 OK, great. Can everybody hear me OK? OK, it's great to be here. Great to be back at Berkman. And I am going to speak just a little bit about OER for scale through the lens of some of the trials and tribulations we've been going through with connections over the last 13 years or so. So we're going to talk about scale. I think everybody in this room is united by a vision, by a drive to move away from the yesterday and today's factory-based education world and move to a new world and a world much more organized like an ecosystem where we have tremendously rich, diverse resources that are available for all to use and to improve. And this is really, I think, the ethos that the OER movement is trying to pursue. I would argue, though, that in spite of our best efforts over the last, say, decade, there is still a lot of room for improvement. There's a lot of room for us to grow and to improve this movement and to scale it up so that we can eventually make this factory model a thing of the past. And the critical thing that I'm going to talk about today are some of the main issues that we've been thinking about for the last two or three years with regards to connections and to think about how to really take it from today's scale to a truly world-beating scale. And so hopefully these are some lessons that could be of use for the broader community. So the question is how to scale. So I'm going to argue that in order to really have OER make the impact on the education world that it really truly deserves to make, we need to move from the current model that you could call OER 1.0, that's really fundamentally, I would argue, based on informal sharing of resources that is really focused on the resources themselves are in OER is for resources. A world that, as Kathy talked about, fragmented by both intellectual property, different licenses, and also fragmented by all kinds of different technologies. And also, frankly, all too often having the lack of a sustainable business model behind. So what I'm going to talk a little bit about are some of the things we've been thinking about moving into a world we'll call OER 2.0, where we're thinking almost a step back in the past, thinking about turnkey solutions, thinking about learning outcomes rather than resources, thinking about new kinds of scalable platforms and also new kinds of business models. So let's talk just briefly about turnkey solutions. How many people have read this wonderful book, Crossing the Chasm? Show of hands. Everybody here needs to buy this book. Basically, it's a beautiful book on marketing that talks about the penetration of new ideas and new technologies. People have the concept in their mind of a bell curve of the population, and they think that any new idea or technology is going to just smoothly flow in from the visionaries and the initial enthusiasts just smoothly through the marketplace. The problem is that this is just not the case and that there's a yawning chasm that lies between the techno visionaries and early adopters and all the rest of the population. And I would argue that connections in particular is really something that is caught on with this side of the curve and yet has had a tremendous difficulty over the last 12 or 13 years trying to cross this chasm into the mainstream. And in order to do this, what we need to do is make the road to adoption much, much easier for materials like we have in connections. And so as a result, with some gracious support of the Hewlett Foundation and several other foundations we launched, OpenStax College earlier this year, to develop a system where educators can adopt books first and then adapt them later rather than the other way around. So for those of you who don't know about OpenStax College, it's a library of free community college textbooks that we're building for the highest impact courses. The critical elements are, first of all, that the material is absolutely intended to be a turnkey solution for these courses. So we're viewing the textbook now as a suite of services that comprises not just the book, not just a collection of resources, but everything that is required for an instructor to be able to adopt the book, mobile apps, ancillaries, homework systems, analytics in the back end, et cetera. And the idea is to make it extremely easy for people on this end of the curve to be able to adopt and use OER. The other key difference with a lot of the material in connections, as you know, connections is primarily about user-generated content. But with OpenStax College, we've actually moved towards a professional development model where the materials are actually written by professional authors, peer reviewed, and edited by a very high level set of editorial boards. So that was part one, which is turnkey solutions. The second is the idea of thinking about new kinds of sustainability models. And I would argue that today, the world of OER, and in particular the world of OER producers, really lives off to the side of the broader ecosystem of including colleges, student groups, legislative organizations. And what we really need to do is move towards a world where we can put OER into the center of this much more productive ecosystem. And so one of the things that we have really tried to do over the last number of years is to exploit the CC by license that all of our materials are developed under and actually build a rather large group of sustaining partners who are actually going to be building services, homework services, tutoring services, printing services, around all of these books, and actually be able to send a sustaining revenue back to the project, both to develop new books and also to be able to sustain the project into the long term. So that's just a little bit about different kinds of business models. The other thing that we've really been thinking very hard about for the last three or four years is moving away from simply developing materials and actually move towards systems based on analytics where we can think of building a customized textbook so that every kid gets their own personalized learning experience. So we've been doing a lot of work on building analytic systems in the back end of connections and OpenStacks College in order to tune the presentation of problems, material feedback, so that students can have the perfect book for their individual learning interests. The challenge to building any kind of personalized learning environment is, again, a scale problem. How to scale to very large communities of learners and many, many different kinds of learners. And also how to scale to many, many different kinds of subjects. And so in contrast to the kind of approaches that have been pursued in a lot of the intelligent tutoring world over the last couple of decades or so, we're taking a slightly different approach. So the first is we're going to be developing the large number amount of content, the large amount of feedback information, the large number of assessments using the community development enterprise that we've been building with connections. But the second really critical thing is we're replacing the ontology or top rules based top down systems that are used for mining analytics information to tune the presentation to students and replace these with the machine learning algorithms, the kind of algorithms that you use every day when Netflix or Amazon suggest a movie or suggest a book or when you do a Google search. So just to give you a sense of how this is going to be working, if you think of this as the connections repository of materials, we have built around it a number of supporting infrastructure pieces, including quad base, which you can think of like a CC by connections like repository just for assessments, just for assessments and questions and answers. We've built a similar one for interactive simulations, a similar one for video tutorials. And we have a peer review system that allows professional societies and other organizations to get involved quality controlling all of these. And then what we have around linking all of these various sets of resources is a machine learning layer that basically provides the interface for instructors to develop courses and then students to interact with those courses, tracks the learning progress and direct students to just the right material at just the right time. Actually, do we have the timekeeper? Good. OK. So the last thing I'll mention about this personalized learning system is we're also working with some really, really fabulous cognitive scientists from Duke and WashU to bake in the cognitive science principles into the system. And we are beta testing the system right now as we speak at Rice moving into Georgia Tech and UTL Paso and a few other higher ed schools in the fall and also into the STEM Scopes project. And we hope to have about 500,000 kids using this system this fall, this personalized learning system. The key thing for this audience is that the architecture, all the software, and all the content is open licensed under open software and Creative Commons licenses. And you can check out PLS at RiceDSP.org if you'd like to learn a little bit more or get involved. The other key element here is that we actually did a very, very fine experiment and actually extracted several neurons out of Vick Vucek's brain. And we're actually growing these on a silicon substrate. And this is actually the heart of the compute power of the personalized learning system. OK, so the last thing I'll just mention is talk about platforms. Because we're very, very interested in moving into a world where we're not just serving millions of users via our platform, but tens or hundreds of millions of users. And that's going to require a rethinking of all of our tool set. And the key thing is that there's an engineering tension between ease of use of a platform and the quality of the output. And one of the things that has kept connections away from moving to simpler editing interfaces along the lines of more like a Wiki-like interface is our desire to be able to produce really, really high quality, not just web output, but PDF print books and now EPUBs. So the response to this engineering tension has been we're just embarked on a complete redesign of our entire platform that we expect to have ready in early 2013. This is a project that we're collaborating on with engineers at Google. And this is also graciously funded by both Hewlett Foundation and the Open Society Foundations. The key things that we're looking at with the new platform is moving from our KonexML-based XML markup to an HTML5 markup so that we have better support for mobile devices. That is also going to give us improved both editing and internationalization capabilities. And finally, in order to truly handle hundreds of millions of users, we're gonna be migrating the entire tool set up into Google App Engine for scalability. Okay, so I think I probably should end there. Hopefully this made some sense to you. Hopefully some of the trials that we've been going through for the last number of years and the solutions that we're developing can be useful, not just for our users, but also the community at large. And feel free to check out these URLs if you'd like to learn more. Thanks very much. Super, thank you so much. Thank you.