 I'm Saïd Cherazery, I'm an Associate Dean in the Libraries at Johns Hopkins University. I lead the group called Digital Infrastructure, Applications and Services. I keep expecting all of you to organize into a bunch of squares on my screen, so it's really great to see you not in that mode. Normally, of course, we'd see N.I. having been virtual for some period of time now. It's great to be back in person. And I would have hoped to present some of the things we're going to talk about today in smaller chunks. So I apologize if this feels like a little bit of a fire hose today. This actually covers about two and a half years of things that we've been working on at Hopkins. It's a very interesting journey, at least I hope you think so. And while we're talking about open source programs offices, and if you heard Josh Greenberg's remarks earlier today, this is somewhat of an administrative sort of bureaucratic view on what I'll be discussing. But more than anything, what I hope you can take away from this is open source as a way of working with communities directly to make all sorts of impact. And I'll touch upon that throughout the talk. So a lot of this actually started from a conversation I had with an alumnus of Johns Hopkins named Jacob Green. He's a member of our advisory board for the Computer Science Department. And he and I had a very long talk in early 2019 about open source software. And the point that he made is open source software is now being used by companies throughout the world. Not just tech sector companies, but companies at large. And there's impact that you can see from the private sector. And why isn't this happening in higher education? Which I thought was a good question. Particularly since so much of open source actually started in higher education. And I will give a shout out and a congratulations to John Coonsy, who some of you may know. He sent me an email saying that he is retiring. And in that email, I saw just an amazing amount of things that he has done over his career, including writing some, this I did not know, writing some software when he was a Berkeley that now runs on everyone's Macs. So including the one that I'm using here. So perfect reminder of how higher education actually started a lot of the things that the private sector has picked up on. So one of the ways that we are trying to address this particular question is through what we're calling the open source programs office. And we've launched one of these at Johns Hopkins. It's the first one in a U.S. university. Now open source programs offices have been used in the private sector by companies for some years now. They're organized in something called the to-do group, which is run out of the Linux Foundation. And most of those OSPOs in that setting not surprisingly are focused on things like compliance, licensing, engineering practices, so on. Those things are of course important in a university context. But we're trying to take that foundational OSPO model and expand it and augment it for a university setting. There's actually some effort looking at this also in a government setting. And one way that I think about this is that, you know, a company like Microsoft and universities like Hopkins has an HR department, has a legal department, has a CIO. And many of those functions span both those organizations. But many of them are different, right? Hopkins has different issues in terms of legal or HR or the CIO's office. We have students. We have privacy issues that a company might not and so on. So the OSPO in a university setting is that similar kind. It's spawned, if you will, or inspired by what's happening in the private sector, but we have to think about it in a very different way in the university context. And in particular, there are three things that I think of. I often think of these as the pillars of universities. Research, education, and translation. And by translation, I mean how does the research and education of a university be extended beyond its walls, you know, both physically and virtually? So keeping those themes, research, education, translation in mind, you know, throughout the talk. So one of the first things that I did in terms of this OSPO was how do you activate different parts of the university, you know, to engage them, to get them interested, to find out who's actually doing things around open source. If you ask people in your university who care about open source software, you get different responses. But if you ask everyone, do you care about research? I think you'll get a uniform response there. Yes, that is important. So one of the first cases that I made was open source software is a primary research object. And what you'll see in this slide is on the left are articles and in the middle are data and then on the right are software. And interestingly enough, not only are of course lots of people interested in open source software, I would submit we can learn more about data by looking at software than we can at articles. The connection between articles and data tends to be a mechanical, sort of a citation type link, whereas the link between software and data is much richer. You learn how the data are processed. You learn much more about the inferences or assumptions around the data. And we're seeing evidence of that from researchers and Hopkins on the medical field. There's a community called the Organizational Health Data and Informatics Group. They are creating open source tools and software through which they gain insights into kind of the data they're using. And the Hopkins principle for this is someone named Paul Nagy. And when I spoke to him about it, he said, I don't want to sign any more data use agreements. And I said, what do you mean by that? He said, so typically what we do with our data use agreements is we try to anticipate all the types of data, all the types of uses. We sign the agreement and as soon as we do, we see from the other institution, oh my God, there are all these other types of data and other types of uses we didn't anticipate. God help us if a third institution wants to get involved and so on. So they're using the open source tools, the platforms, as a way of gaining insights into the data. I'm not saying that eventually there won't be agreements, but they feel like this is a very good first step before actually sitting down to write the agreements. So what I'm trying to show in this slide is that type of connection between, you know, around data. And if you look to the right side of that slide, you see those arrows going off. To me, that's the translation piece, right? So the first sort of articles, data, software are very much the objects, the currency, the assets, whatever you want to call it, the outputs, products, lifeblood of universities. But the translation piece I would submit is one where we have not succeeded when it comes to the outputs of what we produce. Some place, one place you can go for translation in your university is the transfer office, but that's a very particular kind of translation, typically around commercialization. I'm talking about different kinds of translation. I'll give you some examples over the course of the talk. And one of the distinction to keep in mind is if you look at those first three boxes and the arrows, those are the project phases of working on research. And that translation piece is when you move to having a product. And I'll make two important distinctions here. So one of the people I've been working with on this journey is Stephen Wally. He's an executive at Microsoft in the Azure office. And he basically said to me recently, the project phase is when you're dealing with a technical debt, right? And when you move it into a product, we're actually moving in into the kinds of services and utilities that people have. Yesterday was an interesting day if you have any kind of platform that was affected by log4j. Anyway, I'll just leave that there. What I found interesting is my group runs a platform and I'll talk about it called the public access submission system. We dealt with that debt, right? We dealt with all the issues around this vulnerability. We also work with a platform. I'll talk about Lutas, which is run by the city of Paris. And in the midst of looking at, I don't know how many of your tickets yesterday, I get an email from the folks in Paris saying here's what you do to deal with log4j with Lutas and I passed that on to our team and we did it within, I don't know, 15 minutes. So when you have technical debt, are you dealing with it or is someone else dealing with it? This is one of the distinctions of the project versus product view of software. Another is that if I showed up at a researcher's office at Hopkins and told them you need to use this open source license and you need to do things in this way and you need to have this kind of practice. It'd be a pretty short, unpleasant conversation. That's not the way things work. In the project phase, the grant phase typically, researchers need autonomy, need freedom, need flexibility. But if they want to have translation, the choices they make have consequences downstream. So what we've argued is we're not here to tell you what to do. But if you choose this license, be aware that it's going to be hard for you to have this kind of translation or it may be easy for you to have this kind of translation. This has been a seminal moment with the tech transfer office. They have not known how to crack this not proverbial not. They show up and say we're here to help and the researchers say don't tell me what to do. You don't tell me where to publish my papers. You don't tell me where to put my data. Don't tell me how to manage my software. We're not here to tell you what to do. We're just here to tell you the consequences of what you're doing and think about the downstream effects. So a few cases that I want to share with you. One is this public access submission or past system or past. I've talked about this previous CNI. I'll just say very quickly it's fundamentally about aligning open access and public access policy compliance by allowing researchers to simultaneously deposit their articles into your institutional repository and into PubMed Central. We have a recent grant from NSF to look at also how they may be submissions into NSF's public access repository. So one of the key things about past is that we create a package of metadata and metadata around authors, articles and eventually data to do some of the kinds of LinkedIn explorations you may have seen from the presentation from DOE. But in this context, you know, what I'd like to say about past is as we're exploring going from, you know, a Hopkins project to a more of a community project, we receive some very good feedback that if this is a platform that's run out of Hopkins, you're not going to get a lot of adoption, right? There has to be some neutral entity to do that. This person I mentioned, Stephen Wally, has this amazing metaphor to cooking about software development. And what he basically says is, you know, if you cook a meal for yourself, you know, that shouldn't be too complicated, right? We should be able to handle that pretty easily. If you're cooking a meal for someone else, it'd be polite to ask, right, you know, what their restrictions or preferences might be. If you're cooking for a dinner party, right, you'd also do the same thing. But you can understand it's a slightly different way of cooking, right? And then imagine that you're a caterer. You don't ask every single person what they want. You sort of have a set of menu choices and you start to customize and so on. And then imagine if you're actually running a large sales set of restaurants, right? And then imagine if you're doing industrial food production. So think about software development and all those different kind of categories. And where do your platforms sit in that metaphor? I'll turn the mirror on myself. I'd say past is like a dinner party, right? But that's not how you provide a service for the community. So based on feedback, we went to the Eclipse Foundation and basically said, we want you to help us go to that industrial food production model, right? Eclipse has a very extensive playbook. Project Cat and Book that we have gotten to know. In the last few months, we have explored how past can be redesigned, re-engineered in a way that I was not familiar with. And we made some very difficult choices based on the feedback from Eclipse. How's your community going to react to this? What about adoption? What about technical debt? What about maintenance? And we talk about specific examples if you want. Stephen also talks about this powers of ten phenomenon. Right? If you have a thousand people interested in your software, maybe a hundred will make some kind of contributions. Ten will make really substantive contributions. Maybe they'll even write documentation for you and then you'll get that one power developer or power user. So if you want to have multiple power developers or users, you have to scale. And that's what we've been doing with past through this Eclipse process. And I think there are a lot of really important lessons there for all of us as a community in terms of getting away from the dinner party or the catered event to something much more large scale. A second example I'll talk about is Lutess. This is an open source municipal services platform developed by the City of Paris. Over several years and hundreds of millions of euros over those years, it is used to provide hundreds of digital services to the citizens of Paris. Things like you might imagine like filing, you know, 311 tickets or filing for permits or paying your parking tickets, things like that. But also some really innovative features like participatory budgeting. So 5% of Paris' budget is actually allocated through direct citizen input through Lutess. Every year they use this platform to gather ideas, gather feedback, and then they make allocation decisions to that 5% based on this platform. So I mentioned Jacob Green, the Hopkins alumnus who asked me that provocative question. He made this connection to the City of Paris around the work we've been doing with Lutess. And basically Paris is also looking to move further into that industrial scale kind of model, if you will, and building a community of developers and of users. So I'm still not exactly sure how, but Jacob managed to convince them that the next place you should try to use Lutess outside of France is Baltimore, Maryland. So that is what we've been trying to do. We did go to Baltimore City directly. Jacob and I, Jacob set up a meeting at the time, the head of innovation for the City of Baltimore, and we talked about this. And there was a lot of interest and there was a lot of engagement. I don't know about your city, but Baltimore officials seem to be just dealing with lots of different fires all over the place. So instead, we are now working with a local community center, the St. Francis Neighborhood Center in West Baltimore. And this is a really incredible community center, just top-notch leadership, long-term relationships with the community. They've been there since 1963. Amazing results, individual stories of transforming people's lives and also community-based impact as well. So we went to St. Francis and said, do you think this could be useful? And they said, we are working on a digital strategy right now. And we are actually building a physical lab, what we call a smart center, where we're going to have laptops and we're gonna have training and so on. We think this platform will be wonderful for a lot of the things that we need to do as part of that strategy. And I'll say that it again goes beyond this sort of transactional types of activities. One of the things that Lutess has is a module for homeless shelters, okay? And one of the questions that St. Francis and other community centers get often is where are their beds available in a homeless shelter this evening, right? And the director of this neighborhood center, Christy Green, told me that a lot of that is done by phone, right? She will call directors of other community centers. She'll know. We can do better. We can use a platform to do this. And there's a group of Hopkins students who actually did an investigation into this question and then found out you can't just ask about whether there's a bed available or not. You have to ask questions like, is it only for women? You know, sadly, do you take kids? I have a medical device. Homeless people have medical devices. I need to be able to plug it in, things like that. These are requirements and these are contributions. They're not committing code, but these are contributions. We did have students write code as well, right? They actually did lines of code, fingers on the keyboard, but this is a whole other set of contributions that I think are important to acknowledge. We folded all of this in to the work we're doing with Lutas and with St. Francis, right? And one thing that's really important to keep in mind, St. Francis has a relationship with this community that Johns Hopkins never will. And I do mean never will. There are more people in this community who do not trust Johns Hopkins than there are who don't, who do. And they have good reason for it. And this is not unique to this community center. This is not unique to this university and it is not unique to this city. So by working with St. Francis and having them say this is our work, this is our platform, here's how we're going to use data about you. And it's an open platform. You can look, you can see, and if you don't like us, tell us. And we have people at Hopkins working with us. We are writing the requirements and they're responding to them. They're not showing up and saying, here's what you need to do, here's what you need to use. This is incredibly important. And think about this work around homeless shelters, right? Suppose we have a network of these Lutess platforms throughout the city. We can go to Baltimore City now and say, can you organize data from homeless shelters as a feed to this network of platforms, right? That's an easier ask than Baltimore. Can you go buy a platform, lease a platform, procure a platform, manage the platform, hire IT staff? So this is the way you can see the communities and the university and Paris coming together to do this kind of work. So this has been something that Paris has shared far and wide. We had an amazing event in November to showcase a demonstration of this. It's really important for me to mention one thing here. Universities typically don't have a lot of ways to work together. And that might sound strange, but I mean from a legal perspective, right? Not so secret. One of the reasons we sign grant agreements, data use agreements, MOUs, MOAs, BAAs, pick your acronym, I'm sure I'm forgetting a few, is risk mitigation, right? If you want to engage an administrator about signing an agreement, talk about risk. So we did all of the work that I've just described under the BSD license of Lutas. And I had several conversations with the university about what's the risk, go look at the license, who owns the IP, go look at the license. Why should we trust this? I don't know. It's a canonical license that the open source initiative is sanctioned that hundreds of companies and thousands of projects have used. Microsoft and Google use this. So if you can identify a risk that they have not, by all means, tell me. And by all means, go to the open source initiative and tell them that this license needs to be rewritten. That isn't gonna happen, right? How difficult do you think it would have been for Hopkins, St. Francis, and Paris to sign an MOU? I'll just let you ponder that for a second. Okay, I would submit we'd still be sharing drafts of what are for monster document I'm talking about rather than demonstrating this platform being used at St. Francis. So let me shift a little bit. These projects are still relevant to this story, but this is more on the education side, right? So I can't thank Jacob in particular, but Stephen Wally as well, who's been incredibly generous with this time. He taught a course this fall. It's just ending called Semesters of Code. It's not the official name, but it basically builds on these kinds of efforts at other universities. And as you might know, Google Summer of Code and Hackathons and so on. But Stephen's fundamental premise really is that the best way to teach students about sort of robust software engineering practices is to use open source platforms. Is to give them all of the context, not just teaching them engineering, but the history, the legal parts, the economic parts. He's a great presentation about how open source software is not altruistic. But there are economic arguments you can make for it being a better choice than building your own or leasing something from someone else, right? These are the things he's been teaching the students over the course of the semester. But in parallel, have them work on open source software projects with developers from those projects as mentors. So PASS is one of those projects. LUTES is one of those projects. Semesterly, which is something written by students at Hopkins. Open Carvat, which is a genomics project funded by NIH and Microsoft PowerShell are all actual projects that these students worked on. And I was on the Slack channel where the mentors and the students were exchanging with each other. And as you might imagine, students began out by thinking, what the hell, what is this? Well, I'm really supposed to work with these platforms directly and hear the developers and they're gonna be, I don't wanna be embarrassed and I'm sure they'll think I'm silly and so on and so on. And Stephen ran a really great workshop with these mentors before this course and we're blessed, I am blessed to have software developers who are exceptionally good communicators and basically mentored these students and by the end of the course, and the metric we used by the way was not did you make a commit, not is the software gonna be better, but where were you day one and where are you now? What's the progress, right? Every one of the students just, it was transformative to watch how this happened. Three of the students want to keep working on these projects as an independent next semester. One of them had to drop the course because he was worried he wouldn't have the right graduation requirements, yet he kept working on the project anyway. There's a level of engagement here that I don't think we've seen quite often. So Stephen is looking at this as something we can continue, maybe at Hopkins certainly, but other universities, right? One of the big challenges we're all going to face is how do you engage people who aren't physically present? So imagine semesters of code as a way of doing that, right? And I'm not talking about, there's an equivalent here of the administrative handshake, right? I'm not talking about students at your university registering at Hopkins or vice versa. I'm talking about students at your universities making commits. You can do that. You don't need to register for a course, right? And your institution might figure out how to give credit for that kind of work. So this is a very interesting model that we may be able to look forward for this kind of cross institutional type of work. And mind you, the mentors put in a lot of effort. We didn't get substantive, if I can use that word, sort of, commits or contributions, but all of the teams said that the software platforms are much better ready now to onboard new people. What we had to do to help the students is going to help with anybody who wants to make a contribution. And I just want to take a little bit of a, maybe it feels like a little bit of a turn in the humanities because I think Cliff made some interesting comments about scientists versus humanities or humanists. And I also want to come back to this point about grants. I said this at lunch yesterday, particularly to any of the program officers in the room, I am not dissing grants, okay? I think they are wonderful things. I'm just saying they are one way in which people can work together. And interestingly enough, what's happened with the story of Lutess and St. Francis coalesced into a grant. So we have a planning grant from the Mellon Foundation to do work around black digital humanities with St. Francis, with researchers at Hopkins. And these open source platforms, including Lutess, are a key part of the work that we're doing through this grant. So there's a very interesting effect of sort of coming back to the more traditional or sort of familiar ways of thinking about these kinds of ways we work together. Someone I really, really admire, Zaynep Tufichki, and I'm sorry for not pronouncing that properly, was interviewed on the Ezra Klein show. And this is a quote or something she said about systems thinking. I have a background in systems engineering. So it resonates with me, but I think you can read it and see why it's so interesting. This idea of looking at something as a whole and its interactions is really at the heart of open source software. The community is the one who looks at it as a whole. The contributors are the ones who look at the interactions, right? So there's a very strong relationship between open source software and systems thinking. And I'll go even a little bit further back in time and maybe grander. Some years ago, I went to the ACMH Human-Computer Interaction Conference and Doug Engelbart, who some of you may know as one of the founders of the internet, was the keynote speaker. And it's a terrible Google term to use, but if you Google mother of all demos, if there's a better term, please tell me. Mother of all demos, it's this demonstration that Engelbart and his colleagues did in the 1960s, and it is just unbelievable to see the vision in this demo. They are demonstrating hypertext, collaborative document authoring, embedded video, audio, so on. It's just mind-blowing to see the vision that's shown in this thing. And so Engelbart gave this talk and then was asked to question someone in the audience where he basically said, when you did that demo in the 60s, could you have imagined, and this was in the late 90s, could you have imagined we'd be where we are today, the web and all these kinds of things? And Engelbart paused, he looked down, and I think more than anything, he was just trying to sort of gently broach his answer to this question. But he just came out and said, actually, I'm completely underwhelmed, which I'm sure this poor person wasn't looking for that answer, but it was very illuminating, right? It was very telling, and this is what he was saying. He said, I didn't think we were going to use it to sell widgets. I didn't think we would do this to get Facebook, right? He wanted to work on collective intelligence to solve humanity's most important problems, right? And we can all reflect on whether we think that's where we are today or not. So, turning my attention to this more sort of impact side of things, I recently said, you know, COVID really sucks. And somebody said, thank you, thank you, Captain Obvious. So, COVID affected every country on the planet. I had the great privilege, I'll put it that way, to play a small role in supporting the JHU global COVID map. And I've looked at that map many, many times and seen the slack channels behind the scene. Every country handled it differently, right? I'm not sending the effects for the same, but every country reported cases of COVID, every single one, every one of them had to think about what do we do? How do we test? Do we wear masks? Do we shut down? Do we vaccinate? How many other events affect every single country on the planet, right? The other thing is we shut the planet down, right? I told somebody this yesterday, I live close to campus Hopkins, there's a street called Charles Street, it's the most major North, South thoroughfare in the city. And right near campus, it's very busy, right? There's always traffic there. So during the shutdown, I was walking outside and I heard a sound that Charles Street had never heard before. I turned around and it was a deer. This deer was just leisurely walking across Charles Street, right? And kind of gave me this look like, what? Why are you looking at me? It was a surreal moment that just led me to realize we've done something we've not done before, right? So when the 1918 flu happened, you didn't get your food through Instacart, right? You had to go to the store. And we are seeing people on boxes on the screen. You didn't do that. You went to work. There was no remote work. So there's some fundamental differences between what happened a hundred years ago versus what's happening today. And a planetary reboot is not something we've done before. So if you're working at an institution that's trying to go back to normal, I'm sorry, good luck, there will not be a normal. We're not going back to the way things were, right? Look at the supply chain issues we have, right? One of my colleagues is our maps and GIS librarian has a fascination with sort of food and how food development happened. And there's this great story around Orida, which I did not realize is a cross between Oregon and Idaho, a company that basically produced a lot of potatoes for the military during World War II, right? Was churning out lots of potatoes baked potatoes, french fries, water for the military. So the war then ends, right? And they have a bunch of potatoes and they have a big supply chain to deliver lots of potatoes. So I won't give you the whole story, but Orida came around because someone had a sort of inspiration to push potatoes through farming machine and create those little Orida nuggets and then went to a big food convention and passed them around and that's how Orida became a thing. So I'm telling you this because a lot of the supply chain issues we're dealing with are because we don't know how to adjust again. All the demand patterns suddenly stopped and now are suddenly shifting and you have large numbers of people who don't wanna go back to the really, really terrible jobs they had, right? So these things are not gonna just sort of magically readjust and there will be transformative kinds of changes. So picking up on Cliff's theme of these kind of changes of innovation, changes of the community and innovation of necessity and innovation that's maybe more transformative. Just a little bit about some things I think are gonna happen in 10 years around artificial intelligence, okay? And I'm not an expert in artificial intelligence or machine learning, but I did do some work on the grad school and I certainly hear from very smart people how Hopkins would do. So machine learning from what I can tell is a subset of AI, right? It's about data being organized in a certain way and then being analyzed, processed to help people make decisions, to help us do things we do now better. AI is a leap, right? It does that, but you can use data in different ways, unstructured ways to make decisions, right? So human beings are not necessarily part of that conversation. And one thing that's really worthwhile sharing, these algorithms have not changed much, right? I used to use genetic algorithms when I was in grad school, all right? They're still used today. Anybody care to guess when John Holland's seminal paper on genetic algorithms came out? I'll give you a hint. It's before 2000, 1975. Now, to be clear, I was not using them in 1975, okay? But they have not changed much. It's the large amounts of data that exists now that allow these things to go to town, right? There are people now studying the foundational mathematics of these things and trying to figure out, but I distinctly remember a conversation in grad school with my advisor. I said, I ran this thing. Here's the optimal solution. I don't understand how it got here. And he said, yeah, don't worry about it. That's fine. So these things are still somewhat mysterious, but we're getting better at learning them and understanding them. So I'm gonna make a few predictions about what AI might do for us in 10 years. And I can do this because what are the odds that all of us are gonna be together in the same room in 10 years that you're gonna remember I made these predictions, that you remember there may be a recording of it? And to close point, there's video debt, so you probably haven't seen it, or it may not be preserved. So with that in mind, I make these predictions. In 10 years, we will have completely autonomous driverless truck fleets, right? So the trucking industry is really, really beleaguered right now, because they can't get enough people to drive trucks. And what is a pedestrian response? One of the complaints is, I don't wanna leave my family, right? I don't get paid by the hour. I get paid by the load, right? So it's a terrible model. So they're hours and hours days away from their family. So the trucking industry said, how about we give your wives commercial licenses so they can go with you on these trucking journeys? They interviewed one trucker who paused. He said, that's not the family I'm talking about. So a more sort of technological approach will be driverless trucks. And in this exchange I saw, somebody from the union said, you tell me if you think in 10 years a truck's gonna be able to drive through downtown Chicago. And I'll tell you, I don't know that that's true, but I will tell you is a truck will be able to drive to someplace very close to Chicago where a bunch of packages with RFIDs that have a bunch of information about those packages will be picked up by drones and delivered to people in downtown Chicago. So you don't have to go to the doorstep. You just have to get close. Another prediction is that human resources, and I recognize the irony of that, decisions will be made by AI. And note that Amazon is already using AI to hire and fire people. I'm not saying that's a good thing. I also think in 10 years, the transactional documentation type aspects of project management will be done by AI. And then in 10 years, the majority of experiments, maybe all of them in STEM fields will be designed by AI. And for humanists who use data, it will be as well. Human beings are always going to be harder to predict than physical systems. I get that. But think in large numbers. If you're a fan of the Foundation trilogy, you're thinking of Harry Selden. And last one, most of the code over 50% will be written by AI. And why it won't be 100% is we're still gonna need humans to generate the ground truth. So ask yourself, let's say I'm wrong about all these, right? But you come up with your own list. Ask yourself if you want these developments to happen in open, in an open way, or in a closed way. So I'm gonna let you read this because the other point that Cliff made that I wanna pick up on and that we've been exploring over this journey is the power of communities, right? I'm sitting here in unceded land, which according to ARL, is where the Piscataway, the Pamunki, the Nentego, the Mataponi, the Chica-Homini, Monacan and Pahatin cultures thrived. I'm sorry if I didn't pronounce any of those correctly. But I want you to look at this wisdom from the Hopi in modern day Arizona. This is about the power of communities, right? So my father was a political scientist and I know less about political science than I do about machine learning. But from thousands of conversations with him about these kinds of topics, he seemed to be interested in how people in an institution organize large numbers of people, right? He studied Islam as a political force. How did it organize people? He wrote the Constitution for United Bacchuslam, how do nations organize people, right? If he were alive today, I'd ask him what's next? Religion, nations, communities? So institutions are being questioned right now, right? In fundamental ways. Sometimes for really terrible reasons. But sometimes for good ones too. I was in Paris when the COP26 climate change somewhat was happening and I saw news from all over the world. Very strong theme. Why are the communities who are actually doing something about climate change and not speaking at this conference? Why are you going back to the same people who made promises 10 years ago that they did not keep, right? There was a conference recently at Hopkins where Dr. Ibn-Ex-Kendi spoke. He said the same thing. He said the only way we're going to be anti-racist and deal with racism is to go to the communities, right? Is to go directly to them and treat them as partners, as equals, as we're trying to do with St. Francis, right? And let me give you a story that I think is relevant in this regard. The United Nations developed software by working with a university in Sweden. And I'm told this university is like the MIT of Sweden, right? And the software was around analysis for the electrical grid, right? So the resolution of the software was 10 kilometers by 10 kilometers, 100 square kilometers. It took 20 hours to run, right? They issue one of their open source challenges. Open up the code, prepare it, tell people, please, how about it? See if you can do better. One month later, a single woman in India submitted code and they looked at it and discovered that the resolution was much finer, much more detailed, 2.5 kilometers by 2.5 kilometers. But the speed of the processing was faster. It was 30 minutes. That is phenomenally impressive if you've ever done that kind of modeling. So they were so impressed. They brought her to the United Nations in New York. This is pre-COVID. She met the secretary general. Apparently on her way out the building, she and her dad, not a dry eye in the house, walked out and she said, this is the greatest day of my life. There is talent everywhere, right? Everywhere. And it's in individuals, it's in communities. She is a power user, according to Stephen Wally. They discovered this without the full-scale kind of industrial model I talked about, but the UN has a considerable pull, right? An inconsiderable brand and considerable ability to attract large numbers of people to work together. We're gonna have to find a way to engage our communities together, to find this talent, to not assume it has to be credentialed, by the way we usually do through universities, if we're gonna solve the kinds of problems we're facing. If you wanna look backwards at how some of these things have happened. There's a book from David Graber, The Dawn of Everything, where he questions this idea that we went from hunter-gatherers to farms to cities and that's the only way to organize large numbers of people. He says, it's bunk. We've tried lots of different models over the past, some of them seasonal, some of them different, some of them familiar to what we have today. If you wanna look forward, donate economics, like Kate Rowerth. There is an active movement about refreshing recycling at the local level. I'm gonna start to try and give you some more practical, actionable types of things to do, right? I'm certainly not gonna revise code from the U.N. the way this incredibly talented woman did. Ospo plus plus is a network that's been formed around universities and government ospos. I'll have the website address at the end for you to look at. This is a group that's come together to basically create a network of these ospos. And we think this network of ospos can be the way through which universities and governments and companies can work together, but also work with local communities. Also start to do this. Open source is a way of working together rather than sort of an object. Open source is a way to sort of achieve this goal of systems thinking or the collective intelligence that Doug Engelbart spoke of. And one of the things that Ospo plus plus has done recently is lead an effort to update one of the sections of the toolkit from the National Academy's round table on aligning incentives for open science. So that toolkit was just released recently. The Ospo plus plus group had a series of efforts with people from industry, from universities, from government, including from Europe to take that original code and software section and create a new primer that's focused around open source software. This is really intended to be a document you can take at your institution and do something very practical in terms of moving forward with open source software. There's a recommendation there to create a network of 20 Ospos within the US, primarily university-based. This mirrors a recommendation from the European Commission to create a network of 20 Ospos in universities and governments in that case, but as promising and as encouraging as that sounds, we don't want 20, 40 Ospos all off on their own doing things separately. We have to think about how they get organized as a network and that's something that's actually being thought through and talked about. So some things, this is my list, this is Hopkins' list, whatever lens you wanna put on it. Come up with your own list if you want. You don't have to think about this one, but these are some of the things I talk about. We are actively moving past into being an Eclipse product. So in addition to the functionality aspects of past that I think are important for open scholarship, this is a tremendous opportunity for you to get involved with an open source foundation that does this for Fortune 500 companies. They're going to take past and turn it into a community kind of product in a way that they do with their largest private members. Consider joining Ospos, there's no fee, there's nothing. It's easy to go to the website, sign up. We have a series of regular events and special events. Some of the luminaries from open source have committed an enormous amount of time and effort to this and I'll just be honest, one of the ways they're looking at it is could the Hopkins Hospital be a model for other Ospos, right? Because they too want to engage with universities in a more systematic way. Right now when they come to a place like Hopkins and say can you tell me what open source software you have, what licenses do you have? Who do I talk to about getting to the students? It's too opaque, it's too difficult, right? So the Ospo can become sort of that organizational API if you will. The National Academies Roundtable that I just described is launching a communities of practice and I believe there are about 57 institutions I've already agreed to do this work but I would encourage yours if it hasn't already to think about that. Start an Ospo. It can be done. And there are now resources through Ospo Plus Plus and other entities that can help you do that and if you'd like to talk to me about that, certainly please do so but Ospo Plus Plus is a great way to do that more systematically. And I'll just reinforce this point that I've had the chance to visit the St. Francis Neighborhood Center now a few times and I've had the pleasure of getting to know their staff. It's humbling in a very profound way to see people in these communities basically taking ownership of the problems they're facing, right? And I'd like to think universities can be amplifiers. Not that we're the experts and not that we tell them how things should be done but take that tremendous talent, creativity, energy, spirit and amplify it because I don't think billionaires flying into space is going to get us anywhere. We are also the ones we're waiting for, right? So just some resources. I know the slides will be put up so I don't expect you to commit this to memory and I can leave it up but I do think we have a few minutes for questions or comments or any clarification might. So thank you for your attention. And if you could come up to the mic, that'd be great of course. Or if you want coffee, that's fine too. Okay, so I just have to ask. So in your scenario about what all AI is going to be doing in 10 years, what, because humans are always innovating, always trying something new, trying something different. So AI will be taken care of some kinds of things but then what might humans be focusing on? Yeah, that's a really good and important question. And I tried to pick examples that, well, I'll speak for myself, somewhat selfishly, maybe somewhat callously. They affect me the least, do they affect me the most is the way I sort of went through that list, right? Of course I depend on people driving trucks and getting me things but truth be told, am I doing anything to lobby for truckers and giving them better working conditions and so on? I'm not. And if that were to happen, it probably wouldn't impact my professional life in any way. But experiments and code being written by AI, that will affect my life professionally. So I think one thing we'll see with AI is it will sort of rise up the labor market to use a term that Bloomberg News talks about a lot. The different layers of skills, right? And the different levels of professional, quote unquote professional education that are required to do those things. So there's a tremendous amount of things that could go terribly wrong with AI, right? So clearly one of the things we have to do as people is assess them, vet them, so on. The biases that exist in the data, the biases that may exist in the algorithms, right? Those are two different things but they may both exist. I had a very interesting exchange with a humanities faculty member, Hopkins, where she actually said, I don't mind the bias, I'm a humanist, I wanna explore it, I wanna study it. So understanding those biases is the first step toward actually doing something about it, right? But the other thing that's a bigger kind of question, right, is if you feel threatened, and I use that word deliberately, by AI sort of taking on more and more of these functions, we're going to have to have multiple constituent communities working together, right? And I think open source platforms are a way to do that, to have the transparency, sort of the trust, the many eyes on the code, on the data, that will be necessary to make sure these things don't just go off and one fine day we feel like they're out of our hands. I'm sorry if that didn't answer your question directly but some thoughts. Thank you for the question. Thank you. I'm Shimo from the University of Cincinnati. I wonder if you can reflect for this office creation, there are many of the other university unit can be also the home, like the VPR office, CIO's office. How this attached to the library's mission and how do you articulate to the university administration this part of the library's mission? Yeah, really good question Shimo, thanks. So I'll begin by saying I don't assume that an OSPO needs to be within the library in every context, right? So that's not a given, put it that way. What I will say is in our case, there's a couple of things that I think have been really important in this conversation. So I mentioned Stephen Wally, he has over 40 years of experience in software engineering, open source development and I met him at a dinner again when we used to do things like that where he kind of blew my mind when he basically said, you know, it's not that we don't have enough software, we have too much, right? Go look at the GitHub repositories and the exponential rise, right? And you know, he said I don't want to sound like that old cranky man, get off my lawn, whatever but everybody can become a developer now, right? But not everyone can write good code. And he said what we have is a curation problem. And he literally said, I mean, aren't libraries good at curation? So, you know, that's a fair question, maybe it's a challenge, I don't know. So one of the questions is if you think this is a curation problem, what's the right entity in your organization to approach that? The other is that diagram I shared of Articles Data Software, to me, those are the core components of Open Scholarship, right? And Open Scholarship is not something the library owns per se, but I would like to believe and I do think it's important, the library at every institution is involved in it. So if you want to take that idea of an open scholarship kind of lens, then it's another way the library can at least be at the table. But it is really important and I think this is embedded in your question to be aware that other entities in your campus need to be involved, right? So the first conversations I had with Tech Transfer, they weren't sure, what is this? What's happening here, right? But then when you talk about it's about different kinds of translation or it's about different kinds of licenses and the implications you have, it's very different. When research administration hears, this is why we should sign this open source agreement because this is what we're gaining. I know that you're worried about what we're losing but here's what we're gaining. And when you talk to the researchers and say, we can help you manage your software. So one thing that's important, a lot of the things that we've been trying to get our faculty to do around data, right? Document it, deposit it somewhere where it can be cited and identified, share it, so on. All of that with software is very different. The social norms are quite different around software even with our researchers. So there's different ways that messages can be sort of conveyed to different parts of the university but obviously I'm biased but I think a library is a great place for much to do that. Thank you very much for a very exciting and thoughtful presentation. Thank you Brian. And as a futurist, I appreciate your tenure's lookout. But I had one specific question. Sure. How can academia at a global level use open source platforms in collaboration to tackle the climate crisis? Let's see, I have five minutes. That's assuming no more questions. So there's a couple of things I would think about, right? Sadly, one of our really talented software developers is leaving my group for a small startup that is creating a platform to help companies, organizations of any size, track their requirements for climate change. Reporting requirements on carbon emissions and so on, things like that. After getting over the begging him to stay conversation, I basically said, are they using open source? He said, that's something they're starting to think about. Yes. Stephen Wally has been involved in an effort in I believe Denmark around building open source tools for how energy gets tracked and distributed in the country. One aspect of climate change and maybe this is kind of the, I don't want to be involved in the climate change because it's expensive or difficult, whatever is security, right? So we had all those hacks with the oil refinery system in the US, not that long ago. 85% of oil refineries in the US write their own software, it's bespoke. So do a security audit of that versus an open source platform that's transparent, that's open, right? So we had the vulnerability yesterday. Matt Green, who's a researcher at Hopkins cybersecurity expert, tweeted a bunch of stuff where at the end of it he basically said, why isn't there a database of the vulnerabilities that we know about or potential vulnerabilities we know about and how it gets addressed and then open source way of doing that, right? So that model, I think is how we're going to do it. There are plenty of specific examples where you can look at how climate change might be addressed, but if it's done in this open source way, like open source as a verb rather than a noun, I think that'll be key. I don't know the specifics of the communities that were complaining about the COP26 climate change and not having their voices heard, but if there are ways to use open source to get their voices heard, to get them to coordinate and work with each other, then I think that would be wonderful as well. Any other questions? Well, as we like to say on Zoom, I can give three minutes of your life back. So thank you again.