Beyond Institutions: Personal Learning in a Networked World

by Stephen Downes

Presented July 9, 2014, Network EDFE Seminar Series, London School of Economics. For audio, video and slides, click here. 

Introduction: What learners Need

The talk is called "Beyond Institutions Personal Learning in a Networked World" and I want to begin with a story that came across the wires recently and I thought was very appropriate for this venue. This was a manifesto that was authored by economic students demanding that the way their profession be taught be changed.

They made observations about things like the global economic collapse and global climate change and other things not really being addressed by current economic theory. They suggested, not so much that current theory is wrong, although current theory is wrong, but that they should be given alternatives or different ways of being able to look at the world. They wanted, in other words, from my perspective, more control over their education.

Professors, meanwhile, far from embracing this Renaissance of student-led learning, are sticking to the tried-and-true traditional way of lecturing in the classroom to the point where they want laptops banned from the classroom. Dan Rockmore, in an article in the Atlantic as I recall, New Yorker, said, "These digital assistance are more suitable for play and socializing."

Not getting the point that learning today is about play and socializing. It's interesting to the study he cites, there's a little study that says, "Taking notes by hand creates better memory recall than taking notes by typing.'' Again, completely missing the whole point of what learning is about. Learning is not remembering.

Pretty much anything works better than the lecture method that traditional institutions defend. It's funny we're doing the lecture method here. Actually, what we're really doing is we're in the process of creating a learning resource that we hope will be used, shared, cut, clipped, and otherwise abused by people around the world through the years that follow.

This isn't so much about the content of this talk and you remembering what I say as it is about creating the possibility, the potential for dialogue and interaction. Iyadunni Olubode: “Everyone knows that learning is growing at an increasing depth and an increasing breadth, so you need people who can constantly learn and bridge that gap, even when they're in their current jobs.”

This is the shape of learning in the future. Not learning where you go to a lecture, you remember the gospel wisdom that your professor has told you, and you go out forth in the world and propagate it. It's a world where a person is constantly learning before they get to university, while they're in university, after they're at university.

It's a world where the content, the nature, and even the means of learning is changing almost on a daily basis. Look at the setup that we have here which really would have been practically, not technically, but practically impossible five years ago. Five years from now, it will be different again.

People are looking for learning that isn't so much the repetition of their professors' ideas, but learning that they can apply, that is a part of their life, whether it's part of their life in work, part of their life in their hobbies or their avocations, or part of their life just in what interests them. They expect universities to be flexible.

We were having a conversation this morning in the Director's Dining Room, which, if you haven't been in there, isn't as posh as it sounds, although the walls are very nice. I love the wood panel walls, the little sandwiches, anyhow, eggrolls. We were talking about how the university is structured Bachelor's, Master's, PhD, maybe a couple of other names for these degrees. That's pretty much it. That's not what people are looking for. In survey after survey, they want learning that is directly and immediately applicable to what they're doing.

One of the discussions that came up yesterday... there's all kinds of surveys about what students want. You can go out and site surveys of students saying, "We want the lecture method." The discussion centered around the idea, at least my discussion did, that you can survey students, but students are the people who are already being successful using the current method. If you're going to survey what people need from education, it's important to survey both those people inside the institutions and those people outside the institutions, particularly those who the institutions did not serve well, those who failed, who dropped out, whatever.

Economists, on the other hand, have their own view of what academia needs. We've been hearing a lot about it, both here in Europe as well as in North America. The economists, those paragons of virtue (that's cynicism) when they say something is good, I really begin to worry. What they are talking about these days is the destruction of the university at the hands of the massive open online course. As one of the people who invented the massive open online course, I feel a little personally involved here.

It wasn't our intent, I just want to be clear about that now, it was not our intent to destroy universities. That's not why we did it. We want to change universities, and we want them to work for the better.

Thinking in Models: for Design, for Learning…

A large part of this talk is about that change. It's interesting. We go from the first slide about people wanting to be relevant, wanting universities to be relevant, all the way to the last slide about what's going to replace universities, without doing all the thinking that we need to do in between. We need to do this thinking in between.


Let's begin our thinking with where the current trends, we're told, are going. We're told there will be tiered service models at universities. We're told there will be analytics and data-driven management. We're told there will be alternative credentials. To a certain degree, all of these three things are true.

To a certain degree, none of these three things are going to work themselves out in the way that the economist or economists or education reformers predict. When you look at that, basically it's like they have this model or design in their head of how we could rebuild the university system, wipe it all out, start over, and we'll have a new model.




Figure 1 - workflow process employed to assist LMS selection

This model of accountability and cost frameworks and all of that will solve all the problems that the current system has. Models are popular in education too. Here's a model (Figure 1) of a workflow-processed employee to assist LMS selection. You can't really read the small writing there. It goes from enrollment to program administration to learner interactions to content creation to assessment.

It's a fishbone diagram. If you're in economics or business, you're probably familiar with it. Models of how to select educational technology including customized lists of LMS features, a way of picking among those 305 features of a learning management system that you might want to solve the educational problems at your institution.

Models of how to do learning, learning design patterns: Grainne Conole has done a lot on this. Diana Laurillard, who I really wish had been there yesterday because I really wanted to have a chance to discuss some of this she's been working on. If you get the pedagogy right, that will solve all the learning problems.

Best practices for typical learning tasks: this is a reference to a paper that talks about the conditional release of materials, what we used to call back in the day programmed learning. You do some learning, you do a test. If you pass the test, you get to see the next learning. You still see the old professors with their overhead projectors and their slides and their little piece of paper that they slowly work down the slide. Goodness, you can't have people reading the bottom of the slide first. It would be just wrong.

Even models of how to offer courses, there's all kinds of discussion like this in the literature. The model of online, hybrid or traditional models, of course, is broken down into things like the types of tools, whether you're using discussion boards or white boards or websites or videos.

There's the selection matrix. Out the other end comes, again, it's hard to read on this, level two decisions. You have all of these inputs and outputs. It's very much a system's theory kind of approach. You get the system, the process right and everything flows out the bottom the way it should.

In these models, these designs are being implemented as educational technology. This is what education reform is about. Wow, as I said yesterday. It's also about making a lot of money for some people. It's about standardizing and rationalizing the educational system to fit into a certain set of models or designs. We have Google coming out with Classroom just as part of Google apps.

It's Google apps for teachers. It does all the really useful stuff that teachers need to do like marking and scheduling and assigning learning tasks and all of that sort of stuff. It's education done by software application, basically. It's being commoditized and being standardized and being packaged and delivered. This is education of the future.

It's interesting. The MOOCs that came out, not the ones that we did, but the ones that came out after us, are, again, very much in that same model. Get some videos, get some exercises, get some tests, step them through it week by week by week. You don't even need a professor. It's nice to have them to do the videos, but otherwise you just deliver this as a content package.

Everybody gets the same thing. That's what works. It's Dan Willingham and Paul Kirschner and they say, "There are no individual differences in how we learn. The way we learn depends on the content, not the learner." That's a pedagogical approach that I feel is incorrect. I think it is obvious that people learn differently. Learning styles, as a theory, and especially as a design theory, may be wrong.

Models are not Reality

The fact is that people learn differently, that they have different objectives, different priorities, different goals, different times that they want to learn, different pets sleeping on their keyboard, all of these impact how people want to learn. That's immediately obvious to anyone who actually looks at people learning. Even as I look around this room, he's on an iPad, she's typing, she's writing on a notepad, he's asleep. Everyone learns differently.

This is more from Google. This is the work of education, creating and collecting assignments, making announcements, asking questions, and, of course, a folder, school folder for each assignment and for each student, which aside Google will mine. They've said they will no longer mine student data for commercial purposes. They recently came out with that pledge. They did not say they will no longer mine student data. They just said they won't do it for commercial purposes anymore. Interesting.

With models, again, maybe I'm talking to economists here, maybe I'm not. I'm not sure, but this was actually the subject of my Master's thesis, which maybe three or four people have read. The model is not the reality. That's my 235-page thesis in one sentence.

The model is not the reality. The model has never been the reality, and worse, when you're doing any kind of research, if you use a model, typically the answer to the questions you're researching have been defined by the employment of the model in the first place.

That's what happens here. If we use these models, or other models, any kind of model design predetermines structure to define how we're going to understand what learning is, we've predefined what the outcome will be, but learning needs to be open-ended.

Learning needs to be an exploration and a discovery, not the output of predefined, standardized products. The adaptation of these models to computerized learning is no more effective than the use of these models in the classroom.

New Versions of Old Models

If your teacher walked in and spoke from a script and answered every question in the same standardized way, we would not consider that effective education. The same is true if it's done on a computer.

Again, we're told that these MOOCs are a new pedagogy. We're told that the stuff that's being done at Stanford, MIT, EdX, the rest of it, is going to change education, but it's a continuation of the same models and the same strategies that have defined education for decades, despite the fact that people are asking for something different.

It's not even that the new models are the old models with new names. The new models that we're seeing today being done on a computer are the same models we saw being done on a computer a decade ago and two decades ago.

Audrey Waters talks about Fathom. Fathom had a plan. What their plan was, to take learning materials, put them on a computer, and make them available, even openly, to people who wanted to learn.

That was 20 years ago, well, not quite 20 years, but almost 20 years ago. People talk about EdX and Coursera and the rest of them as being new. It's like they've written off the previous experiences. Interestingly, the president of Coursera is the president of Yale who had the wonderful idea of putting courses online and charging money for them.

Guess what he's going to do at Coursera. It's the same model being repeated over and over and over again. Universitas 21 was invented something like 15 years ago to monetize online learning. It's one of many initiatives to take a course, charge university-level tuition for it, and sell it online.

It's not what people want, and these initiatives continuously fail. Even LRMI (Learning Resource Metadata Initiative) is a system of standardizing the descriptions of learning resources, but it's a clone, in many respects (I'll talk about a way in which it's not a clone in a number of slides) but it's a clone, in many respects, of the standards-driven efforts that have come before. AICC, (Aviation Industry Computer-Based Training Committee): they had a set of learning resource metadata standards in the 1990s. IMS, Instructional Management Systems. IEEE, which is the IEEE, learning object metadata. Shareable coursework, object recognition model.

Again, over and over and over again we see, you take standardized resources. You create standardized descriptions, standardized search mechanisms. The standard is the golden standard of learning, it seems, and it's always though if we could just get this precise standard right, it'll all work.

The results have been, over the years, pretty much what you expect. Here's LRMI again, a Phil Barker presentation on LRMI. He listed the institutions - there's something like six or seven institutions using LRMI. As with all of these standards initiatives, the easy part, although you'd never know by listening, is to define the standard. I've designed dozens of standards. The hard part is getting people to use the standard, because everybody does these things differently.

Even the terms within the standards, you have a term, title, or no, even better, author. You'd think, author, how could you go wrong with author? But you get everything from people, lists of people, organizations, associations, sometimes pets, sometimes nothing at all. Right now, I currently have an issue with my own system because people are putting hypertext markup language inside the author field in their RSS.

Why they feel the need to put markup in what should be a simple string of characters, I don't know, but they put anything and everything inside the author tag. New versions of old models don't produce new results. I'd like to go down to Canary Wharf and tack that onto one of the buildings. If you do the same thing, even if you do it on a computer, you're going to get the same result. The same result, even the people who are producing the result say, isn't sufficient.

The Right Model is No Model

I criticize Coursera. I criticize the Stanford MOOCs and all of that, but when Norvig and Thrun launched their artificial intelligence MOOC, in the first week, 150,000 people signed up. Overall, I think it was something like 250,000 people signed up for one course, a really hard course that's really difficult to understand, in artificial intelligence.

Forget the fact that a lot of them dropped out. A lot of them didn't. Tens of thousands finished. This, by itself, indicates that the old model wasn't working. There was such a pent-up demand for upper-level university courses in artificial intelligence that, when one was finally made available, people knocked down the doors trying to get to it.

George and I launched our MOOC on connectivism, which some of you may have heard of. Most of you may not have heard of it. If you talk about a niche subject, this is as niche as it can get. It's an unknown theory in the field of educational technology.

Try going out onto Fleet Street and advertise that. No, nobody's interested. We got 2,200 people without advertising. That was our first MOOC, and that was when we realized we were onto something, because again, people were beating down our doors.

Not as many people, but they weren't very big doors. You see that offering these courses to 10 or 12 people at a time in a seminar, whether it's online or offline, isn't going to work. Here's where I go meta. The right model is no model.

The right model is to do away with the models. Think of non-standard-based systems. Think of non-standard designs. Think of courses where there are no defined learning objectives. Think of a learning environment where there is no common core of content. Think of a conversation where you and I have not first established a shared understanding of the meaning of all of the terms.

That's reality. That's this room. There is no model for what's happening here. If there is, I'm probably breaking it, although I'm probably following...you can come back after the fact, look at what I did, and say, "Oh, yeah, that's part of the model."

The slide, because the article that I'm referring to referred to Sugata Mitra ( Sugata Mitra has almost been commoditized these days) – but the concept, the idea behind what he did - and also, too, there's been a lot of criticisms because people went back years later to see these computers, and what they found were nothing but holes in the wall. Computers had been vandalized, the Plexiglas stolen.

That happens. It doesn't mean he was wrong. It just means that that experiment for that time was finished. People were looking for a model that would always work, but things don't always work. They work for a time, then move on to something else.

That's what worked with what Mitra was doing. There were no roles. This is David T. Jones, with a cameo by Michael Jackson. "What's missing," he says, "in the standard-based models is what we used to think of as BAD." BAD. (Bricolage, Affordances, Distribution).

Bricolage: I remember when IMS Learning Design first came out and they did a tour to promote it. They used, in their documentation and in their presentation, the metaphor of actors on a stage, and the teacher, of course, would be the director, and then everybody would play their roles.

My question was, because I'm from Canada and we have these, you probably have these here, too, improv, I said, "What about improv?" You can't do improv. That's what's missing with these models.

Affordances: When people built the Internet, they did not intend to design a system that would store, I forget the exact number. I referred to it yesterday, 680 million cat photos. I think that's the number. That was not their purpose, and had anyone anticipated in the 1950s and '60s that they were building a system for storing 680 million cat photos, first of all, they would have thought it's ridiculous.

Then they would have thought, "Who would want such a system?" That's the beauty of the Internet, is that, although it was designed for academic research and to survive nuclear wars and things like that, it turns out to be the perfect place to share your cat photos.

That's what makes it beautiful, the affordances, the possibilities of technology that come up that you didn't plan on ahead of time that you can use for other things. History is full of these things, from the first person to use duct tape for something other than to repair ducts. The misuse of technology is what makes technology great. I wish I could think of more examples, off the top of my head, because there are dozens and dozens of them. I can't, because I'm in the middle of a talk, but you can, so you should.

Distribution: The idea that you have to be here to get my wisdom is ridiculous. What new technology, new learning allows is for this to be not only here, but available online to web video-streaming people (Hi, web video-streaming people. I always like to make sure the video people know they're important) (Actually, that's not true, that’s something I started yesterday). Also, the audio-recording people. Hi, audio-recording people, who are enjoying a less than perfect experience, but still, you're there. The idea that things don't have to be in one place anymore.

Content

We need to question the presumption -- and it is a presumption -- that there's too much content, too much data. We have information overload. It has to be organized. It has to be standardized. It has to be categorized. Must be delivered in packages.


Figure 2 - Apple pie recipes

David Weinberger says, "We do not, interestingly, feel overloaded by the effects of 1.3 million apple pie recipes or 7.6 million..." (He is dramatically underestimating. I searched yesterday. I actually checked). "7.6 million cute cat photos."  (Or maybe he was just referring to the subset of cat photos that is cute, but clearly, it's much larger than that). It's interesting. We're not overwhelmed by it at all. I don't walk out in the world, wondering, "What am I going to do? I got so many cat photos."

When we talk about learning, it's almost the first thing people say. We started our MOOC. What we did is we got people to bring in content, suggest content to us. We brought in content. First thing people said was, "There's too much content. How am I supposed to remember all of this?"

I wish I'd had the cat photos line back in 2008, but he didn't say it until 2014, so I couldn't. Not without time travel. You're not supposed to remember all of it. That's the old way of thinking that you're supposed to remember the content that the instructor is delivering. For all kinds of reasons, not the least of which, you won't remember it, anyways, you're not supposed to remember it.

The whole idea of our course, any course, this course, this talk isn't to get you to remember what I said. Just so you know it's OK if you forget every word I've said. It's to stimulate in you the sort of mental experience that will create in you the sort of mental structures that will, at some point in the future, be useful to you.

It's to stimulate this environment and this sort of interaction. That's why we want questions and discussions after, and that's why we want the real reason people are here, the wine and cheese (I'm assuming cheese) at 4:30 (I was told there would be wine). We're not expected to master every one of the 1.3 million apple pie recipes. Mastering one will actually be enough. Mastering three to five will be more than most of us ever do.

Actually, mastering one will be more than any of us ever do. The main point is we pick and choose them as we need. The interesting thing is take a room full of 50 people, and 1.6 million apple pie recipes, people will choose, not all the same recipe, but different recipes, because they look different. They seem different.

They appeal to us in different ways. I have a whole talk on this. How you will read "Perfect Apple Pie Recipe" from Pilsbury.com, and that will set off one set of mental associations for you. Someone else will read it very differently. Maybe you have a favorable impression of Pilsbury, or maybe you think of the dough-boy and say, "No, that's not for me."

Maybe taste.com.au appeals to the Australian nationalists in the room. No American would use that one. Australian apple pie isn't as American as apple pie (at a certain point, my talks devolve into nonsense; I'm sorry, it happens).

Personal and Personalized

I want to draw a key distinction here, and it's a distinction that the model builders don't get. It's the distinction between personal learning and personalized learning. It's the same difference as pretty much anything like this. Custom tire, customized tire. Chocolate, chocolatized. Maybe that metaphor doesn't quite work.

The idea here is that personalized learning, personalized anything, is you take something that's off-the-shelf, and you tweak some variables in it, and that has thereby made it personal. That's what's offered in programmed learning. That's what's offered in customized learning solutions, personalized learning, adaptive learning. Any of these design-based systems, they're personalized. Really, it's one package with a bunch of options in it.

Personal learning is made to order. Personal learning is even learning you make yourself. Personal learning is where you build your learning, not from a kit, but from scratch. See the difference? People don't want customized, necessarily. Sometimes, they do, but typically they don't. They want personal. They want custom.

That's the expensive part of learning. We were talking yesterday about the Oxbridge model. "The Oxbridge model is so much better," said the Deputy vice-chancellor at Greenwich. Why is it so much better? He has a point. It is better, in many ways, because it's personal.

The problem is it’s also really expensive. To provide that for everyone would cost more money than there is in the world, but boy, it sure is nice having learning that's tailored exactly to your needs. If you can build it yourself, that's better. If you can design your food or choose your food from an infinite array of choices, that's better than going to McDonald's. Even if they offer to take the pickle off for you.

Institutions, I would argue, understand personalized. They don't get personal. There's so many ways in which this is manifest. Even in some of the discussions over the last few days about personal websites by institutional staff. The first response that comes up is, "But will they follow institutional standards?" The answer, of course, is, "Well, no."

There's the concern that widespread adoption of social media brings shared interactional practices that do not match university arrangements for learning. They talk about classes to people who aren't in the class. That's just wrong. You can have a personalized class, but it's still this thing in a box. Something that's personal can go beyond the box.

Autonomy vs Control

Autonomy, for many reasons, rather than control, is essential in education. This is a bit of a digression, but I want to be really clear about what I mean here (by the way, I'm a huge Bill Waterson fan). Autonomy does not mean no structure. It means choice of structure.

Think of touring a city. The way the autonomy-versus-control thing is typically sketched is, if you're visiting a city for the first time, either you wander around with no idea of where anything is or where you are, or you're taken on a guided tour. If you actually want to get someplace in the city, you pretty much have to do the second. You can't do the first, because you'll just wander around aimlessly.

That's a dumb distinction. Those are not the choices that are given to people. If you're visiting a city for the first time, you have a number of choices. You could -- and I do this frequently -- wander around, aimlessly, not knowing where you are. Or you could wander aimlessly around with a map on your phone.

Or you could wander aimlessly around with a dead battery on your phone, but using maps that are put up in the city. Or you could ask someone for directions. Or you could, as I did this morning, take the train, which will drop you close to where you want to go, although you might have to ask for directions to do that, too. Or you could get on one of those hop-on, hop-off buses.

Or you could get a friend to drive you around the city and show you things. They offered to do that for me in Greenwich. Or you could join a guided tour. That's choice. That's autonomy. The other option is they kidnap you and take you around the city, no matter where you want to go. That's control. Those are the real choices.

Ironically, we do education the second way. Control, really, is an illusion. Really. When you manage and control a work for outcomes, they tend not to result. This is the interesting thing about these designs. They're really abstractions of the actual process. They're not useful, as prescriptions of what should be done. If they're useful at all, they're useful as descriptions of what was done, but only partial representations of what was done.

The personal, by contrast, is not designed. It's based on -- and the photo of the murmuration should a big clue here -- based on self-organization.


Figure 3 - Murmuration


It's based on the idea that people can manage themselves and manage their interaction with others, including learning, for themselves. The murmuration. They've done studies on the murmuration, and of course, there's no head starling. What's interesting is there's no mass communication, either.

It's rather more like a mesh network, in which each starling is reacting only to the seven starlings around it. Anytime a starling changes position, the seven starlings around it change position. That's what produces the cohesive movement of the whole.

It's interesting, because when you think of it -- and that's not even in this article -- when you think of it, a murmuration is a perceptual system for starlings. It's a way a whole flock of starlings can magnify the perceptions, say, of a hawk, by any individual starling. The great wildebeest migration, same sort of thing.

They're not actually going somewhere. We think they're going somewhere, but they're not. They're just searching for food and water, and their migration follows the natural patterns of the climate and the environment. It takes them around in a big circle, over time.

Complexity, Cause, and Murmurations

It's interesting. This whole concept of design, organization, planning, et cetera, suggests that we can cause these events to happen.

But for anything that's remotely complex, there is no cause, properly, so called, of the event. Landmark ideas are created not necessarily -- not at all, pretty much -- by individual people, but by societies, by this large murmuration of people interacting with their community.

The modern technological world is giving us new examples of that. The hashtag is a way of creating self-organizing networks. Imagine if we tried to plan the Internet so that we could account for and index and abstract all of the conferences that will happen from now on, before they happened. It'd be ridiculous. Couldn't happen. A new speaker series like this would be impossible to anticipate. Hashtag networks can be seen as self-organizing ideas. The hashtag is a murmuration of tweets (that's a neat line; I'll have to remember that).

Mary Meeker -- if you're not familiar with Mary Meeker, you should be, if you're at all interested in education technology and markets -- has observed the proliferation of apps, not just in education, in everywhere. What the app world does is facilitate this kind of network interaction. What she's noticing is that the edge, which is jargon-speak for the link between two nodes, is more important than the node, where the node is the person, the computer system, the learner, the starling, whatever. It's the connections that are interesting. That's what's interesting in education, as well.

The starlings, in education, are the students. The university, the learning institution, properly conceived, should be organized like a murmuration. Should be a self-organized assemblage of students. But then, of course, you don't need that institutional structure, at all, and it becomes really difficult to justify a $20,000 or 9,000 pound tuition rate.

A Reclamation Project

It's happening. There was a critic at yesterday's talks saying, "Do you know there are no students involved in these conferences?" I did a search. The search for "student panel" actually yielded 199,000 results on Google. The search for "ed tech student panel" in quotation marks, so it's the union set of those words in that exact order, yielded 2,000 results.

The students are doing this. They are organizing themselves. Audrey Waters says -- and I think she's quite right -- "The future of educational technology is a reclamation project." The idea here is that we, the learners, the people who need to learn need to reclaim the management and organization of learning for ourselves.

It's interesting. It's like the whole Facebook thing. Facebook has taken over conversations with our grandparents. They're using it to run tests on emotions and to sell stuff. We need to reclaim our conversations with our grandparents. We need to reclaim our interactions of Plato, Socrates, and the person next door. We have to reclaim control of the data, the content, and the knowledge we create.

The idea that this belongs to the university that it belongs to the institution is ridiculous. The idea that the university has any say over what students or even its professors would produce in this Internet is absurd. Lucy Gray does what I do. She puts her presentations up on Slideshare. One day, Slideshare deleted everything. No explanation. No recourse. She couldn't even contact them.

She had to tweet it to get any attention from them. No notice. They were just gone. That's why we have to be the owners of our own education. I read this morning -- I didn't have a chance to put it in the slides -- there's a guy who started a course, a Coursera course. Went two weeks into the course, and then he deleted the entire course. It was, he said, "An experiment in causing confusion." (It’s not the first time this has happened).

One of my colleagues, Ben Werdmuller, is creating something called Known. Known is an application where, as he says, you can still share selfies, make friends, listen to music, et cetera. Put up cat photos, very important. But in a space that's yours and that you get to have control over.

Ben Werdmuller, with Dave Tosh, built a thing called Elgg a while back, which is now widely respected as a social networking environment for learning. They also built something that really should have been much more successful than it was, called Explode, which was a similar sort of thing to Known, except I think Known will be better produced. David Wiley gushed when he heard of the "publish on your own site, syndicate elsewhere" anti-model, it's called model in the article but I'm calling it an anti-model because it's not really a model.

There have been (if you check out the #indieweb hashtag you'll see) indications of this, from Diaspora (disclosure: I invested a hundred dollars in it - that was my only investment; I'll never see a return on it) to App.net, which I actually pay money to as well.

Even to syndication itself, it's this idea that what’s today a silo (which is learning) is going to become the syndication end point. These applications, these services, these resources, are the things we reach out and touch but not where we invest our entire lives.

Reclaimed learning is network learning

Jim Groom has been running something called, "Reclaim your domain." There have been various other wordings, "Reclaim innovation, reclaim learning." Starting now, he writes, “a technology that allows for limitless reproduction of knowledge resources, instantaneous global sharing and cooperation. All the powerful benefits of digital manipulation, recombination, and computation.”

That was the potential of the Internet twenty years ago and it was basically stopped by the institutions that decided it should be organized a different way. The idea of "reclaim all of this stuff" is to bring back that idea of the Internet. That begins with personal control over your own resources and your own access to external services including leaning.

I've outlined the model in our discussion earlier today. It used to be the case that you would go to one institution, maybe two institutions and you did all your learning there. It's changing now so that you access learning from multiple institutions. Not just multiple universities, but multiple types of institutions, from colleges and universities, potential employers, current employers, past employers, to pet food stores, to fringe networks, to special interest groups, to hobby groups, to the government, to whatever.

They're all sources of learning. The idea is, you are at the center of this network of learning. Reclaimed learning is network learning. Reclaimed learning is having access to the tools and the mechanisms to freely author and create your own learning and share it with others and to access and use learning that was created by others and shared with you.

It's your mechanism for talking to the starlings that are nearby (I really love that murmuration example. I'll try not to beat it to death, but just happen to be beating a starling).

That's something like what we were building when we were building the first MOOCs. Our MOOCs are called "connectivist MOOCs" or "cMOOCs." What makes them distinct is that the people, the individuals are at the center and the learning resources are all distributed.

You might think, "Well, how do you build a course where the center isn't your course?" What we did is, we pointed students to mechanisms on the current internet where they could build one of each. We said, "Create a blog on Blogger. Create an account on Delicious and do that. Put photos on Flickr. Add videos to YouTube. Create a Google group." Do any of these things.

In the future we'd say, "Use your own personal web space to organize and coordinate your resources and then tell us what these resources are." You create your space, we'll create a space like this too, and then we'll join them together.

That's what we did. It wasn't a course where we had a pedagogical model in mind where we tried to step people through. It was this ridiculous no-rules mess that turned into a murmuration, that turned into a MOOC, that turned into something that can attract hundreds of thousands of people.

Technology Behind the Reclaimed Web

Some of the technology behind the reclaimed web, technology that allows us to have comments on our sites without having to author a comment management system...

There's a tool set these days is called the distributors' developers stack, where you can build your own website and access external services like storage. The old stack was called LAMP - Linux, Apache, MySQL, and Perl. Perl's programming language. That would be where you manage all your data. Today, the stack is your website, but then all the remote system that you can access with your website.

Making it easier for search engines to index (and this is the promised clarification of LRMI) is Schema dot org, where you manage and create your own metadata. You don't have to adopt and adapt IMS or IEEE or whatever. Schema is being set up by the search engines. The search engines are saying, "Here's what you can do. It's almost like tagging with tags for websites.''


Figure 4 - Bitnami Apps

Or this is one called Bitnami, it's an app store for server software. These are all different apps. Again, they're kind of hard to see. But Word Press, Joomla, Redmine (which I don't know anything about), WAMP stack (Windows something, something, something, probably Python) Moodle, Magneto, for e-commerce, just to name some of them. There are actually 50-60 different applications.

You want to run a survey. The old way to run a survey on your website is you download and install software like LimeWire, configure it, set it up -- and hope it works -- and launch your survey.

The modern way to do it is you get an account with Amazon web services or something that will give you some Cloud hosting. You use the app store you rent a LimeWire for $1.99. It installs in Amazon web services, you put a link to it, you have a survey ready. You're not even using any disk space and Amazon is taking the hit for all the traffic. It cost you a little money but it so much better than Facebook.

Take your data back from Google. This will be a thing, a personal web server, preloaded with open-source software that lets you run all of your web services from home, your home website.

If you don't think it's going to be a thing, think again. People use to go to the Western Union Wire Office to send messages to each other, then the fax machine was invented, and Western Union installed a fax machine on each one of their offices and figured, "This is great. Now, we can charge people for sending messages, and now they can send facsimile images too."

But what happened instead is, people bought fax machines and put them in their homes. You would think who would put a message sending device in their home? Now, we carry them in our pockets.

This is the future of this technology. The personal learning that I've been talking about isn't just personal learning in a conceptual sense, it is personal learning in a concrete hardware sense. Your university will be a box in your living room.

The modern web is distributed, interactive, murmuration of services and people -- 2RCode, OpenSearch, Windows Live Tiles, touch icons, RSS and even a thing called human dot text. All of these are little pieces. Your flavors will all be different as they should be. The idea that every website must be exactly the same is absurd. Only an economist would come up with that.

Social Networks and Neural Networks

This changes learning. This is what George and I are getting at is of the theory of Connectivism. “Connectivism repositions media as type of content” - but content is, remember, the McGuffin; it's the thing that gets us talking to each other.

We use media. We use our own services. We use our interaction with each other to create links with each other. These links with each other, these connections between people, between neurons, between concepts, between ideas. That's the actual learning. I could go on. I have two hour long talks about that, which I won't do.

One question that's always asked is what is the connection between social networks and neural networks? What is the connection between tweeting each other, or sending email, or skyping, and learning, where learning is the formation of connections between your neurons? Learning is, manifestly, the second. What is the link between this and the first?

There are two ways of looking at it. Connectivism embraces both ways. These are not alternatives, although they're alternatives, but they're not exclusive alternatives.

George's answer is that it's a multi-nodal extension. What that means is when you learn it's literally the formation of connections between your neurons. You have a network in your brain. This network extends out of the brain and into devices, into the Internet, and, eventually, through to other people. It's an adaptation of the old McLuhan idea that a communication system is like an extension of the body. An information system is an extension of the mind. Pretty smart.

My answer is just as smart. My answer is pattern recognition. My answer is that neural networks and social networks are, in fact, ontologically different, and one is not an extension of the other, but they're related.

They're related by, first of all, a common set of underlying principles described in the mathematics and the methodology of networks. I talked about underlying network principles like autonomy, diversity, and that sort of thing.

The other aspect of it is that networks learn by pattern recognition. The learning in a network is literally the formation of connections. A society learns by forming connections between its people. A human learns by forming connections between their neurons.

What these connections are actually doing is creating a capacity on the part of the network, as a whole, to recognize characteristic patterns. Just like a murmuration of starlings can recognize a falcon. Not because it has falcon-like content in its collective mind (when you put it that way, it's pretty absurd, right?) but, because it, as a whole, is a system that can react to the presence of a falcon.

(I'm assuming that falcons predate on starlings. I could be completely wrong, but I assume a murmuration would move away from a falcon. Anyhow, I don't know a lot about starlings, but you get the idea, right?)

The same principles underlie social networks and personal networks. A social network is a perception mechanism for a society. A neural network is a perception mechanism for a person. Persons can recognize patterns in society. Societies can recognize patterns in persons. The interaction begins to flow.

You can see that the Downes answer and the Siemens answer are really two sides of the same coin. Different ways of seeing the same topic. That's really common in network learning.

Even if we're examining the same thing, we're all looking at it from a different perspective. Our understanding of it is never going to be the content of any individual's mind. Again, that would be ridiculous. Rather, the combination, the pattern created by the multiple perspectives that come into play as we all look at this common object.

If I put a chair in the room, our understanding of the chair is the totality of our perceptions of the chair (which is why Wittgenstein was right and Moore was wrong - but that's an aside).

Network Learning

Connectivism can be thought of as a learning theory. Personally, I don't care whether you call it a theory or not. But, it accounts for existing theories, it explains where we are, and it we can make predictions.

One thing I do a lot of is make predictions. The predictions based on connectivism can be tested. I've got a history of making predictions, and I'll continue making predictions. One response of connectivism was the MOOC. We built a course in this network style. What we discovered (and frankly, we did discover this, we did not know this going in this) was that building a course as a network allows you to accommodate a lot of people with very few resources.

We had a budget for our first MOOC of nothing, yet we still managed 2500 people. I shouldn't say nothing - George wrangled the free Elluminate account. What we were doing is we were testing connectivism by using connectivist theory to create a course, and that course resulted in the MOOC. My verdict is the experiment was a success. Participants seemed to agree.

Creating networks, developing professional connections, studies of MOOCs - This one is a study done by a couple of my colleagues, Helene Fournier and Rita Kop. The people who actually took these MOOCs report that the really important part wasn't the content, because the content was just the stuff that George and I sent, but the creating of networks, the developing of connections, the networking, building on the affordances of this particular network.

Where are we going? Here comes the prediction part. Although, it's not just me anymore. I've been talking about personal learning and personal learning environments for a number of years ago.

The Aspen Institute - they're actually one of these right-wing think tanks, but we'll leave that aside - even they are saying learning has to be personal. Learners have to be empowered to learn any place any time. The idea is to use networks to support and guide learners and, most importantly, build operability across learning networks. Grain of salt: they're thinking of this management design perspective. You can't do exactly what they say, but they have the right in saying learning needs to be personal. Learning needs to be connected. Learning needs to be networked.

Learning control is moving beyond computer-assisted programs “towards authentic learning context, mediated by technology.” If you think about it, if learning is a network and not an on-site event-based kind of process, it can happen anywhere. It will happen anywhere. It will happen and be managed and controlled by people using their own devices, wherever they happen to be. The devices that are implicated in learning will multiply.

This is an interesting one. I love this one. Reading and networking will become one and the same thing. This is not exactly what Steve Pettifer is saying. Steve Pettifer developed a program called Utopia. It's an Adobe Reader, but when you read it, a sidebar opens up and gives you all kinds of resources from other services.

We built something similar to that, called Plearn. It was an in-house proof-of-concept project that we did between 2010 and 2012. I've seen similar sorts of things in Microsoft Word. The norm will be to have, if you're consuming (terrible word) consuming content, the norm will be that you have a sidebar experience. Even watching television. It used to be you just sit there (and watch). Remember that? But now, we have what they call the second screen experience.

Reading, watching television, all of these things will be, are becoming networked experiences. In the workplace, connective learning is already changing the workplace. It is going to really change the workplace when our learning becomes present in our devices. I used to talk about the fishing rod that teaches you how to fish. Now, recently, I saw an advertisement for a tennis racket that teaches you how to play tennis. Somebody actually built it. It exists. I wish I had the link for it. You have internal sensors. The internal sensors know what a good swing looks like. Feeds back to your device. Your device says, "You really ought to work on that backhand." Or whatever.

Cooperation

Teams and collaborations will be transformed. The old way, the design way, the management way, the control way is to form teams and collaborations, and put people in groups and get them all marching to the same tune, singing the same song, et cetera. The new way is to connect, to interact, but to work autonomously. In software development, they're calling the oscillation principle, where you get together and connect, and go away and do your thing. Get together and connect, go away and do your thing.

Cooperation is basically defined as a set of interactions in a problem space. The problem space might be anything. The idea here is that you can achieve results without actually having all the overhead of a collaboration. A murmuration is cooperation. Each starling is autonomous. Each starling decides for itself where it's going to go. There's no shared vision.

"Hey, let's have this really neat kind of amorphous mass." It's like one starling's saying, "That's a falcon," and making its own decision. In cooperation, we don't share models. We don't share designs. We don't share goals. We don't share objectives. Axelrod talks about cooperation. All cooperation requires is a durable relationship.

All the overhead that we typically associate with managing activities on the web, including learning -- things like centrality, commonality, learning objectives, learning management, controlled outcomes, even trust -- all of these are unnecessary for self-organizing systems. They're overhead. They make the institutions rich. They don't do the kind of job that the students need.

Cooperation means working with others. Working with them directly, without the overhead. Doing away with the negotiations, the discussions, the accommodations. All you need is to be able to interact and communicate with people. That doesn't mean you can't have negotiations, discussions, and accommodations. A lot of people like that stuff, and it's OK. But it's like the guided tour. A lot of people like the guided tour. If you want to get on the guided tour and share an experience with people, you can. The point of cooperation is we can run a society where you don't have to.

The new skills, therefore, both in teaching and learning are network skills. The new skills, pretty much in any discipline now, are network skills. This is a reference to Coding For Journalists so that journalists will understand the real meaning of things like lists, loops, and application programming interfaces. The whole idea here is to understand the concept of how individual entities are related to form patterns, data structures, and entities.

People forget about things like Codeacademy, which have proven, very successfully, through millions of users, that people can do things like learn to program on their own, without being told how to do it. It's like I mentioned at the beginning of the talk. The model for learning is like socializing and playing games.

A New System of Learning

If you've ever been to the media center, I visited the media center once at MIT, the Media Lab - it was a really interesting experience, because the place is a mess. It's utter shambles. It's probably a fire trap. But it's brilliant because people can interact any way they want, using whatever kind of device they want. If you want to build a robot, that's cool. It's all play, but I'm sure it's not all play.

We have these models. One model is called the super-university. It's going to respond to government directives or commercial imperatives. It will be designed. It'll produce outcomes. It'll create jobs. Economic development. Employment for graduates. Even manage immigration. Commercialize research. This is a line; people have told me this: "It isn't innovation unless it's been commercialized."

Again, if you're in economics or business, you've probably heard this. I really don't think that's true. You might say it's not innovation unless it's used, but that's something distinct. That's the one model. That's the kind of thinking that (is typical of) the people's that are saying, "There will be 10 universities left in the world." That's the kind of thinking that goes into that sort of model, that sort of design.

They talk about the importance of universities because we need them. They don't talk about what it is, in fact, we need. Think about the topics I've talked about in this presentation. Do we need more models and more designs? Does the world really need another theory of learning, honestly?

Do we need more standards and more measurements? I showed you half a dozen ways of standardizing learning resources. I could go on about standards and measurements. Do we need more centralization and control? Are the people out there yearning, "Control me! I don't know where to go." Do we need, quite frankly, the same mistake repeated again?

We're able now to rebuild our system of learning. Why on earth would we do it the old way? What is it that we need? What we need is the mechanism to support learning itself. When you ask the people what they want, they don't want immediate economic development. They want better lives.

They see things like learning - All of those people who went to the artificial intelligence course, all of those people who flooded into our MOOC, they're doing the same thing that the people of Leiden did when they opted for a university instead of lower taxes. They're doing what the people of Tublingen did, when they said, "We want a university, not industrial development."

We have an alternative. We do have an alternative. There is a model. It's not a model. I shouldn't call it a model. We have an anti-model. Maybe I should be anti-anti-model and call it...Never mind. I won't go there. We can, as they say, reclaim learning.

We can have a way of looking at learning where learning is not structured, designed, and set up to create outputs, but rather run, operated, and controlled as an unorganized, unmanaged system by individuals. I say we're moving beyond institutions in learning, toward a cooperative model, toward a knowing society, based on network knowledge. That's the model of the future.

It'll be based on software, technology, resources, systems, interactions, communities, and the rest that take learning well beyond formal education. People talked about, "What is the role of formal education and institutions in all of this?" It (the role of formal education and institutions) is to serve that.

Institutions need to adapt and get out of the mindset that they control and manage learning, and now think about how they can serve many different people in many different ways, with the resources, the learning, the coaching, the mentoring, et cetera, that they need when they need it.

We're going to get the opposite of these large, control-based universities. People say, "This is the death of the university, these MOOCs." It's not. This is the beginning of the university. The shift of the university from something big and large and available only to a few, to something much smaller, much more nimble, much more independent, a lot like community music artists -- that line was for George Siemens -- that will cater to specific learning needs. They will number in the hundreds of thousands, not the tens. They will be everywhere.

I thank you for your time and your attention. We do actually have a little time for questions, but I won't say how little. Thank you.

Peter:  Thank you very much, Stephen, for that. We do have time for questions, therefore the floor is open. We do have a microphone for those of you who do have questions. That's so that the feed can get you, and we can keep you on tape forever. The floor is open for questions.

Male Audience Member:  Reclaiming the web. Reclaiming learning. Given that the Internet has been around for a while, how come the institutions took over? Why hasn't this happened already if your predictions...? You have been making those predictions 25 years ago, right? What went wrong?

Stephen:  14 years ago. The predictions I made back in 1998, and you can check them, they're online, have come true. This is new. These are the predictions for 2014. It's not that the predictions I made back then are wrong.

Back in 1998, I wrote a paper called "The Future of Online Learning." I predicted a device that would be widely used by students. It would about this big by this big. It would cost around $300, and it would be called the "PAD."

Male Audience Member:  I suppose what I mean is we're all part of institutions, so we're part of the problem.

Stephen:  Oh, yeah.

Male Audience Member:  Why do institutions come about if what you're saying is so brilliant?

Stephen:  Why was there 2,000 years of history before there was the Internet? Why didn't they have the Internet back in creation? That's the logical structure of the question, but I'll be a bit more charitable.

If you look at the Internet, and by look at the Internet, I mean look beyond Facebook, and look beyond Twitter, and look beyond your universe in the intranet, you see exactly what I'm describing.

You see a billion web pages, 10 billion web pages. I'm not sure the exact count. You see tens of thousands of Google Groups of people getting together. Google Groups, Yahoo Groups, other groups. People getting together to talk to each other about the things that are important to them. You see informal learning like crazy.

You see things like the Khan Academy before Microsoft bought it. Where some guy just creates a whole bunch of videos.

Male Audience Member:  Microsoft did buy it though, didn't they?

Stephen:  They bought that.

Male Audience Member:  Wouldn't it be the case that all of these things will get consumed.

Stephen:  They bought that, but not the 90 other things. What you're doing is you're focusing on the big media thing, and you're missing the other 90. The reclaim the web thing is ignore the big media thing. It's the other 90 that are important.

There's another thing too. It's hard or has been hard to setup things like your own web server. It's much easier to simply sign on to Facebook, but the technology is being developed. It used to be really hard to publish a book. It used to be the only books that ever got published were published by publishers. Now, anyone can publish a book online. Yeah, there are still publishers, and they still exist.

You can say, "Why are there still publishers if anyone can publish a book online?" Because there's still a business model there. People still make money off of it. That does not negate the fact that anyone who wants can publish online.

Do you see the point? Are you happy with it?

Male Audience Member:  I think it's part of a larger discussion. I'll resign to the argument.

Stephen:  Good enough.

Female Audience Member:  If I may challenge you on a rather distinct hypothesis. Namely that people really want to restructure their own learning, or do want to define their learning in the first place. I have made the experience that a lot of people, in general, in life are actually quite happy to have control, to a certain degree, taken over and that they can rely on structures that have established themselves as successful in the past.

Even though I, personally, very much agree with you, and I'm doing a PhD, so I actually am in that position that I can restructure my learning. I can sit here and not somewhere else. I still believe that there's a lot of people who don't want to do that and actually find it very stressful.

Are you not basing, in a way, your very argument on a very small, I'm going to say, elitist group? Basically, saying that this is not very representative of everybody.

Stephen:  This is why I was really careful and even used the extended analogy, and I felt very badly at the time for going on so long about it...The analogy of touring a city.

You're quite right. Even apple pie, right? Many people would just rather buy the pie from a store that produces pies. That's kind of where I'm at. Other people if they're actually going to make a pie, fools that they are, would rather follow the directions. Then, there's this small set of people, this miniscule, elite set of people who wing it and design an apple pie for themselves by scratch.

The argument is not that everyone wants to design a pie by themselves by scratch. The argument is that people want the choice. If the only way in your life you could have an apple pie was to go to McDonalds and order an apple pie, you would find that limiting. Even if you like McDonalds apple pie...For me, the only thing I'll eat at McDonalds, the very few times that I go in there, is the apple pie, because it's good, and it's delicious, and it's got this nice hot filling.

But, if that was the only way to have an apple pie, I would be up in arms and so would everyone else. You see the difference? It matters. The nice thing about doing a PhD is you're willingly putting yourself under the direction of people who are experts in the field. It's a powerful experience, and I know you're enjoying it.

If that was the only you could do to get a job, period, now, all of a sudden, your willingness to do this takes on a different character. That different character is what people are worried about.

Male Audience Member:  On that point, I was just reminded about Chomsky's lecture "Education for Whom and for What," which was talking about the stratification, political side of things. I'm reminded of Zizek's problem solving and creative thinking or thinking as he makes the distinction. Institutions and economies depend on people executing...It's like serving the Industrial Age. You need to read the manual, so you can operate.

Psychologists come to diffuse dissent. Urban planners come to organize cities, so the dissent doesn't erupt and can be contained. Nobody really wants thinking in a political sense in our institutions anyway. Let's assume that's the case. You said perception on a collective level is somehow sort of uncorrupted or sort of clean.

What about perception of society? BBC shapes opinions in this country. Private's use a reason. Public's use a reason. CNN...It's like holding up a cardboard [inaudible 1:31:01] instead of [inaudible 1:31:03] , which is not a real [inaudible 1:31:05], but they react in a certain way. I applaud your idealism. I'm part of it.

I had the privilege to talk to Richard Storman about privacy not far from here at IET at Savoy, and he was using a MacBook for his publishing. He had to succumb to all the issues he as partially against in some ways. We have to obey gravity, I guess.

Stephen:  A lot of people say, "I applaud your idealism, but it could not possibly work." Yet, at least, in my life, it has worked. It's not a sample set of zero. When you say nobody wants the people to be educated. The universe of discourse of nobody here is limited to managers and those above managers. I don't think most people have their ambition in life to work in a factory and be told what to do.

I don't think so. They may choose that, but we have to consider the choices that we're offered. You're quite right. When you point to the influence of media in our thinking, that's...Yeah, absolutely. It's one of the really interesting consequences of the Internet. That we now have at least a billion choices of points of view and perspective.

We don't need to depend on the BBC or the Guardian or any of the newspapers that have recently been closed. You know what I mean? We have many choices. I get my news now from a wide variety of sources. One of the things the Internet has done is it's lifted this impact of particular vehicles of ways of looking at the world. I think this is really valuable.

I think it's replacing them with others, but, at least now, it's the choice thing. It used to be if you wanted news, you had to watch BBC or maybe ITV after a certain date, and I understand you have other channels now. Now, we can get our news from almost anywhere. Literally, almost anywhere.

Where I live in New Brunswick, there's a guy called Charles Leblanc, who's a self-proclaimed ADHD individual, who's been permanently banned from the provincial legislature for various indiscretions, who is one of the major sources of news about political affairs in the province. The guy is crazy. Certifiable. But, he's wonderful, and he's just as accurate as the local newspaper.

Male Audience Member:  I've got a question from Sonja Grussendorf. Looking forward 200 years, how sustainable is the change you're proposing if it's dependent on things like energy and technology? Can self-organization happen without technology?

Stephen:  Without technology, we'll be depending on self-organization, because there will be no means, except a club, to impose control, and the effective limit of a club is the circumference of the arm. OK. That was badly put, but I tried.

What doesn't go away, presumably, is a lot of the technological capacity that we have today. We know that there are other sources of energy besides fossil fuels. Wind, solar, et cetera. There is no a priori reason why 200 years from now there would be no energy. The bigger problem is probably the lithium shortage, because it makes portable batteries tricky. We'll have things like carbon nanotube capacitors as a way of having stored, portable energy.

I don't think that energy is the big problem 200 years from now. The fact that many of our major cities are flooded is a big problem, but energy, I don't think so. I think, in broad strokes, self-organization is sustainable in that way. As long as we have the capacity to communicate with each other to any significant way in a free and open way, self-organization is possible.

The bigger threat to self-organization in the future is the reemergence of authoritarianism, non-democratic or post-democratic forms of government, nuclear war. That kind of thing. Those are threats. Corporatism is a bigger threat than the lack of energy, quite frankly. Corporations are not democratic. They're explicitly anti-democratic.

Communication within a corporation is very typically monitored and controlled, and that's extending into government. Both senses. Those are more concerns. Interestingly, we're to the point in technology now where we can build alternatives. What's it called? There's an alternative point-to-point WiMAX network being setup that operates outside the boundaries of the traditional telecom system.

It's not very popular. There's no particular need for it. The thing is the possibility exists with existing knowledge that we can form a mesh network without the existing telecommunications infrastructure. Even things like silicon chips. There are enough silicon chips in the world now that even if we stopped making them, we could go out and mine them from garbage dumps and have a supply for years in the future.

It's a novel I want to write one day. Mining for silicon chips. Got to find the knowledge.

Female Audience Member:  I was intrigued very much by the assertion you've made about institutions understand personalization but don't understand personal. I would like to ask, do you think that through what they understand as personalization can they get to the personal? Can they reach the personal and through what way, perhaps?

Stephen:  That's a really interesting question, and it's interesting in an ontological sense. If it's true that we can completely define a person as a set of options, where completely means not fully completely or universally completely but completely enough, then personalization could take us to personal.

But, if there are aspects of individuals that could not be defined, a priori, as sets of options, then personalization can never become personal. You see why, right? Personalization requires defining the set of possibilities for an individual service or resource by these sets of options. If the set of options is not complete, at some point, some percentage of people will find a barrier or a limitation.

Not that all barriers or limitations are bad, but when they're imposed on you by the structure of the learning or production network, they become something people struggle against. That's an empirical question. That's not a question I'm going to say, "Well, it's conceptually impossible that this could happen, or it's practical." it could happen. We could do it.

Humans are not infinite. Nothing is infinite. Not even space. Probably not even time. Conceptually, it's possible, a priori. Whether it's practically possible is a matter of just how detailed can our technology get. People like Neal Stephenson and "The Diamond Age" with the notebook that reacts to your every learning need. That's the sort of picture that they have in mind where the notebook learns enough facts about you and throws enough switches that it becomes completely personal.

Maybe we could design a system that's sufficiently complex, but if I had to pick one, I'd say probably not. I'd say maybe it could be done, but it would probably be too expensive. It would cost as much as a professor, and it would require more money than there is in the world to provide one for everyone. I think personal is the way to go.