In sum, IA has not met its “burden of proving that the secondary use doesnot compete in the relevant market[s].” Warhol I, 11 F.4th at 49. Its empiricalevidence does not disprove market harm, and Publishers convincingly claim bothpresent and future market harm. Any short-term public benefits of IA’s FreeDigital Library are outweighed not only by harm to Publishers and authors butalso by the long-term detriments society may suffer if IA’s infringing use wereallowed to continue. For these reasons, the fourth fair use factor favors Publisher

In this case, the relevant market for purposes of analyzing the fourth fair usefactor is the market for the Works in general, without regard to format. TheCopyright Act protects authors’ works in whatever format they are produced. Seegenerally 17 U.S.C. § 106 (granting copyright owners the exclusive right toreproduce their works in numerous formats). Even if we consider the printeditions to be the “original” forms of the Works in this case―and we recognizethat the concept of a singular “original” form may be difficult to grasp in the digitalage―the Act recognizes that presentation of the same content in a differentmedium is a derivative work subject to the same legal protections as the purportedoriginal. 17 U.S.C. § 106(2). Here, Publishers obtained from authors the exclusiveright to publish their Works in numerous formats, including print and eBooks, andit is this exclusive right that IA is alleged to have violated via its Free DigitalLibrary. For that reason, the relevant harm―or lack thereof―is to Publishers’markets for the Works in any format.

Here, IA makes unauthorized digital copies of the Works and reveals thosecopies to the public in their entirety. The “amount and substantiality” of thecopying is not necessary to achieve a transformative secondary purpose, butserves to substitute Publishers’ books. 17 U.S.C. § 107(3). For that reason, the thirdfair use factor favors Publishers.

Here, while the nonfiction Works undoubtedly convey factual informationand ideas, they also represent the authors’ original expressions of those facts andideas—and those “subjective descriptions and portraits” reflect “the author’sindividualized expression.” Harper & Rowe, 471 U.S. at 563. Thus, because theWorks in Suit are “of the type that the copyright laws value and seek to protect,”the second fair use factor favors Publishers.

We conclude, contrary to the district court, that IA’s use of the Works is notcommercial in nature. It is undisputed that IA is a nonprofit entity and that itdistributes its digital books for free. Of course, IA must solicit some funds to keepthe lights on: its website includes a link to “Donate” to IA, and it has previouslyreceived grant funding to support its various activities. App’x 6091. But unlikethe defendant in TVEyes, who charged users a fee to use its text-searchabledatabase, IA does not profit directly from its Free Digital Library. 883 F.3d at 175.It offers this service free of charge.

But this characterization confuses IA’s practiceswith traditional library lending of print books. IA does not perform the traditionalfunctions of a library; it prepares derivatives of Publishers’ Works and deliversthose derivatives to its users in full. That Section 108 allows libraries to make asmall number of copies for preservation and replacement purposes does not meanthat IA can prepare and distribute derivative works en masse and assert that it issimply performing the traditional functions of a library.

We conclude that IA’s use of the Works is not transformative. IA createsdigital copies of the Works and distributes those copies to its users in full, for free.Its digital copies do not provide criticism, commentary, or information about theoriginals. Nor do they “add[] something new, with a further purpose or differentcharacter, altering the [originals] with new expression, meaning or message.”Campbell, 510 U.S. at 579. Instead, IA’s digital books serve the same exact purposeas the originals: making authors’ works available to read. IA’s Free Digital Libraryis meant to―and does―substitute for the original Works.

IA hosts over 3.2 million digital copies of copyrighted books on its website.Its 5.9 million users effectuate about 70,000 book “borrows” a day―approximately25 million per year. Critically, IA and its users lack permission from copyrightholders to engage in any of these activities. They do not license these materialsfrom publishers, nor do they otherwise compensate authors in connection with thedigitization and distribution of their works.

Total casesadjudicatedConvicted Not convictedTotal Guilty plea Bench/jury trial Total Bench/jury trial DismissedAll offenses 71,126 92.6% 90.9% 1.7% 7.4% 0.3% 7.1%

Right. So let's start here: an Alford plea is fundamentally a form of coercion, because it's basically telling a person, "Admit to this crime or else we'll kill you."

Johanna Hellgren, who has for years has been researching this Alford plea. JOHANNA HELLGREN: Yeah. PETER: So could we just start—like, where did the Alford plea—where did it even come from? JOHANNA HELLGREN: Yeah, the way it came to be was this guy, Henry Alford, was accused of first-degree murder. PETER: This is, like, the early 1960s. So first degree murder meant he was facing the death penalty. JOHANNA HELLGREN: But he took a plea for, I believe, second-degree murder. PETER: Which meant instead he'd get life in prison. JOHANNA HELLGREN: Yeah. PETER: But when he gets up to enter his plea in front of the judge, Alford says—and I'll quote from the transcript right here, hold on one second, he says, "I just pleaded guilty because they said if I didn't they'd gas me." JOHANNA HELLGREN: "I'm just pleading because I don't want to get the death penalty, but I didn't do it." PETER: And then later he said, "I'm not guilty, but I plead guilty."

The realization that you have to live in a society with people who are harmed by injustice, even if you personally escape that justice? It's the whole basis for solidarity.

But actually a physical library cannot make a physical copy of a book and circulate it in place of the original copy. It can't do that, right? So if it can't do that, then you're not doing just what a library is doing. So your argument actually depends on the physical, the digital copy being transformative and the physical copy not being transformative. And that's the thing that makes the difference, not the one-to-one ratio and the analogy to physical books.

The struggle I'm having with your response to these questions is on the one hand you want to say, look, this is transformative because it's efficient and we can get people to read more books faster. They don't have to go to libraries. The efficiency is the value of this to the public. But at the same time, you're saying, but that efficiency has absolutely no impact on whether the publishers can sell the e-books or the hard copies. And it sounds wonderful when you're saying it, but when I step back and listen, I'm having trouble reconciling those two.

that efficiency may or may not have an effect on either the number of copies that get sold or on the market for the overdrive service, which has a variety of different sort of different aspects and benefits over and above CDL. I mean, CDL is largely sort of image-scanned images of pages of paper books because it's the paper book. The overdrive service has a lot of many. You can flow the text. You can do different features and that is one reason why that is one explanation for the data that you see that there is no reduction in demand for overdrive.

you're reducing the market from the number of people who might want to read... Let's look at even the paper books. They'll pretend like take out the digital market for a second. The number of people who might want to read it ever, down to the number of people who might want to read it simultaneously. And if you put digital books into the mix, it's the same idea, right?

That IA's brief and amiki try to create the impression that the public interest is on their side. And it is not. The protection of copyright is in the US Constitution and federal law because it creates an incentive for writers and artists to create new works to benefit our broader society. Internet archives, control digital lending is in direct conflict with that basic principle. And as I previously... You don't really think people are going to stop writing books because of the control digital lending to you? Well, I think publishers are going to go down the tubes if they do not have the revenues. I'm not going to publish your books. You think that that's really... I do, Your Honor. There's no question. I mean, and the standard here is not, will this eliminate... No, I understand. ...the... It's just a part. But this question about balancing the incentive to create a work with the larger distribution of it, that is the question to be decided in this case.

In the same way, control digital lending is a contrived construct that was put together at the behest of Internet Archive back in 2018 when they confronted the fact that libraries didn't want to deal with them. Libraries didn't want to give copies of their works to be digitized because they were concerned about copyright arguments. So they got in a room together with various people and contrived this principle of control digital lending to rationalize what they were doing.

that very point has been made by this court that you transform a book when you create or work when you create it into a new format. But that is not the type of transformativeness that the first factor looks at. You're converting it to a derivative form, not a transformative form.

Really in question that if the comparator on the transformative issue were between the digital book and the ebook, that it really wouldn't be transformative because it is just a different version of a digital book. But he's saying, well, we have a right to use our physical copy and this is a transformative use of the physical copy. Is that the right way to think about it? Are you suggesting that actually the book is something more than the physical copy? And so when we think about the CDL version of the book, we should compare it both to the physical copy and to the ebook because those are both versions of the book that the publisher produces. They are not distributing the physical copy you're on. That's the whole point to control digital lending. That is not the right way to think about it. They are taking the physical copy and transforming it into a nude and different format with different capabilities that has a different market.

You are still distributing the physical copy of the book. And as Your Honor recognized, there's a lot of friction involved with distribution of physical copies that is significantly different than what is capable with digital copies. That's why they are two independent markets with very distinct capabilities and the law and the digital economy button with books as everything else turns on the fact very key principles that the copyright owner owns the right to distribute their works in different formats and to distribute them under the terms for which they deem to be appropriate.

there is no market in control digital lending. But of course, there's no market in control digital lending. Control digital lending is predicated on infringement and the nonpayment of any fees.

this is so utterly transformative and so utterly, it's substantive. I'm sorry, it's utterly derivative is your position. Exactly, it's a derivative work. And there's, and it's doing nothing than repackaging and repurposing the derivative work.

let me focus for one moment on just a little bit of the evidence here on the commerciality of Internet Archive. Unlike most libraries, Internet Archive is owned by an individual largely who has funded this, and he, that Brewster Kale, as you well know. And every, virtually every page of the Internet Archive has a button that says you can buy this on better world books, which is giving an incredible amount of PR and certain revenue that drives between them.

What about the evidence that certainly the district court cited to it, admissions or undisputed evidence in the 56th statement about pitching CDL or pitching, joining the open library project as a way to save money. And are you relying on that at all as a basis to show? Absolutely, Your Honor. It's very rare in a record that you have actual admitted evidence that shows that a party is intending to supplant your market. And that's what we have in this record. The Internet Archive on this appeal tries to dismiss this as rhetorical flourishes. These pitches were made to hundreds of libraries and hundreds of slide decks provided to libraries with this exact same pitch. You don't have to buy it again.

ASTM versus public resource decision

then your argument actually is that, yeah, okay, there's a statute that talks about fair use, but there's a more specific statute that says when libraries can digitize books and that should control the fair use statute. Correct, your honor. That should control. And that's what was decided by this Court. They said that if you want to change the law, your job is to go to Congress. We're not in the position to change the statute towards you. In terms of what the statute says, the statute says, okay, you could do whatever you want with a physical book, but you can only create a digital copy for archival purposes or other limited purposes, but you're not allowed to distribute it. Correct, your honor. It does not envision in any way the practices of Internet Archive, which is digitizing literally millions of copies of books and making them available around the world to users.

So what was not recognized in the argument you just heard is that the copyright office and Congress over the last decade has repeatedly been approached to say, you need to think about the digital economy. You need to think about digital works. You need to think about the first sale doctrine and whether that should apply in the digital world. They have consistently rejected the changes to the law, both by the copyright office as well as Congress.

So if Congress had not codified the first sale doctrine, it didn't have Section 109 that authorizes libraries. It only had to rely on the fair use doctrine. Would it be obvious that you could do whatever you want with a physical book? Like would that, would libraries fall under fair use if we didn't have the first sale doctrine in the statute? Yes, I think it would, Your Honor. I mean, I think the Supreme Court, the physical books, the Supreme Court, is recognized that you have an unlimited right to distribute physical books. It's not simply because it was codified in Section 109.

I want to start by reframing and step back to really focus on the practical realities of what Internet Archive is doing and what is before this court. Internet Archive is asking this court to disregard the controlling law of this court as well as the Supreme Court. And what it is seeking is a radical change in the law that if accepted would disable the digital economy. Not just for books, but for movies, for music, for TV and the like.

You have not addressed the National Emergency Library. That's been sort of silent today. So given your statement now, you would agree that the National Emergency Library was a violation of copyright, because it wasn't one-to-one, correct? I would not agree. I mean, you were allowing multiple users to use the same digital copy of a hard book. The National Emergency Library does present different facts and different justifications you're under.

With constraints that impact the value of the library's ability to do that that are very much tied to the physical instantiation of the book, right? That's right. You can't rely on that one book to serve the serial needs of people globally because the costs of sending the book would exceed the costs of just getting another book on the other side of the world. I don't think I agree with that, Your Honor, and I don't think the record supports it.

Like the Wikipedia links where we are, where people are able to... Yeah, but the way that works is like snippet view, right? You can click on it and go to the particular part of the book. But if you want the whole book, you have to do it through CDL. Again, this is not... That's not really part of CDL, the Wikipedia links, right?

This is exactly what was going- this is exactly what the plaintiff said and what was going on in Sony as well. They said, well, you don't need to tape these movies off the air. We'll rent you a tape. We'll sell you a tape. You can get those benefits this other way just by paying us. You shouldn't be able to use technology yourself with the access you already have to get those benefits. That is the same thing that's going on here.

if they already have a physical copy and they want a circulated digital copy now, in the absence of your program, they would have to license an ebook. But once the program is available, they don't need to and they can just digitize or rely on you to digitize the physical book they have, right? This offers them another way of using the access they've already got, the right they already have to lend it to one patron at a time. In exactly the same way that the VCR and Sony allowed the person to access the material later instead of right now

You are right that the license terms that the publishers offer to libraries do not allow them to have electronic materials in their prominent collection, which is these libraries have print materials in their prominent collection. And if they want to use CDL as an alternative to rental, right, the overdrive scenario, they need to buy those books.

When they buy those books, they buy the physical copies to lend to their patrons one at a time or through an interlibrary change. They also buy e-books to make those available to their patrons. We're focused here on e-books and impacting e-licensing. I have a hard time reconciling those two, specifically as to e-licensing. Why would libraries ever pay for an e-license if they could have internet archives, scan all the books, hard copies they buy and make them available on an unlimited basis?

under factor 4 you say that actually there's one reason there's still be a market for e-books is because e-books are more attractive than digitized versions of physical books. Right? Because they have features and they're more user friendly or whatever. So what that kind of means is what you're saying is that your digital copies are more convenient or more attractive, I guess more convenient than physical books, but less convenient than e-books.

So that statute, Section 109, talks about you can lend out the physical copy, but then it also specifically delineates when you can make a copy of it or a digital copy and it limits when you can distribute that. So why wouldn't it conflict with what Congress has specified to say, well, this is really just the same as the physical copy.

That's right, and that is why we have been doing this without molestation by the publishers since 2011. But it's your position that you could lend it out during the first five years and that would still be fair use? That would be a different case, Your Honor. And we've very... I think the answer to that is... My only reason is if you're just doing it in your discretion. So the answer to that is we think it would... We think that would be fair use, Your Honor, because we don't think that would have a market effect either. There might... If they could show there was, or if the facts were different, that's why fair use is case by case, and if there were a case presenting those different facts, that might be different.

in the real world, there's a lot more friction in the sort of market for passing a paper book from one person to another. And I'm imagining that that's priced into the price of the paper book. Your premise is that a scanned digital version of that paper book is nothing more. It is tantamount to the same thing as the book. But we know there's a distinct market for those digital books. They're priced separately. So you're taking something from one market and you're inserting it into another market without ever having paid the premium in that new market.

If the forms are considered distinct things in their separate markets for them, why shouldn't the law recognize that converting the paper book into a digital book isn't just the same thing as passing around the paper book?

Google translate is generative AI

the various contracts or license agreements that publishers require libraries to sign to access research content for our users. And this is generally described as the problem of contractual override.

why training artificial intelligence in research context is and should continue to be a fair use

is the output copyrightable

let's go to the second question. Does the output

And then Singapore and Japan also have provisions and they basically allow the reproduction necessary for text and data mining.

So, let's look at the first question.

three different issues that are being implicated by artificial intelligence. And this is true with, you know, all artificial intelligence, not just a generative but particularly generative.

And one way we've seen artificial intelligence used in research practices is in extracting information from copyrighted works. So researchers are using this to categorize or classify relationships in or between sets of data. Now sometimes this is called using analytical AI and it evolves processes that are considered part of text and data mining. So we know that text data mining research methodologies can but they don't necessarily need to rely on artificial intelligence systems to extract this information.

the hardest part of doing this has not been the technology, it's not been the APIs, it's not been training engineers. It's been working with the practitioners with the change management doing it this way requires. Because when we do this recontextualization, when we cross the streams between various disciplines where we take the information that someone has spent their career cataloging and throw a 90% out of it to present it to an audience who doesn't care, that really, really irks people who have spent their entire career doing it in a way that they were trained is the best way you could do this. And some of the things that you can do with technology run into conflict with longstanding practices that were put in place before that technology existed.

The other cool thing that we've added is the knowing what records of change is really valuable, but the other question, really a more scholarly question is, how did it change? How has information changed over time? And so we built our infrastructure in support in Memento, which is one of the protocols behind the net archive, to give our data the equivalent of that way back machine, be able to say, what did that record look like a year ago? What did that record look like two years ago? To give you that audit log, so you can go and look at how that changed over time, and to give ourselves the ability to audit what changed when?

You need something that will help you show documents that will help you map to those understandings, those contexts that people bring to the information very quickly. And documents are also the way the internet works, right? You want to be able to use the sorts of affordances that the engineers understand. Rest APIs, JSON documents, cache control. Because those are things that engineers know that will let this stuff work fast. It will also let you hire people who know how the internet works and don't have to understand all the complex crazy stuff that we call children, heritage people do, they make it possible.

In a context based on the sort of context that people would understand, we say, this is the record about an artwork because you have a shared idea of what an artwork's information might be. It's probably not deep in the weeds around like Van Gogh had a brother named Theo Van Gogh and Theo Van Gogh was married and that wife turns out that's not part of what we described the artwork, that's part of the ecosystem knowledge around it. But graphs, on the other hand, are really optimized for asking those sorts of questions, for saying, I have a context, maybe I'm interested in siblings of artists and I want to be able to ask that kind of question.

But when we want to provide access to that data, what we end up doing is recontextualizing that data. We have to change the lens from that around someone who knows the discipline, who knows the form into one that reaches the way that users expect to see that data, which is often a really different shape to different flavor because their needs are different than the needs of catalogers. And so when we do that, we end up displaying records that are aggregations of data coming from multiple different systems because the tool that you would use to catalog the source is different from the tool you'd use to capture digital media, which is different from the tool you'd use to capture collections records, which is different from where you put the audio for the audio guide.

There's this research use case where there are people who are not looking for information, but they're looking for questions that haven't been asked, patterns that they haven't seen before. Things in the data that we don't already know and couldn't share with them, but that they could discover with their expertise.

Digital infrastructure isn't designed to make computers happy because computers aren't happy. We do it to empower people to be more effective in meeting their mission. And so when we think about who those people are, the reason we have the ecosystem we have is because we have many different constituents with many different kinds of needs.

over the past five years, what we've done is said, what if we take those sort of technologies, a single standard in linked art to linked things together, the vocabularies as connective glue, and used it to power our entire discovery ecosystem, including our archival collections, but also our museum collections on top of that. What if we used it also to power the audio guide that we provide to our visitors, or to provide interesting novel experiential interfaces on top of the collections that we have? What if we worked with third parties to use it, both large parties like Google, Arts, and Culture, and small projects like the Spanish Art in the US, which is a project of the cultural office of the embassy of Spain defined Spanish artworks in American museums and bring them together? And so we've worked across projects at all of these scales to pull records together and say, what would happen if you really tried to build this system out? So we end up with a unified API model, a way to access that data that spans across all of these collections and brings them together into a single data model.

American Art Collaborative. And this was a 2017 project to take 14 art museum collections together and use these sort of linked data principles that these things were built on to say, could you bridge across 14 institutions? Could you find connections? Could you provide a unified discovery environment for that many institutions at the same time? And what came out of that was a project that is called linked art. Rob Sanderson, who I'm sure many of you know, he's a regular here and a good colleague of mine. He and I worked together to create this data model called linked art based on the data model of the American Art Collaborative, which was the underlying sort of connective data tissue, building on years and years of work in the academic community under C-Dark to say, what if we had a tool that could bridge these things together?

So I'm not surprised, and perhaps the last question about user privacy.

I quite often would like to ask my research assistant to go and explore for me. Part of what they would be looking for is who is mentioned, what reference, what arguments do they mention, what school of thoughts. So this would be a very simple way and it would certainly make their lives easier.

we can see like there's a whole body of things that aren't working well.

So they say, I really love this tool. It is great how it attempts to relate the search query back to the content and the article. I find that I'm spending more time capturing and downloading the AI summaries than the downloading the PDFs.

And then we have, and this is a work in progress, we have automated mechanisms for assessing several different aspects of the responses.

So how does this work? I wanted to give this picture of what's actually happening behind the scenes, especially with this question and answer. So first, I will say that we're using a combination of OpenAI's GPT 3.5 to do this as well as some open source, smaller open source models to generate the vectors for the semantic search.

For any instance where, in most cases, the users will be able to trace back from the response to the point in the article where that information was pulled from. So you can see arrow pointing to the highlighted text that is the start of the segment that was used to generate that answer.

And in this on the side, you see we have this new chat box where the user can engage with the content and this very first action. The user doesn't have to do anything. They land on the page and as long as they run a search, we immediately process a prompt that says what in your voice, how is the query you put in?

you can see across the bottom we have a set of like preset prompts these are highly optimized to generate um responses that we've we've crafted so the first one is what is this text about this is a summary of the entire document

In approaching this material, suspend your disbelief, avoid choosing apreferred scenario, and embrace the full set of possibilities included in this material.Remember, the future will not be as described in any one scenario but will be made upof components of all four scenarios

AI not used tobenefit society’sbetterment,drivercapitalism

Alex was a couple years in now as a codirector of HIF with her partner, MITA. MITA hadbeen her assistant for several years ahead of their promotion, but at the time of thepromotion it was so clear that they were more partners than a boss and assistantrelationship. So, HIF made the decision to opt for the codirector approach and so far ithad been a rousing success.

The most advanced libraries operate almost exclusively on an AI platform

These emails — which I encourage you to look up — tell a dramatic story about how Google’s finance and advertising teams, led by Raghavan with the blessing of CEO Sundar Pichai, actively worked to make Google worse to make the company more money. This is what I mean when I talk about the Rot Economy — the illogical, product-destroying mindset that turns the products you love into torturous, frustrating quasi-tools that require you to fight the company’s intentions to get the service you want.

A grander sense of partnership is in the air now. What were once called AI bots have been assigned lofty titles like “copilot” and “assistant” and “collaborator” to convey a sense of partnership instead of a sense of automation. Large language models have been quick to ditch words like “bot” altogether.

Norman, now 88, explained to me that the term “user” proliferated in part because early computer technologists mistakenly assumed that people were kind of like machines. “The user was simply another component,” he said. “We didn’t think of them as a person—we thought of [them] as part of a system.” So early user experience design didn’t seek to make human-computer interactions “user friendly,” per se. The objective was to encourage people to complete tasks quickly and efficiently. People and their computers were just two parts of the larger systems being built by tech companies, which operated by their own rules and in pursuit of their own agendas.

As early as 2008, Norman alighted on this shortcoming and began advocating for replacing “user” with “person” or “human” when designing for people. (The subsequent years have seen an explosion of bots, which has made the issue that much more complicated.) “Psychologists depersonalize the people they study by calling them ‘subjects.’ We depersonalize the people we study by calling them ‘users.’ Both terms are derogatory,” he wrote then. “If we are designing for people, why not call them that?”

We often think of software development as a ticket-in-code-out business but this is really only a very small portion of the entire thing. Completely independently of the work done as a programmer, there exists users with different jobs they are trying to perform, and they may or may not find it convenient to slot our software into that job. A manager is not necessarily the right person to evaluate how good a job we are doing because they also exist independently of the user–software–programmer network, and have their own sets of priorities which may or may not align with the rest of the system.

I guess her own self-description that it doesn’t actually matter where she stops, that the important thing in the making of the painting is the making and destroying and making and destroying, that that’s actually what the whole thing is about.

There are, in my view, three stages of making art. One of them is the imagining, and the final one is the shaping. But in between, there is the judging, which is kind of what we’re talking about here, the editing.

Now, many people, when they read, listen to anything, when they take in media, they don’t necessarily even know where it was from.

The Work of Art, How Something Comes From Nothing

But the National Emergency Library (NEL) refutes that.

Properly understood, controlled digital lending simply enables modernlibraries to carry out their time-honored missions in the more efficient and effectiveway digital technologies allow.

Publishers’ claim that the scope of IA’s lending is too small to calculatemarket harm (Resp.Br. 52-53) is equally unfounded. To start, Publishers ignore thehuge increase in fixed costs—purchasing and storing books, building and expandingdigital infrastructure, and more—that would necessarily accompany (and imposelimits on) any expansion of controlled digital lending.

IA’s lending serves additional transformative purposes by enabling innovativeinteractions between books and the Internet, such as cite-checking online resourceslike Wikipedia.

Controlled digital lending permits libraries to build permanent collections andto archive and lend older books in a form that preserves their original printing.Publishers’ ebook licenses cannot serve this preservation mission because they arenot photographs of the original editions, and ongoing access depends on Publishers’discretion and is subject to change without notice.

Publishers erroneously claim that controlled digital lending does notexpand utility because their ebook licenses already provide the same efficiency.

In any event, IA’s controlled digital lending program serves a differentpurpose than physically lending the books it owns. Although both ultimately enablepatrons to read the content, that does not mean the purposes are the same—most usesof a book involve viewing its content. Controlled digital lending serves a differentpurpose by permitting libraries to lend the books they own to a broader range ofpeople for whom physical lending would be impractical.

But what matters is whether theborrowers are “entitled to receive the content.” Redigi, 910 F.3d at 661 (emphasisadded).

Publishers mischaracterize BWB’s ownership. Resp.Br. 17. As explained(IABr. 23 n.8), BWB is not owned by IA or Brewster Kahle, but by Better WorldLibraries, which has no owner. A-6087-89. BWB and Better World Libraries areoperated by a three-member board. A-6089. IA has no control over either entity.While Kahle has leadership roles in each entity, some overlap in personnel does notundermine the separateness of corporate entities.

Second, Publishers present an erroneously cramped view of libraries’missions. Libraries do not acquire and lend books solely to make them physicallyavailable to patrons within a restricted geographic area. Contra Resp.Br. 4, 9.Libraries provide readers more egalitarian access to a wider range of books,overcoming socioeconomic and geographic barriers by sharing resources with otherlibraries through interlibrary loans.

First, Publishers disregard the key feature of controlled digital lending: thecontrols that ensure borrowing a book digitally adheres to the same owned-to-loanedratio inherent in borrowing a book physically. Publishers repeatedly compare IA’slending to inapposite practices that lack this key feature.

Vaughn says the temperatures along with carbon dioxide levels have naturally fluctuated over earth's history inside lasting between 144,000 years. Well, over the last million years, co two in the atmosphere has never really gone despite its ups and downs never gone above maybe 280 parts per million. Until now. As of January 2024 the amount of heat trapping carbon dioxide is a whopping 422 parts per million. We've had a wonderful party with fossil fuels for a couple of centuries. We've, we have changed the world at a cost that's now only becoming evident.

Each ice core is kind of unique and shows you a different climatic window. Vaughn uses water isotopes to determine what the temperature was when each layer of ice was formed. Isotopes are molecules that have the same number of protons and electrons, but a different number of neutrons affecting their mass. For example, water h2o has oxygen that has either a molecular rate of 16 or 18. And so it's a heavy and light water precipitation that falls in warmer temperature tends to be heavier water. He says, but in colder air like at the poles, the snow that falls is generally lighter water by looking at these ratios of ice tops in ice cores. We were able to infer the temperature from when it fell. As snow.

many of those servers are of course honeypots some of them are staging and demo environments of the vendors it's interesting to not that this protocol and technology is also used for animals so many of the records on the internet actually are for cats and dogs and many of the records are exposed via universities or research centers that they just share anonymized data with other research centers

many hospitals started moving their dcom infrastructures to the cloud because it's cheaper it's easier it's faster it's good so they did the shift and they used the Legacy protocol dcom without sufficient security

dcom is the standard that defines how these images should be digitally structured and stored also it defines a network protocol that says how these images can be transferred in a network

109. On information and belief, in addition to her extensive online presence, she has aGitHub (a software code hosting platform) account called, “anarchivist,” and she developed arepository for a python module for interacting with OCLC’s WorldCat® Affiliate web services.

99. This includes metadata unique to WorldCat® records and created by OCLC. Forexample, the Anna’s Archive’s blog post indicates that Defendants harvested metadata that denotesassociations between records, such as between an original work and a parodying work. OCLCadds these associations data as part of its enrichment process.

1792. In total, WorldCat® has 1.4 billion OCNs, meaning Defendants claim that theywere able to harvest, to some extent, 97.4% of unique WorldCat® records

78. The bots also harvested data from WorldCat.org by pretending to be an internetbrowser, directly calling or “pinging” OCLC’s servers, and bypassing the search, or user interface,of WorldCat.org. More robust WorldCat® data was harvested directly from OCLC’s servers,including enriched data not available through the WorldCat.org user interface.

38. This includes adding OCLC’s own unique identifying number, the “OCN,” whichenhances queries and serves as an authoritative index for specific items or works.

1476. These attacks were accomplished with bots (automated software applications) that“scraped” and harvested data from WorldCat.org and other WorldCat®-based research sites andthat called or pinged the server directly. These bots were initially masked to appear as legitimatesearch engine bots from Bing or Google.

58. When an individual searches on WorldCat.org, the individual agrees to the OCLCWorldCat.org Services Terms and Conditions (attached here as Exhibit B).

51. The information available through WorldCat.org on a result page includes data thatis freely accessible on the web, such as title, publication, copyright, author, and editor, and limiteddata “enriched” by OCLC, such as OCN, International Standard Book Number (“ISBN”),International Standard Serial Number (“ISSN”), and pagination. This enriched data is moredifficult to find outside of WorldCat® and varies by each result in WorldCat.org.52. Most WorldCat® data available in a WorldCat® record is unavailable to anindividual on WorldCat.org. This is because a full WorldCat® record is part of a member library’ssubscription for cataloging and other library services.

39. Of the entire WorldCat® collection, more than 93% of the records have beenmodified, improved, and/or enhanced by OCLC.

OCLC is a non-profit, membership, computer library service and researchorganization dedicated to the public purposes of furthering access to the world’s information andreducing the rate of the rise in library costs

By hacking WorldCat.org, scraping and harvesting OCLC’s valuable WorldCat

In the blog post announcing their hacking and scraping of the data viaWorldCat.org, Defendants publicly thanked OCLC for “the decades of hard work you put intobuilding the collections that we now liberate. Truly: thank you.”

7. When OCLC member libraries subscribe to WorldCat® through OCLC’sWorldCat® Discovery Services/FirstSearch, the subscription includes the WorldCat.org service.Libraries are willing to pay for WorldCat.org as part of their WorldCat® subscription

6. To accomplish this, WorldCat.org allows individuals to search member libraries’catalogs as represented by their corresponding WorldCat® records in the WorldCat® database.When an individual views a search result from WorldCat.org, they see a more limited view of aWorldCat® record, i.e., with less metadata than is available for the record in the WorldCat®database for cataloging purposes

Proactively manage staff and user wellbeing: Cyber-incident management plans shouldinclude provisions for managing staff and user wellbeing. Cyber-attacks are deeply upsettingfor staff whose data is compromised and whose work is disrupted, and for users whoseservices are interrupted

Manage systems lifecycles to eliminate legacy technology: ‘Legacy’ systems are not justhard to maintain and secure, they are extremely hard to restore. Regular investment in thelifecycle of all critical systems – both infrastructure and applications – is essential toguarantee not just security but also organisational resilience

Enhance intrusion response processes: An in-depth security review should be commissionedafter even the smallest signs of network intrusion. It is relatively easy for an attacker toestablish persistence after gaining access to a network, and thereafter evade routinesecurity precautions

The need to embed security more deeply than ever into everything we do will requireinvestment in culture change across different parts of the Library. There is a risk that thedesire to return to ‘business as usual’ as fast as possible will compromise the changes intechnology, policy, and culture that will be necessary to secure the Library for the future. Astrong change management component in the Rebuild & Renew Programme will beessential to mitigate this risk, as will firm and well considered leadership from seniormanagers

a robust and resilient backup service, providing immutable and air-gapped copies, offsitecopies, and hot copies of data with multiple restoration points on a 4/3/2/1 model

The substantial disruption of the attack creates an opportunity to implement a significant number ofchanges to policy, processes, and technology that will address structural issues in ways that wouldpreviously have been too disruptive to countenance

our reliance on legacy infrastructure is the primary contributor to the length of time that theLibrary will require to recover from the attack. These legacy systems will in many cases needto be migrated to new versions, substantially modified, or even rebuilt from the ground up,either because they are unsupported and therefore cannot be repurchased or restored, orbecause they simply will not operate on modern servers or with modern security controls

some of our older applications rely substantially on manual extract, transform and load (ETL)processes to pass data from one system to another. This substantially increases the volumeof customer and staff data in transit on the network, which in a modern data managementand reporting infrastructure would be encapsulated in secure, automated end-to-end

our historically complex network topology (ie. the ‘shape’ of our network and how itscomponents connect to each other) allowed the attackers wider access to our network thanwould have been possible in a more modern network design, allowing them to compromisemore systems and services

viable sources of backups had been identified that wereunaffected by the cyber-attack and from which the Library’s digital and digitised collections,collection metadata and other corporate data could be recovered

The Library has not made any payment to the criminal actors responsible for the attack, nor engagedwith them in any way. Ransomware gangs contemplating future attacks such as this on publicly-funded institutions should be aware that the UK’s national policy, articulated by NCSC, isunambiguously clear that no such payments should be made

The lack of MFA on thedomain was identified and raised as a risk at this time, but the possible consequences were perhapsunder-appraised.

The intrusion was first identified as a major incident at 07:35 on 28 October 2023

The criminal gang responsible for the attack copied and exfiltrated (illegally removed) some 600GBof files, including personal data of Library users and staff. When it became clear that no ransomwould be paid, this data was put up for auction and subsequently dumped on the dark web. OurCorporate Information Management Unit is conducting a detailed review of the material included inthe data-dump, and where sensitive material is identified they are contacting the individualsaffected with advice and support.

Bobbs-Merrill Company v. Straus, the 1908 Supreme Court case that established the First Sale Doctrine in United States common law, flowed directly from a publisher’s attempt to control the minimum price that the novel The Castaway could be sold for on the secondary market.15 In that case, The Castaway’s publisher, the Bobbs-Merrill Company, added a notice to each copy of the book that no dealer was “authorized” to sell the book for less than $1. When the Straus brothers purchased a number of copies and decided to sell them for less than $1, Bobbs-Merrill sued to enforce its $1 price floor. Ultimately, the US Supreme Court ruled that Straus did not need “authorization” from Bobbs-Merrill (or anyone else) to sell the books at whatever price they chose. Once Bobbs-Merrill sold the books, their preferences for how the books were used did not matter.

Over the years, publishers have made many attempts to avoid this exchange, controlling both the purchase price and what purchasers do with the books after they are sold. For example, in the early 1900s, publishers tried to control resale prices on the books people bought from retailers by stamping mandatory resale prices on a book’s front page.6 (That attempt was rejected by the US Supreme Court).7 Publishers also tried to limit where people could resell books they bought, in one case claiming that a book sold in Thailand couldn’t be resold in the US.8 (That attempt was also rejected by the US Supreme Court, in 2013).9 These attempts failed because the publisher’s copyright does not give them absolute control of a book in perpetuity; the copyright system is a balance between publishers and purchasers.10 If publishers want the benefits of the copyright system, they also have to accept the limits it imposes on their power.

Moving scanning from the server to the client pushes it across the boundary between what is shared (the cloud) and what is private (the user device). By creating the capability to scan files that would never otherwise leave a user device, CSS thus erases any boundary between a user’s private sphere and their shared (semi-)public sphere [6]. It makes what was formerly private on a user’s device potentially available to law enforcement and intelligence agencies, even in the absence of a warrant. Because this privacy violation is performed at the scale of entire populations, it is a bulk surveillance technology.

Many scanning systems make use of perceptual hash functions, which have several features that make them ideal for identifying pictures. Most importantly, they are resilient to small changes in the image content, such as re-encoding or changing the size of an image. Some functions are even resilient to image cropping and rotation.

The alternative approach to image classification uses machine-learning techniques to identify targeted content. This is currently the best way to filter video, and usually the best way to filter text. The provider first trains a machine-learning model with image sets containing both innocuous and target content. This model is then used to scan pictures uploaded by users. Unlike perceptual hashing, which detects only photos that are similar to known target photos, machine-learning models can detect completely new images of the type on which they were trained.

In what follows, we refer to text, audio, images, and videos as “content,” and to content that is to be blocked by a CSS system as “targeted content.” This generalization is necessary. While the European Union (EU) and Apple have been talking about child sex-abuse material (CSAM)—specifically images—in their push for CSS [12], the EU has included terrorism and organized crime along with sex abuse [13]. In the EU’s view, targeted content extends from still images through videos to text, as text can be used for both sexual solicitation and terrorist recruitment. We cannot talk merely of “illegal” content, because proposed UK laws would require the blocking online of speech that is legal but that some actors find upsetting [14].

Less discussed than these broader cultural trends over which educators have little control are the major changes in reading pedagogy that have occurred in recent decades—some motivated by the ever-increasing demand to “teach to the test” and some by fads coming out of schools of education. In the latter category is the widely discussed decline in phonics education in favor of the “balanced literacy” approach advocated by education expert Lucy Calkins (who has more recently come to accept the need for more phonics instruction). I started to see the results of this ill-advised change several years ago, when students abruptly stopped attempting to sound out unfamiliar words and instead paused until they recognized the whole word as a unit. (In a recent class session, a smart, capable student was caught short by the word circumstances when reading a text out loud.) The result of this vibes-based literacy is that students never attain genuine fluency in reading. Even aside from the impact of smartphones, their experience of reading is constantly interrupted by their intentionally cultivated inability to process unfamiliar words.

And those light sticks aren't handed out as part of the event, they're mementos that fans will sometimes spend more than $100 on.

But the more advanced wristbands, like what you see at the Super Bowl or Lady Gaga, use infrared technology.

Let's start off with the simplest, RF wristbands, that receive a radio frequency communicating the precise timing and colors for each band.

it seems that one Wilshire has always had a crisis of location curiously the building isn't actually located at the address one Wilshire Boulevard it actually sits at 624 South Grand Avenue when Wilshire was a marketing name developed afterward

in 2013 this building sold for 437 million dollars the 660 dollars per square foot of leasable space that's by far the highest price paid of any office building in Downtown LA

Part of it is that old cardboard can't be recycled indefinitely. The EPA says it can only go through the process about seven times. Each time it goes through pulping and blending, the long, strong pine fibers get a bit shorter and weaker, and eventually the degraded paper bits simply wash through the screens and out of the process. So recycling is very important, but even if 100% of boxes got reused, making new ones would still mean cutting down trees.

The southern US, sometimes called "America's wood basket," is home to 2% of the world's forested land, yet it produces nearly 20% of our pulp and paper products.

Forester Alex Singleton walked us through an area whose trees were sold to International Paper two years ago. Alex: It has since been replanted with longleaf pine. Narrator: But it will still take decades for the new crop to mature. For many foresters, we only see a site harvested once during our careers. From this stage to there would probably be around 30 years.

Law enforcement contends that they want front dooraccess, where there is a clear understanding of when theyare accessing a device, as the notion of a back door soundssecretive. This front door could be opened by whomeverholds the key once investigators have demonstrated a lawfulbasis for access, such as probable cause that a crime isbeing committed. Whether front or back, however, buildingin an encrypted door that can be unlocked with a key—nomatter who maintains the key—adds a potentialvulnerability to exploitation by hackers, criminals, andother malicious actors. Researchers have yet to demonstratehow it would be possible to create a door that could only beaccessed in lawful circumstances.

Some observers say law enforcement’sinvestigative capabilities may be outpaced by the speed oftechnological change, preventing investigators fromaccessing certain information they may otherwise beauthorized to obtain. Specifically, law enforcement officialscite strong, end-to-end encryption, or what they have calledwarrant-proof encryption, as preventing lawful access tocertain data.

With respect to medical marijuana, a key difference between placement in Schedule I and ScheduleIII is that substances in Schedule III have an accepted medical use and may lawfully be dispensed byprescription, while Substances in Schedule I cannot.

In each budget cycle since FY2014, Congress has passed an appropriations rider barring theDepartment of Justice (DOJ) from using taxpayer funds to prevent states from “implementing their ownlaws that authorize the use, distribution, possession, or cultivation of medical marijuana.” Courts haveinterpreted the appropriations rider to prohibit federal prosecution of state-legal activities involvingmedical marijuana. However, it poses no bar to federal prosecution of activities involving recreationalmarijuana.

Congress placed marijuana in Schedule I in 1970 when it enacted the CSA. A lower schedulenumber carries greater restrictions under the CSA, with controlled substances in Schedule I subject to themost stringent controls. Schedule I controlled substances have no currently accepted medical use. It isillegal to produce, dispense, or possess such substances except in the context of federally approvedscientific studies, subject to CSA regulatory requirements designed to prevent abuse and diversion.

Cannabis and its derivatives generally fall within one of two categories under federal law: marijuana orhemp.

So we have 50 independent electoral systems that kind of work in conjunction in tandem, but they're all slightly different and they're all run by the state.

Images of women are more likely to be coded as sexual in nature than images of men in similar states of dress and activity, because of widespread cultural objectification of women in both images and its accompanying text. An AI art generator can “learn” to embody injustice and the biases of the era and culture of the training data on which it is trained.

“Information has become a battlespace, like naval or aerial”, Carl Miller, Research Director of the Centre for the Analysis of Social Media, once explained to me. Information warfare impacts how information is used, shared and amplified. What matters for information combatants is not the truth, reliability, relevance, contextuality or accuracy of information, but its strategic impact on the battlespace; that is, how well it manipulates citizens into adopting desired actions and beliefs.

Harvard had posture tests as early as 1880 and many other colleges would follow suit Harvard's program was developed by da Sergeant who created a template of the statistically average American and measured Harvard students in an effort to get students to reach a perfect muscular form women's college Vasser began keeping posture and other physical records in 1884 the ideal test was for a patient to stand naked in front of a mirror or have a nude photo taken and for an expert to comment and offer remedies for poor posture

Lord Chesterfield's widely published 1775 book lord Chesterfield's advice to his son on men and manners recommended new standards advising against odd motions strange postures and ental carriage

Santosh Vempala, a computer science professor at Georgia Tech, has also studied hallucinations. “A language model is just a probabilistic model of the world,” he says, not a truthful mirror of reality. Vempala explains that an LLM’s answer strives for a general calibration with the real world—as represented in its training data—which is “a weak version of accuracy.” His research, published with OpenAI’s Adam Kalai, found that hallucinations are unavoidable for facts that can’t be verified using the information in a model’s training data.

As with Midjourney, DALL-E 3 was capable of creating plagiaristic (near-identical) representations of trademarked characters, even when those characters were not mentioned by name.DALL-E 3 also created a whole universe of potential trademark infringements with this single two-word prompt: “animated toys” [bottom right].

Put slightly differently, if this speculation is correct, the very pressure that drives generative AI companies to gather more and more data and make their models larger and larger (in order to make the outputs more humanlike) may also be making the models more plagiaristic.

Moreover, Midjourney apparently sought to suppress our findings, banning Southen from its service (without even a refund of his subscription fee) after he reported his first results, and again after he created a new account from which additional results were reported. It then apparently changed its terms of service just before Christmas by inserting new language: “You may not use the Service to try to violate the intellectual property rights of others, including copyright, patent, or trademark rights. Doing so may subject you to penalties including legal action or a permanent ban from the Service.” This change might be interpreted as discouraging or even precluding the important and common practice of red-team investigations of the limits of generative AI—a practice that several major AI companies committed to as part of agreements with the White House announced in 2023. (Southen created two additional accounts in order to complete this project; these, too, were banned, with subscription fees not returned.)

One user on X pointed to the fact that Japan has allowed AI companies to train on copyright materials. While this observation is true, it is incomplete and oversimplified, as that training is constrained by limitations on unauthorized use drawn directly from relevant international law (including the Berne Convention and TRIPS agreement). In any event, the Japanese stance seems unlikely to be carry any weight in American courts.

Such examples are particularly compelling because they raise the possibility that an end user might inadvertently produce infringing materials. We then asked whether a similar thing might happen in the visual domain.

After a bit of experimentation (and in a discovery that led us to collaborate), Southen found that it was in fact easy to generate many plagiaristic outputs, with brief prompts related to commercial films (prompts are shown).

We will call such near-verbatim outputs “plagiaristic outputs,” because if a human created them we would call them prima facie instances of plagiarism.

Newspaper and magazine publishers could curate their content, as could the limited number of television and radio broadcasters. As cable television advanced, there were many more channels available to specialize and reach smaller audiences. The Internet and WWW exploded the information source space by orders of magnitude. For example, platforms such as YouTube receive hundreds of hours of video per minute. Tweets and Facebook updates must number in the hundreds of millions if not billions per day. Traditional media runs out of time (radio and television) or space (print media), but the Internet and WWW run out of neither. I hope that a thirst for verifiable or trustable facts will become a fashionable norm and part of the soluti

So, Sam decided, why not count the adverts he himself saw. It's just one number and applies only to the editor at a marketing magazine living in London on one arbitrary day. And what I saw were 93 ads I tried to be as open as I could about the fact that it's likely that I didn't notice every ad I could have done, but equally, I didn't miss that many, I don't think Sam also persuaded other people in the industry to do their own count. The most I've seen is 100 and 54. And I think I was quite generous. The lowest I've seen is 26. The most interesting version of the experiment was that I tasked someone to see as many as he could in a day and he got to 512 what a way to spend the day. And you will have noticed it's nowhere close to 10,000 ads.

Sam Anderson didn't marketing gurus have been making claims about advertising numbers for a very long time and Sam followed the trail all the way back to the 19 sixties. The very start of it was this piece of research by a man called Edwin Abel, who was a marketer for General Foods. Edwin Abel wanted to do a rough calculation on how many adverts people saw. He looked at how many hours of TV and radio people watched or listened to every day and worked out the average number of ads per hour on those mediums. So he multiplied those two numbers together to come up with this number. And this is our 1500 this 1500 ads a day number is still kicking around today. Often as the lowest of the big numbers in the blogosphere. And it is potentially a kind of legitimate calculation for the number of ads seen or heard, albeit from a quite different time in history. But there's some fine print to consider that is a number for a family of four. So if you divided that between a family of four, you'd actually be looking at something like 375 ads in this estimation.

One of the numbers we quoted was a high estimate at the time, which was 5000 ads per day and that's the number that got latched on to. So this 5000 number was not for the number of adverts that a consumer actually sees and registers each day, but rather for the clutter as Walker calls it, these large numbers are not counts of the number of ads that people pay attention to.

One of the arguments that they make is that the original parents' rights crusade in the US was actually in opposition to the effort to ban child labor. The groups really driving that opposition, like the National Association of Manufacturers, these conservative industry groups, what they were opposed to was an effort to muck up what they saw as the natural state of affairs that would be inequality.

Well, there are some people who have never liked the idea of public education because it's the most socialist thing that we do in this country. We tax ourselves to pay for it and everybody gets to access it. That's not a very American thing to do. Then you have conservative religious activists. They see a real opening, thanks to a whole string of Supreme Court cases to use public dollars to fund religious education. Then you have people who don't believe in public education for other reasons. Education is the single largest budget item in most states. If your goal is to cut taxes way back, if your goal is to give a handout to the wealthiest people in your state, spending less on education is going to be an absolute requirement. The same states that are enacting these sweeping school voucher programs, if you look at states like Iowa and Arkansas, they've ushered in huge tax cuts for their wealthiest residents. That means that within the next few years, there will no longer be enough funds available to fund their public schools, even at a time when they have effectively picked up the tab for affluent residents of the state who already send their kids to private schools.

As unregulated as these programs are, that minimal data is something we have access to. There's been great coverage of this, including a recent story in The Wall Street Journal by an education reporter named Matt Barnum, that what we are seeing in state after state is that in the early phases of these new programs, that the parents who are most likely to take advantage of them are not the parents of low-income and minority kids in the public schools despite that being the big sales pitch, that instead they are affluent parents whose kids already attended private school. When lawmakers are making the case for these programs, they are making the Moms for Liberty arguments.

The Heritage Foundation has been an early and very loud backer of Moms for Liberty. There, I think it's really instructive to see that they are the leader of the project 2025 that's laying out the agenda for a next Trump administration. You can look at their education platform, it is not about taking back school boards. It's about dismantling public education entirely.

I would point you to something like recent Gallup polling. We know, it's no secret that American trust and institutions has plummeted across the board, but something like only 26% of Americans say that they have faith in public schools. Among Republicans, it's even lower, it's 14%. Groups like Moms for Liberty have played a huge part in exacerbating the erosion of that trust.

mail trap

Papercut is a Windows application that just sits in the corner of your screen and your system notification tray and it's an SMTP server that doesn't send email every email that you send it intercepts

mailjet markup language

John Gilmore John is a very interesting person he's one of those people who I agree with everything he does right up to the point where I think he turns into a bit of a dick and then he kind of stops just past that point John was employee number five at Sun Microsystems he was one of the founders of the Electronic Frontier Foundation he is a uh I've seen him described as an extreme libertarian cipherpunk activist and uh the most famous quote I've seen from John is this one the net interprets censorship as damage and Roots around it if you start blocking ports because you don't like what people are doing the internet is designed to find another way rounded and you know he has taken this philosophy to an extreme he runs an open mail relay if you go to hop.toad.com it will accept email from anyone on the planet on Port 25 and it will deliver it doesn't care who you are doesn't care where you came from which is kind of the libertarian ethos in a nutshell

I went through to see how many email addresses I could register

in 1978 Gary Turk was working for the digital Equipment Corporation he was a sales rep and his job was to sell these the deck system 20. now this thing had built-in arpanet protocol support it was like you don't have to do anything special you could plug it into a network and it would just work and rightly or wrongly Gary thought well I reckon people who are on the upper net might be interested in knowing about this computer and digital didn't have a whole lot of sales going on on the US West Coast they had a big office on the East Coast but West Coast you know California Portland those kind of places they didn't really have much of a presence so he got his assistant to go through the arpanet directory and type in the email addresses of everybody on the American West Coast who had an email address 393 of them and put them now at this point they overflowed the header field so all the people who got this email got an email which started with about 250 other people's email addresses and then right down at the end of it it says hey we invite you to come and see the deck system 2020

one person whose Innovation is still a significant part of the way we work with it was this guy it's Ray Tomlinson and he was working on an opponent Mail system in 1971 and Rey invented at Rey is the person who went well hang on if we know the user's name and we know the arpanet host where they host their email we could put an at in the middle because it's Alice at the machine