1,430 Matching Annotations
  1. Aug 2026
    1. Anumber of other options were considered atthe 1990 Harvard workshop. For example,NSF could have turned over backbone opera-tions to a nonprofit operator. Wolff even sug-gested ‘‘privatization by function,’’ in whichcommercially viable services such as mailand news would be blocked from the NSFbackbone, while more specialized researchapplications would continue to use it.

      Alternatives to privatization.

    2. First, because NSF did not imposeperformance requirements on the NAP oper-ators, the NAPs became congested andimpeded backbone traffic flow. 57 Second, inthe absence of a regulatory requirement fornondiscriminatory interconnection amongISPs, large backbone providers or privateNAPs were able to exclude smaller playersand thus undermine competition, increaseconcentration, and drive up prices. The lackof rules specifying universal unmetered inter-connection led to an unequal two-tier back-bone peering system, in which large ISPsagreed to exchange traffic for free butcharged smaller ISPs for the same service. 58Third, the lack of regulation means that un-like phone companies, backbone providerscan, and sometimes do, harm subscribers bycutting off their connectivity on short no-tice. 59 Fourth, NSF did not require neededupgrades to network service, such asimproved security. Other commentatorsraised a host of concerns ranging from tech-nical to social and political.

      Issues raised in the internet's privatization.

    3. Thebackbone’s communications links had al-ways been leased from commercial providers,the regional and local networks that made upthe bulk of the Internet did not change own-ership, the TCP/IP protocols had been andremained nonproprietary, and control overrouting and technical development stayedwith nonprofit groups. Privatization was lessa change in ownership than a change inwho could serve NSF’s users, how thoseusers were subsidized (directly throughgrants, rather than through a subsidizedinfrastructure), and who made policy forthe backbone.

      End result of the internet's move to privitization.

    4. Internet connections were tobe commercially provided, with governmentefforts focused on developing applicationssuch as educational software, subsidizing ac-cess for libraries and schools, providing train-ing for librarians and educators, and makinggovernment information available over thenetwork.

      Federal investments move higher up the stack.

    5. TheGore Bill, as it was known, stressed the threatof foreign competition as its political ratio-nale and reflected three beliefs: that high-speed (gigabit) networks were essential toresearch and as an economic driver, thatindustry would not provide such networksbecause the market and technology werenot yet proven, and that government over-sight was needed to ensure equitable accessand proper network use.

      Again motivation from perceived international competition.

    6. Merit informed NSFthat it planned to create a nonprofit corpora-tion called Advanced Network & Services, forwhich IBM and MCI would provide severalmillion dollars in funding as well as person-nel and equipment.

      Bridge nonprofit to foster commercial development.

    7. ome saw the networkas a public trust, others as an opportunityfor profit; some wanted a testbed for high-performance technology, while others hopedfor an economic driver and an edge in inter-national competition.

      Competing motivations for advancing the internet in the 1980s.

    8. Merit, a consortium of Michi-gan universities founded in 1966 to promotecampus and state-wide networking. Meritserved as project manager, provided engi-neering support, and ran the Network Oper-ations Center. Merit’s partners were the Stateof Michigan Strategic Fund (which offered$5 million), IBM (which supplied the packetswitches), and MCI (which provided the cir-cuits).

      Partners in the first NSFNet management delivery.

    9. Therefore, NSF didnot address the issue of commercial use of thenetwork by the awardee in the solicitation orthe Cooperative Agreement. Similarly, noneof the offerors addressed commercial use inits proposal.1

      Lack of commercial elements in the first NSFNet backbone agreement.

    10. For scientists who cared littleabout networking per se and simply wantedto get on with their research, applicationswith consistent, user-friendly interfaceswere as important as the network itself:

      Primacy of UX among some early network users.

  2. Jul 2026

    Tags

    Annotators

  3. Dec 2025
    1. This is not a new critique. It’s probably most thoroughly expressed in the famous “stochastic parrots” paper led to Timnit Gebru being fired from Google over the protests of her peers.

      Is there a word missing from this sentence?

  4. Mar 2025
    1. I asked our friend Dr. Oblivion, Why is it better to refer to AI hallucinations and AI mirages? His response.

      I'm assuming this is some kind of ✨sparkling intelligence✨ and given that Dr. Oblivion seems to miss the point of the paper and our discussion here, I found it more illustrative than helpful ;)

    1. It’s not the universities’ fault, but they have certainly played their part in creating and perpetuating these inequalities.

      hm, does EDU only perpetuate wealth inequality? I'm not sure that the the question of EDU's role in US economic mobility is so fully settled...

    2. The incident encapsulated the entitlement, the arrogance and the unbearably petty grievances of a generation who seemed to find their voice and power in the taking of offense.

      Stephens starts out by suggesting that the issue is generational.

    3. He said presidents and professors had taken too many things for granted — they thought they’d always be seen as a “public good” benefiting society, but came to be seen as elitist and condescending toward regular Americans. And Americans hate a lot of things, but they really hate elites condescending to them. Now we are seeing a big reckoning for higher education — ideological, cultural, financial — driven by Donald Trump and the right.

      This statement draws a direct connection between what the Trump administration is doing to EDU and a vague description of an alleged popular opinion, while providing no proof of the connection. While Trumpists may be fomenting and harnessing anti-EDU sentiment, they may have other reasons for their anti-EDU actions.

      On popular opinion about EDU, see https://www.pewresearch.org/politics/2024/02/01/colleges-and-universities-k-12-public-schools/

  5. Feb 2025
  6. Jan 2025
    1. student voices, perspectives, and agency should be more prominently featured in OER decision making

      This is perhaps the most compelling recommendation here. And bringing in the student voice not only in OER, but also in the other ways in which OER support additional open educational practices.

    2. While there is no one size fits all model for running and growing OER initiatives, certain elements of institutionalization could prove beneficial. For instance, a succession plan would ensure that when an OER leader leaves their role, whether on campus or at the state level, the initiative could continue. OER efforts also require the professionalization of roles that manage content development, dissemination, and support. Alternatively, responsibilities for OER implementation could be dispersed across multiple individuals or coordinated through a centralized office, possibly at the state level, to ensure long-term success and accessibility.

      This recommendation seems key as clearly, the work to increase student success will be long.

    3. The fact that we have saved our students more than half a million dollars now with our products gives me great joy, because we are at an institution where our population, a lot of our students, have very little means. Most of our students are under some sort of financial help. So many of them are first generation, so many of them have jobs and families.

      Is this a quote? If so, it seems like it should be marked and attributed.

    4. the departure of key stakeholders could disrupt ongoing efforts and threaten the continuity of OER initiatives

      This is absolutely key: The OER movement has too long relied on individuals and not realized institutionalization.

    5. OER initiatives flourish when they benefit from multiple-framings, such as advancing equity goals, instructional design innovation, or student success outcomes, in addition to cost-savings benefits.

      This seems like the headline: OER is not an end in itself, but as one tool in enabling progress in these larger and more impactful goals.

    6. there is clearly no one size fits all model for the OER initiatives

      This heterogeneity seems like a symptom of efforts that are working against dominant commercial practices.

    7. most initiatives have been funded through federal and/or state grants

      At issue: while individual learners may provide the financial support for commercial endeavors (eg, textbook costs), the costs for OER seem to always need to be underwritten at a higher level, often with public funding.

    8. the equity and accessibility of OERs

      While I concur with the need to continue to examine the equity and accessibility of OER, I wonder if it is being held to a standard that proprietary commercial resources are not held. equity and inclusivity goals? What equity and inclusivity goals are commercial proprietary educational materials setting and/or meeting?

    9. most financial models described in the literature come from the corporate world, putting the need to maximize revenue in tension with higher education’s mission-based goals that are harder to quantify

      interesting: perhaps the overall impacts of education are hard to measure against the brass tacks of commercial endeavors...there are likely studies in the field of economics that could help weigh education's less monetary measures against commercial activities...the following paragraph suggests some avenues for exploration

    10. If OERs save students money, while also acting as catalysts for faculty to make improvements to their teaching, it is no wonder that colleges and universities are trying to expand them on their campuses.

      This is a core insight — while the cost savings and related accessibility of OER are compelling, there is also the collateral effects of instructors and instructional designers revisiting course materials and practices that can lead to other improvements.

    11. OERs

      I'm not sure why this even matters, but I would have loved to see this report use just "OER" in both the singular and plural cases. Maybe it's a bit like "deer" versus "deers"...

      So this sentence would be: "In those instances, OER is framed as both a cost-savings tool and also as a means to advance student success and equity."

    1. In an era where trust in science and concerns about misinformation areparamount, publishers could leverage anxieties about the risks of sharing "non-peer-reviewed" material to cast doubt on the reliability and credibility of preprints.

      A deep concern, especially in a wider context where the value of expertise itself is being undermined...

    2. allows authors to declare a VOR should they wish to, treating it as a ‘state’ within acontinuous process rather than a definitive endpoin

      another possible necessary cultural shift for PRC to succeed? where scholarly progress is recognized as a more dynamic activity with crystalized moments (ie, VORs)?

    3. This monetization strategy highlights a potential risk of the PRC model being co-opted by commercial publishers to prioritize profit over the principles of openness andaccessibility. By charging for submissions, even those that do not proceed to full curationor publication, journals could exploit the publish-first ethos of PRC for financial gain,further complicating the already high cost of academic publishing for researchers andcausing even more inequities.

      good to see this focus on the way money flows in the scholarly communications system...I wonder if it might not be brought up earlier as a concern to more fully addressed in its proper place as it is here...

    4. an open, decentralized ecosystem where innovation flourishes,traditional hierarchies are dismantled, and the community takes control of its ownnarrative and processes

      maybe a good point to reflect on the ways in which the open source software community did not fully open and dismantle hierarchies, and how maybe scholarly communication is also not yet as open and non-hierarchical as we might want it to be

    5. curation emerging in many forms as anecessary outcome of this shift

      maybe it's not surprising that the part of PRC where most of the money is located is also the part that's hardest to redefine...

    6. dynamic post-publication evaluation

      "dynamic post-publication evaluation" is a key phrase in this work...perhaps "dynamic post-distribution evaluation" to differentiate between depositing work in preprint form and publishing it in a curation?

    7. By using termslike “deposit” instead of “submit,” “distribute” rather than “publish,” “preprint” instead of“article,” and “server” instead of “publisher.” These terms signaled that the preprintprocess was no challenge to the journal system, presenting itself merely as a mechanicalprocess akin to hosting files on an FTP server—“not really publishing.”

      See my note above on possibly renaming the PRC idea...

    8. The parallels between these two models and scholarly publishing are striking.

      The characters of the two systems seem well-aligned with the cathedral/bazaar metaphor, but what of the financial and motivational aspects of the two systems?

    9. Publish-Review-Curate (PRC

      I have to admit, when I first saw "PRC" my immediate thought was the People's Republic of China — even though that's clearly way off topic — and it took me a few pages in to the document to see it defined. I realize insiders may get PRC right away, but I wonder if there might be a more distinct shorthand for this topic, like maybe "PuReCu", which could maybe be pronounced "Pure Queue"...

  7. Jan 2024
  8. Dec 2023
    1. Beyond simpleassociations it acquires high-level abstractions like expressive structure, ideology or beliefsystems, since these are all embodied in the corpora that make up its training sets.

      hm, I'm not sure how LLMs acquire these higher-level concepts out of the probabilistic relations just described.

  9. Nov 2023
  10. Jun 2023
  11. Mar 2023
  12. Feb 2023
    1. It means that everything AI makes would immediately enter the public domain and be available to every other creator to use, as they wish, in perpetuity and without permission.

      One issue with blanket, automatic entry of AI-generated works to the public domain is privacy: A human using AI could have good reasons not to have the outputs of their use made public.

    1. a commercial entity would probably find itself at the end ofsuccessful litigation if it used such images without authorisation

      I had to read this a couple of times before I concluded that the meaning was a commercial entity would likely lose a case where it had infringed a trademark. It was unclear which commercial entity was mentioned: the trademark holder or the alleged infringer?

    2. It could be argued that the training of an AI model does not fulfil all of these requirements,but it is likely to be something that will be argued in court by future defendants incopyright litigation. It is clear that making a temporary copy for training is transient, butit could be said that it is not incidental. The copy is also part of a technological process,but it may not be considered a lawful use. Similarly, one could argue that the resultingmodel does have economic significance, but specific copies do not, a model such as StableDiffusion can be trained with billions of images, each individual copy used in training maynot count as having “independent economic significance”.

      Key questions around exceptions.

  13. Dec 2022
  14. Nov 2022
  15. Jul 2022
  16. Jun 2022
  17. May 2022
    1. “Establish a core distribution requirement focused on the history and legacy of racism in the country and on the campus.” There would be wisdom in this time of disunity in suggesting (not, in my view, requiring) that students take courses in American history and constitutionalism, both of which almost inevitably consider slavery and race, but that is not the same thing. Not incidentally, if you believe anti-blackness to be foundational, it is not a stretch to imagine that you will teach the 1619 Project as dogma.

      This section demonstrates how calls to "objectivity" undermine critical thinking. Objectivity becomes its own dogma.

    2. In 1776 there were “united States” but there was not yet the “United States”; in these past two months, by contrast, at a time when we are increasingly un-united, “black” has become “Black” while “white” remains “white.”

      A focus on "proper" writing: recognizing the historical specificity of why "united" might not yet have been capitalized in 1776, but avoiding an understanding of why today one might capitalize Black and white differently.

    1. Subsidiarity, which uses “data cooperatives, collaboratives, and trusts with privacy-preserving and -enhancing techniques for data processing, such as federated learning and secure multiparty computation.”

      Another value of the data cooperative model might be that each individual might not have time to research and administer possible new data-sharing requests/opportunities, and it would be helpful to entrust that work to a cooperative entity that already has one's trust.

    2. In general, I would say that I think there are only a few circumstances in which markets produce good incentives and distributions, and that these depend heavily on publicly accountable governance that set up their rules.

      Amen. This resonates with my concern about systems that want to govern human relations without humans at the center. It's not like I believe having humans at the center of our relations guarantees good practices and outcomes (we have ample proof it does not), but rather that NOT centering humans in human relations may cast aside what's good about humanity along with what's bad.

  18. Apr 2022
    1. One of those factors is globalization which has helped lift hundreds and millions out of poverty, most notably in China and India, but which, along with automation has also ended entire economies, accelerated global inequality, and left millions of others feeling betrayed and angry at existing political institutions.

      An awareness of other structural, economic issues that are weakening democracy: Globalization, Automation, Inequality.

    2. We’ll have to come up with new models for a more inclusive, equitable capitalism. We’ll have to reform our political institutions in ways that allow people to be heard and give them real agency. We’ll have to tell better stories about ourselves and how we can live together, despite our differences.

      Obama's three strategies: better capitalism, better politics, better stories/communications.

    3. National School of Journalism and Public Discourse

      "National School of Journalism (NSoJ) is a highly selective J-school that identifies and trains India's best journalistic talents. Our newsroom-focused curriculum, in-house digital news portal, expert faculty members and unparalleled industry connections prepare our students for successful careers in broadcast, print and convergence journalism." Visit website >

    1. the universal principles of equality, dignity, and freedom

      It's easy to reach for these "universal" principles, which have always been universal in theory only. It doesn't surprise me that folks heretofore excluded challenge universal principles that have not served their purposes.

    2. An extensive survey of American political opinion published last year by a nonprofit called More in Common found that a large majority of every group, including black Americans, thought “political correctness” was a problem.

      This is where it gets really interesting. I'd have to explore what branches out from here, but the term "political correctness" is not really just one thing that everyone agrees on, and from where I'm sitting, mostly seems to arise around areas where people who historically have had less of a voice start having one that disturbs established POVs.

    3. True meritocracy came closest to realization with the rise of standardized tests in the 1950s

      Interesting, I'm ready to buy that the post-WWII period had the biggest opening to education in the USA — tho far from truly open or meritocratic and definitely unevenly distributed in many ways, including between K12 and higher ed — but I'm not sure I'd put standardized tests first in a list of reasons for the opening. I'd want to hear more about that.

    1. For students, historical habits of mind constitute majorintellectual hurdles. Students see their professors' thoughts asfinished products, tidied up and packaged for publicpresentation in books, articles, and lectures

      This is where social annotation can be powerful, by making the habits of a reading mind visible, and open for discussion.

    1. Students should be directly involved in campus conversations and decision-making about social reading technologies.

      Love this! It's hard to make happen, but learners voices are so often missing from EdTech conversations and may well be the most important voices to be heard. Given how tech decisions can have huge impacts on learner success and well-being, how can we ensure that they are a bigger part of the conversation?

  19. Mar 2022
    1. the permanence and agency of their work

      This seems like essential digital literacies work that should be part of all educational levels...otherwise, we run the risk of always being mere consumers of knowledge/content services that may not support our agency.

    2. boutique file types wither into obsolescence

      This is such a crucial point...how much content and knowledge is buried in obsolete file types? In the end, HTML may be one of the most robust file types, as well as being one of those that is most easily made accessible. Even the PDF we are reading here has layers of complexity beyond HTML that may make it less available in the future.