak as this year has been (a large part of our celebration is wishing 2020 good riddance), there is another side. And that has been the spectacular performance of the scientific enterprise. Science has acquitted itself with honors this past year, rethinking processes and committing capital and labor to the development of a COVID-19 vaccine at a pace heretofore unthinkable. Some of the new discoveries and processes of the past year will remain with us, enabling new science and the better world that can grow from it. This is a cause for celebration.
 
All of us at C&E wish the members of our community a happy end to a difficult year and a wonderful new year. We hope to see you at a better time and in a better place.

Avoidance of Doubt

Sometimes when no one likes a policy, it means that the policy represents a shrewd compromise — while it may not be ideal for anyone, it balances competing priorities and interests. But other times when no one likes a policy it is because it is just a bad policy. The Plan S Rights Retention Strategy (RRS) looks increasingly to be among the latter category. 
 
As we noted in the November 2020 issue of The Brief (see Item 2), RRS is a scheme by Plan S in which authors of research funded by the coalition include an ultimatum in the cover letter of their submission to a journal (or in the acknowledgments to the paper) that asserts that the authors will retain the copyright to the author accepted manuscript (AAM). In addition, they (the authors) will release the AAM to a public repository with a CC BY license immediately concurrent with publication in the journal (a journal, of course, may simply decline the ultimatum and refuse to review the paper or refuse to review it under those terms). Here is the proposed language from the Wellcome Trust, a Plan S signatory:

This research was funded in whole, or in part, by the Wellcome Trust [Grant number]. For the purpose of Open Access, the author has applied a CC BY public copyright licence to any Author Accepted Manuscript version arising from this submission.

A group of notable physics societies has raised concerns about the policy, arguing in an open letter that RRS undermines the transition to OA for many hybrid journals, which today publish 85% of physics papers: 

The ability of these journals to transition sustainably [to full OA] is challenged by the prospect of free and unrestricted distribution of accepted manuscripts without concomitant funding for the peer review and publication costs involved. We are concerned that policies such as the proposed Coalition S Rights Retention Strategy would undermine the viability of high-quality hybrid journals and mean that many physics researchers no longer have an adequate range of options or freedom of choice in where they publish their work.

When physics societies are concerned about your green OA policy, something has gone very wrong, as the vast majority of physics papers are posted to arXiv prior to publication in a journal — which is to say that the physics community has long been comfortable with papers being openly accessible in advance of publication. 
 
It is not just physics societies that are concerned. A broad coalition of publishers has issued a statement about RRS on the OASPA blog. The coalition includes Hindawi, a Gold OA publisher, as well as Springer Nature, a publisher that has announced its intention to make all of its journals Plan S compliant. It also includes the Biochemical Society, Cambridge University Press, IOP Publishing, Oxford University Press, The Royal Society, The Royal Society of Chemistry, SAGE, Taylor & Francis, and Wiley.
 
The publishers make two arguments. The first is that by emphasizing the AAM, green OA “confuses the scholarly record with multiple inferior versions of manuscripts.” They write:

Do we want researchers to have to search through repositories for an earlier version of a manuscript, and then spend further valuable time seeking out accompanying data, or checking whether there have been post-publication corrections? Or would we rather that they have immediate access to the trusted and enhanced VOR [Version of Record] on the publisher platform, with links to relevant data and other outputs?

The second argument the publishers make is that RRS will slow the growth of Gold OA and of transformative agreements by providing a “good enough” means of compliance with funder mandates:

Green becomes “good enough” and compliant at the expense of proper OA to achieve open research more broadly. This “good enough” solution is even more attractive under the current budgetary constraints that the covid-19 pandemic has brought with it. Why spend your research grant money on the cow if you can get the milk for free?

RRS, in other words, threatens to undermine both hybrid and Gold OA journals, which is quite a feat. Through RRS, cOAlition S appears to be moving, whether intentionally or not, from a position of insisting on Gold OA and agreeing to pay to support it, to suggesting that researchers should publish OA without paying anything at all. Publishing is free on the Internet, don’t you know? 
 
But it is not just publishers that RRS fails to adequately take into consideration. Consider the researcher who, A. may not wish to distribute an unedited AAM to siphon traffic away from the VOR, B. is now placed in the awkward position of making demands (on behalf of Plan S funders) to journal editors and publishers, and C. may ultimately not be able to publish in the journal of choice. It is another hoop to jump through, more time spent on rigamarole that is not research. 
 
On the plus side, cOAlition S is making improvements to the Journal Checker Tool. We reported last month in The Brief on some of the deficiencies in the third-party data underlying the JCT. cOAlition S has provided publishers with more detailed instructions on how to supply updated data, particularly with regard to transformative agreements. This is helpful and the JCT may yet prove to provide a useful service to authors and others interested in the scope of such agreements. 
 
In a recent posting to sOApbox (the Plan S blog), Robert Kiley of the Wellcome Trust has also clarified the circumstances under which the JCT will display (or not display) whether the RRS can be used for Plan S compliance with regard to a given journal. Unfortunately, the clarification raises as many questions as it answers. Kiley writes:

For the avoidance of doubt, the JCT is not asserting that the publisher policy supports the RRS. Rather, it indicates that the RRS allows researchers to self-archive the AAM with a zero-month embargo and a CC BY licence. The only exception to this is if a publisher notifies cOAlition S that manuscripts which include the Rights Retention language will be rejected at submission. In such cases, the JCT will show that there are no compliant options, even if the publisher offers Transformative Arrangements, including Transformative Agreements (like Read and Publish deals) or Transformative Journals (TJs). To be clear, if manuscripts are rejected at submission because of the RRS language, they cannot then go on to be published through Transformative Arrangements.

We confess we are lost here. The JCT is not asserting that the publisher’s policy supports RRS, but will indicate if the publisher’s policy does not support RRS? We are also befuddled as to why a publisher that does not support RRS but has signed a transformative agreement (or has been approved as a Plan S Transformative Journal!) that will result in the VOR being OA under a CC BY license would be considered noncompliant. It sounds like Kiley is saying that a publisher that has entered into arrangements (transformative agreements or transformative journals) that would result in an OA VOR are no longer considered Plan S compliant if they do not also support release of an OA AAM. There may be some policy nuances we are missing but it seems counterproductive to go through all the trouble of promulgating transformative agreements and transformative journals focused on making the VOR OA, only to turn around and now say that such agreements (and journals) are not good enough because they don’t also make the AAM OA. As Kiley says, all this was done “for the avoidance of doubt.”


Source: PucOAlition S, IOP Publishing, OASPA Blog, sOApbox

2

Professional and Academic Publishing

Elsevier announces that it will follow Nature’s lead in APC pricing for its flagship Cell portfolio:

…for those [Cell Press] titles that do not currently offer an open access publishing option, the article publishing charges (APCs) will be £7,000/ €7,600 /$8,900 for the majority of journals, and £7,800/€8,500/$9,900 for our flagship title, Cell. We will align our existing Cell Press APCs for hybrid journals with these new publishing charges.

As we noted in the November 2020 issue of The Brief (Item 1), Springer Nature set APCs across the Nature portfolio at €9,500, or approximately $12,000. Nature’s APC includes all the Nature hybrid titles, including the flagship itself. So depending on the title, Cell Press offers a relative bargain! 
 
Elsevier’s announcement triggered many of the same reactions as that of Springer Nature, but this is just math. There is no way to publish highly selective journals using an output-based OA model without costs for said outputs being high. The subscription model spreads the costs around to a wider readership; the OA model (at least any OA model where payment is tied to output) consolidates those costs. This is what funders and national consortia promulgating publish and read deals have asked for — Springer Nature and Elsevier are merely meeting the demands of their customers. The test will be how many funders and consortia actually pay. 
 
While the cost of an APC in Cell or Nature has been compared to the purchase of a small car, that view misunderstands what is being purchased. Paying an APC to be published in Nature or Cell is not the equivalent of purchasing a product. It is the equivalent of purchasing both a credential and a marketing service. If a funder or university were to hire a PR agency to promote the research, how much would that service cost and how effective would it be as compared to publication in Nature or Cell? If an author were to purchase a credential that led to career advancement (tenure, promotion, grant awards, speaking engagements), what would that cost? The automotive comparison is confusing the economies of mass production with that of services and credentials. While we have no point of view as to whether the APCs are priced “correctly,” they are much less than the cost of purchasing bespoke marketing services or other credentials. But this is not something worth debating — there is a mechanism called the market that will tell us in relatively short order whether the APC pricing is optimal. The stakes are high not just for Elsevier and Springer Nature; at issue is whether highly selective journals can work with an output-based business model.


Source: Elsevier, @jeroenbosman via Twitter

3

In other Elsevier news, the world’s largest publisher has signed on to DORA (Declaration on Research Assessment). In making the announcement, Elsevier affirmed it will be making the reference sections of its journals openly accessible via Crossref (in keeping with DORA principles — see number 9). This brings a large swath of the scientific corpus into Crossref’s open citation graph. With the addition of Elsevier’s portfolio, nearly every large commercial publisher in the industry now has openly accessible references (the Initiative for Open Citation, or I4OC, has been tracking participation). The only notable commercial absence we could identify is that of Wolters Kluwer. (Curiously, there remain a number of notable not-for-profit publishers who are absent from the I4OC list). An open data graph is helpful not only for enabling new bibliometric services (which explains why Elsevier, which owns Scopus, has long held its citations back) but in improving basic article-level metrics for all journals that employ them.


Source: Elsevier, DORA, I4OC

4

What every academic publisher is wondering about is the health (meaning the budgets) of libraries and their parent institutions as the consequences of the pandemic come into sharper focus. Ithaka S+R did an extensive survey of library directors and got a surprising number (638 or roughly 43%) to participate. The outlook is not good, but it is not as stark as feared. For example, “The scientific research enterprise has thus far faced essentially no cuts and even benefited in some cases from targeted stimulus.” That suggests to us that the feared decline in article submissions a year or two out may not occur. Another interesting, and unexpected, development is the surge in the value of endowments for those institutions and philanthropies lucky enough to have one, as Wall Street has gone on a real tear since March. On the other hand, enrollments are down, taking tuition payments with them; and international enrollments, which generate a higher margin, have dropped sharply (because of the reluctance to, and sometimes outright ban on, travel). But more than half of the library directors expect to see cuts, though the cuts are not equally distributed; 4-year liberal arts colleges, for example, anticipate more stable budgets than do doctoral institutions. The striking table in the Ithaka report is the one that asks if cuts will be temporary or permanent. Almost half of the respondents say they are not sure, a number large enough to put all forecasts in doubt. The report’s conclusion is sobering: “Some might be tempted to assume that the sector is bottoming out and will soon recover. To the contrary, we see risks of a further slide. In the current set of findings, nearly a quarter of responding directors report not having a budget for the current fiscal year: this often means that they are subject to expenditure controls and will spend less, in some cases far less, than they did in previous years. And while many directors are not sure what the future will bring, we have heard anecdotally from some that the cuts they are experiencing in the current fiscal year are part of a university effort to smooth anticipated reductions over two or more years, meaning that they anticipate further cuts next year.”


Source: The Scholarly Kitchen

5

eLife has announced in an editorial by its editors that it will henceforth (with some caveats for authors who opt out) only review papers that are first posted as preprints. The editorial by Eisen et al. notes that 75% of the papers it publishes now are posted as preprints, so really it is a policy that is describing what more or less is already happening. While this may not be as novel a practice as the editorial suggests — claiming to be “the first major journal to move to only reviewing preprints” ignores the fact that nearly all papers that appear in physics journals have long appeared first in preprint form — the journal is nonetheless breaking some new ground. 
 
eLife has made reviewer reports public for some time on its website. It will now be posting reviews back to the relevant preprint server. See, for example, the paper “Toxoplasma gondii Infection Drives Conversion of NK Cells into ILC1s” by Park et al. as published in eLife. Clicking on “Decision letter” on the left menu will take you to the peer review report and the authors’ response. The preprint version of the paper appears on bioRxiv. If you click on the “Peer Reviews” button on the right, it will bring up a version of the decision letter. Curiously, it is not exactly the same letter. The version on bioRxiv is redacted, and only includes the Summary and some of the Essential Revisions, but excludes the list of 27 items deemed “essential to revise” by the reviewers. The bioRxiv version also omits the authors’ response. It is unclear if the difference between the two reports is intentional or inadvertent (and it is notable that eLife, and not bioRxiv, manages the review posting via an integration with Hypothes.is). 
 
We applaud tighter integration between the version of record (VOR) and the preprint that preceded it. At the moment, however, that integration remains opaque. In looking at the article as published in eLife, there is no indication that a related preprint exists (on the bioRxiv version there is a clear link to the VOR at eLife). It is also unclear to us what the advantage is of two separate versions of the published reviewer reports. Why not just include a link to the review (that resides on the eLife website) from the preprint as opposed to copying and pasting (a redacted version of) the review? 
 
These are quibbles and may simply be a result of the editorial policy at eLife being out in front of the technical implementation (which, we must add, is better than the reverse!). We must register an objection, however, to eLife’s characterization of “publishing.” Eisen et al. write:

… for all practical purposes eLife is no longer a publisher: rather, eLife is now an organization that reviews and certifies papers that have already been published.

This view misrepresents the role of a publisher. The growth of preprints in the life sciences has been a positive development for science (perhaps less so for science journalism), but posting a preprint is not publishing. Reviewing and certifying papers are essential aspects of scholarly publishing, not activities divorced from it — after all, it is the journal providing the certification. While print distribution may no longer be fundamental to scholarly publishing, bringing papers to the attention of an audience is (an audience cultivated by the journal and anchored by its reputation). Publishing also entails taking responsibility for the paper. This means being responsible for archiving, managing retractions, and providing ongoing technical support (providing the “home” for the version of record of the paper and managing all of the integrations with the scholarly ecosystem as they stand now and as they develop in the future). 
 
Were eLife a different journal, this characterization of publishing would not warrant comment. It is difficult to read the passage quoted above, however, without it being colored by the knowledge that eLife was founded and is supported by research funders who appear increasingly allergic to the evaluation of the research they fund by independent third parties. Eisen et al. want to reinvent scientific communication and that is not necessarily a bad stance for an editorial team to hold. It can lead to bold experimentation, novel approaches, and the questioning of sacred cows. It also happens to align well with the motives of eLife’s financial backers, who would not shed a tear were the role of publishers to be diminished.


Source: The eLife, bioRxiv, The Scholarly Kitchen, @mbeisen via Twitter

6

Editors and publishers of journals that receive a higher 2020 journal impact factor (JIF) may want to hold off on celebrating until they look at the JIFs of their competitors. An analysis by Phil Davis published in The Scholarly Kitchen suggests a rule change at Clarivate will result in an inflationary JIF effect for many journals. The rule change pertains to what papers will contribute to the impact factor in what year. At present, for some journals, Clarivate assigns the year of publication based on when the article was published online. In other cases, it goes by the publication date of the issue in which the article ultimately appears. This is notable as some papers are published online months in advance of assignment to an issue. A paper published in September 2020, for example, might later be assigned to the January 2021 issue (and indeed that is a strategy some publishers have used in order to boost their JIF by essentially adding several months to the citation window for these papers). Clarivate’s rule change simplifies things and brings the calculation of the JIF into the modern era of publishing by going by the date of article publication only. 
 
While this is a welcome change, it will have the effect of increasing the number of papers considered “published” in 2020. This is because 2020 is a transitional year for the policy, and hence includes both any papers published online in 2019 but assigned to an issue in 2020 and any papers published online in 2020 that in the past would have been assigned to 2021 issues (but are now assigned to 2020). As a reminder, the 2020 JIF (which is released in 2021) is calculated by adding up the total number of 2020 citations to articles published in a given journal in 2019 and 2018 (the numerator), divided by the number of papers (citable items) published in the same journal in 2019 and 2018 (the denominator). In other words, for 2020, the increase in papers results in an increase in citations (the numerator) but not an increase in citable items (the denominator), resulting in higher JIF scores. 
 
The inflationary effect should be a “rising tide lifts all boats” situation. However, some journals may receive a modest one-time boost due to self-citation. If, for example, a given journal has a bolus of additional articles assigned to 2020 that heretofore would have been assigned to 2021, and cites itself to a certain extent, its JIF may be boosted further. Any inflationary effects caused by this rule change, should, however, even out over the next few years. Additional citations in 2020 are likely to mean fewer citations in 2021. 
 
It is likely, however, that this inflationary effect will be masked by a much larger inflationary effect: the pandemic paper surge (which may be why Clarivate chose to announce the rule change toward the end of 2020, after the magnitude of the COVID-19 surge has come into focus). Christos Petrou estimates that paper output in 2020 will increase by 17%–26%. While papers related directly to COVID-19 are certainly a contributor to this surge, much of the increase is due to researchers being at home with time to focus on writing up a backlog of work. While Petrou did not use Web of Science (WoS) in his analysis (and hence his results will vary somewhat as compared to WoS), assuming WoS is directionally similar we will see some eye-popping numbers in the 2020 Journal Citation Reports. However, unless the present output is maintained into 2021 and 2022, the surge will likely depress impact factors as the 2020 papers move from the numerator into the denominator of the JIF equation. Just remember before you uncork the champagne (or, in future years, drown your sorrow in something stronger) that absolute scores are meaningless: it is all relative!


Source: The Scholarly Kitchen

7

In his long and distinguished career, Richard Charkin has worked in both trade and academic publishing. As he puts it, “Looking back over my 48 fulfilling years in the publishing industry, I estimate that I’ve spent 40 percent of my time and effort on trade books, 40 percent on academic books, and 20 percent on education.” He has now written a brief essay in which he compares academic and trade publishing, a piece that is replete with one candidate for a pull quote after another. For example, “Possibly the greatest publisher of the 20th century is also the most loathed, Robert Maxwell. He made a fortune — and lost it — out of scientific publishing. However, making money was necessary but not the main driving force for his publishing. He really cared about the scientists who were served by Pergamon Press. He transformed a sleepy, self-obsessed industry into the vibrant and innovative network it is today.” And he notes that while academic publishing is mission-based and focused on prestige, whereas trade is about sales and profits, “The irony is that for most of the last 50 years, academic publishing has been significantly more profitable than trade publishing and in some respects trade publishing has become more prestigious than academic.” We could quibble with Charkin on a couple points. For example, he does not mention a distinguishing feature of academic publishing, namely, the sheer number of not-for-profit organizations operating in it, something with no parallel in trade (or anywhere else). And we would add that trade publishing may have little growth ahead of it, as it competes with other forms of entertainment (e.g., Netflix) that gobble up people’s discretionary time; whereas academic publishing is likely to continue to grow, albeit unevenly, as the world’s economy is built upon it.


Source: Publishing Perspectives

Dealmaking

8

Elsevier acquires Shadow Health, a developer of virtual simulations in nursing and healthcare education.


Source: Elsevier

The Book Business

9

In addition to being the world’s largest bookseller, Amazon is also a very large publisher, with over 1 million titles in its catalog. These titles have not been available in libraries, to the dismay of librarians, who have been agitating for legislative action to “correct” what they view as monopoly abuse. At bottom, the claim is that people have a right to read a book, and that publishers have an obligation to fulfill that right. Perhaps fearing the current mania for dragging tech companies into court on antitrust charges, Amazon is now deep in negotiations with the Digital Public Library of America. DPLA has said how a prospective arrangement would work: “First and foremost, the discussion covers Amazon Publishing titles only (not titles from Amazon’s KDP program). The current talks also do not include Audible, Amazon’s digital audio service, which does not make its exclusive content available to libraries. And while Amazon is heavily invested in a subscription model for books and reading (Audible, Kindle Unlimited) a subscription model for libraries has not been part of the talks. All titles under the potential deal would be licensed ePub editions managed by the DPLA and its partner libraries and made accessible to patrons via the DPLA’s SimplyE app — meaning library users would not have to go through Amazon to access the titles.” We will be watching this situation carefully, all the while wondering about the multitude of other publishers that do not sell materials to libraries and whether they will be compelled to do so.


Source: The Hill, Publishers Weekly

10

Can James Daunt do it? The stakes are high. Daunt is the CEO of Barnes & Noble; he also serves as CEO of Waterstones in the UK; both chains are owned by Elliott Management. Daunt is credited with accomplishing a remarkable turnaround at Waterstones, at a time when brick-and-mortar bookstores were dying everywhere, victims of high rents, competition for leisure time (think Netflix, Disney+), and that little operation in Seattle that shall not be named. And now there is a pandemic to test Daunt’s mettle. But Daunt is fearless. He sees the future of B&N’s 600 stores in decentralization and the empowerment of local managers; in effect, he is trying to make the still huge B&N (with sales around $3.5 billion — half of what they were ten years ago) operate as a loosely connected set of independent bookstores. No two stores will look alike or display merchandise in the same way; local management will make merchandising decisions, on the principle that customers in Idaho want different things than customers in New York City or Austin. B&N will still hold a great deal of operational control (and the efficiencies that go with it) at the corporate level, but Daunt’s vision is as surprising as it is electrifying. He is not naive or hopelessly romantic: “As you let the stores diverge, a quarter will be brilliant and a quarter will be absolutely terrible…. A significant number of your stores will become worse, not better. Then you teach and encourage them and, in time, everybody becomes better.” Do you mean to tell us that it will actually be fun to step into a B&N bookstore again? We are lining up with our masks to get at this small slice of a civilized life gone by.


Source: The Wall Street Journal

Technology

11

It is a truth universally acknowledged that an author wishing to submit his or her paper to a journal must endure the painful and vexing trial-by-software that is the modern manuscript submission and review system. Short of applying to refinance one’s mortgage, it is hard to point to a more unpleasant submission experience. The reason for this is largely due to the fact that such systems are not built to make the process easy for authors — they are built to make the review process easier for editors to manage. Authors, it has long been assumed, will put up with any inconveniences and indignities visited upon them (or their post-docs) in order to submit a paper to their preferred journal. However, the dynamics of open access publishing have prompted revisiting this assumption. When authors are paying customers, and with a proliferation of publishing venues, it is time to rethink the submission process from the author’s perspective. Enter Wiley ReX (Research Exchange). 
 
Wiley has cleverly separated the submission system from the review system. A journal can continue to use its current review system but authors do not have to interact with it (at least to submit a paper; it is unclear to us whether they must interact with such systems to view reviewer reports and submit a revision). ReX is designed with the author experience in mind and uses entity extraction and machine learning to simplify the submission process. While we do not as yet have a view as to whether ReX does this well, we applaud the concept. An author-friendly submission system is long overdue. 


Source: Wiley

12

Professional associations have long struggled with customizing content and messaging to the myriad segments of their membership and their wider audience that often have very different information needs and interests. Silverchair has launched Hum, a customer data platform (CDP), to address this problem. Hum acts as a sifting and sorting mechanism that integrates with an association’s acronym stack (AMS, LMS, CMS, MarTech) to match people with relevant content. Hum is an entirely new venture from Silverchair and not part of its Silverchair Platform (though, of course, integrates with it). 


Source: Hum

13

The biggest story in tech this month is of course the antitrust suit against Facebook brought by the U.S. Federal Trade Commission and more than 40 states. This follows an antitrust suit against Google brought in October by the U.S. Department of Justice (DOJ), joined by 10 states. While the legal armies that will do battle over these cases are still mustering, some intellectual skirmishes have already begun. For some of us, an engrossing and enlightening intellectual debate on the nature of monopoly among today’s tech companies is akin to watching a playoff game. And so we have broken out the beer nuts and poured ourselves a crisp lager to follow the back and forth between Ben Thompson (Stratechery) and Tim Wu (Columbia University) on the Google antitrust case. 
 
The debate is over the extent to which Aggregation Theory, an important concept of Thompson’s that is essential for understanding the market power of Google, is relevant to the antitrust case brought by DOJ. This may sound arcane, but (and regardless of whose side you might ultimately land on) the debate proves to be an extraordinarily good primer on how the nature of digital distribution has changed markets with implications for anyone doing business online. 
 
Here are links to the relevant pieces in chronological order:
 
Thompson: United States v. Google
 
Wu: Ben Thompson’s “Stratechery”: Smart, but a little too much Kool-Aid
 
Thompson: Is the Internet Different?
 
Wu: Ben Thompson’s Stratechery, Part 2: In which the Kool-Aid gets stirred


Source: The New York TimesStratechery, Tim Wu (via Medium)

14

Okay, we have to admit that everything about AI is weird and reading about it brings to mind images of Arnold Schwarzenegger in the Terminator movies. But here is the weirdest of the weird: just like humans, certain kinds of AIs may need to take a nap. It seems that some AIs experience “instability” when they keep performing the same task over and over. The instability is not universal, however. “This sort of instability is not a characteristic of all AI networks. The issue only arises when training biologically realistic processors, or when trying to understand biology itself. The vast majority of researchers on machine learning, deep learning and AI never encounter this instability because, in the very artificial systems they study, they have the luxury of performing mathematical operations that have no equivalent in living neurons.” So if you come across an error, the explanation is either (as HAL says in 2001: A Space Odyssey) human error or, with biologically realistic processors, human example. Which prompts a kids’ joke: What does the AI say before a nap? “I’ll be back.”


Source: Scientific American

People

15

The British Standards Institution announced that Susan Taylor Martin will be appointed as Chief Executive of BSI in 2021. 
 
Kirsty Meddings has passed away. She served the last 12 years as Product Manager at Crossref and was well-known and respected throughout the industry. 
 
John le Carré, whose real name was David Cornwell, died at 89. He transformed espionage fiction and ultimately transcended the genre. He also may have written the most riveting autobiographical sketch ever published in The New Yorker (he had the material to work with: his father was an international con man who operated on a grand scale). 

Briefly Noted

16

Frankfurter Buchmesse restructures, reducing headcount. 

Implementing a Subscribe to Open model: A case study from IWA Publishing.

College students build their own multispectral imaging system and find hidden text on a medieval manuscript.

Bob Dylan sells all publishing rights to his songs for a price estimated to be over $300 million, but the precise terms of the deal are blowin’ in the wind.

Elsevier’s sister organization, Reed Exhibitions, has not weathered the pandemic well. It anticipates cuts of over 500 jobs. As part of this, Reed will be shutting down BookExpo, the U.S.’s largest exhibition for trade books.

JRR Tolkien’s house is on the market. The listing makes no mention of balrogs in the basement or dwarfs at the door, but it may be worth taking out a rider on your homeowner’s policy to insure against such unwelcome guests should you choose to purchase the home. The not-for-profit Project Northmoor seeks to acquire the property, with the backing of a group of celebrities including Gandalf himself (Ian McKellen), and turn it into a center for Tolkien studies. 

The Economist lists all the books we were supposed to be reading during the pandemic lockdown. 

The U.S. Government Publishing Office will permanently close its last brick-and-mortar bookstore, in Washington, DC. At one point, the GPO operated 27 bookstores across the U.S.

More libraries are going fine-free for overdue books.

The Onion surveys new research on Charles Dickens’s A Christmas Carol.

Nature has assembled a selection of the best science images that appeared in 2020.

How an obscure British PC maker, collaborating with a BBC television show, invented ARM and changed computing

The state of not-for-profit publishing today.

Digital Science releases report on the State of Open Data in 2020.

Richard Poynder on the history of OA (a very long read).

***
Now this is not the end. It is not even the beginning of the end. But it is, perhaps, the end of the beginning— Winston Churchill