Creating Technology for Social Change

Community and Nature: Session Five at #MediaLabIO

  • Design and Humanism–John Maeda
  • Oversimplified–Cesar Hidalgo
  • A Language for Community Computing–Sep Kamvar
  • Social Madness–Reid Hoffman
  • Conversation: John Maeda, Sep Kamvar, Cesar Hidalgo, Reid Hoffman

The second day of the Media Lab’s spring member meeting is a daylong conversation on the theme of “Inside Out”. Featured in the conversation are a wide range of scientists and designers from within and outside the lab, who are learning from and with examples from nature. Hugh Herr begins the conversation by noting that designers are using science to translating nature into technology. From synthetic constructs that emulate biological materials to computation that resembles neural processes. Design is also driving nature in the world of synthetic biological constructs. Too often, today’s technology is overly simple–things which require manuals. But our hands don’t requre a user manual– it’s integrated with our self. We are headed to a world where the boundaries between nature and design will be forever blurred. We understand the things we make better than the things we are–we understand our phones better than our brains. In the future, designers will use living tissue to build our world.

Video of a flock of birds (sorry video embeds not working right now).

John Hockenberry highlights what to expect today: it’s less about making and more about big questions and hard issues. We often think of ourselves as separate from our tools. That may not even the case right now. John shows us a shot from “The Tree of Life” which shows a swarming flock of birds. When you look at it, you don’t see individual birds; you see a complex collective action. How might we understand this flock? We couldn’t look at just one bird. We would have to look at the entire system and maybe describe the algorithm that keeps them from colliding. But would that tell us anything about what they’re actually doing? We’re just now starting to be in a position to see the big picture in biology, psychology, and neuroscience.

John Maeda is a former faculty member at the MIT Media Lab, a celebrated artist and designer and president of the Rhode Island School of Design. His talk on Design and Humanism opens with his memories of the “infamous” sponsor week – the flock of birds Hockenberry showed reminds Maeda of the work that brings together an event like sponsor week.

Maeda is curious to embrace the world off the computer, trying to unplug. Nature doesn’t require a user manual, but it reveals itself to us if we bother to look at it. We’re in a world of information streams, but we also need to stare at things, enjoy them, understand them. At RISD, wherever he walks around, he sees wonderful things. He talks about the RISD nature lab, a collection of things like python skins, horns, rabbit paintings, taxidermized animals, and signs that make no sense. At RISD, students try to learn from these animals and find new insights from them. (Maeda notes that when animals are run over by cars near RISD, the university is called to retrieve them for the collection. ) It’s a facility which has been there since the 1950s and Maeda thinks that it has been the source of inspiration for many of the Caldecott Award winners which have come from the school. Students are allowed to check out animals and sometimes pull all-nighters staring at them. They want to understand their form– what makes them look like that.

The most popular items in the collection aren’t the butterflies, as Maeda had assumed. They are the collection of gourds. Maeda wanted to know why the gourds were so compelling. A student responded: hey don’t appear to have a function, they take up space, they just exist.

When RISD students draw things, they don’t just draw images. They draw shards– a drawing of just one part of the figure. When Maeda asked the same student about butterflies, the student mentioned that when you look at butterflies in a microscope, you can see their feathers. In the basement of the nature lab is a row of filing cabinets with samples from nature which they can look at through microscopes. Thanks to NSF funding, they are able to look at natural forms with very powerful microscopes. Maeda also shows us an image of poppy seeds zoomed to a very large scale– not just to understand why, but also how.

Maeda shows us freshman art projects inspired by nature: a sculpture made from press-on nails which is evocative of the wings of butterflies, cardboard self portraits which are complex, multifaceted faces and forms. It’s hard to believe they’re built by a single individual in a few weeks. There are human artifacts in the work, as well, like acrylic nails and pasta inspired by nature. John tells us “At RISD we teach you to have fun, but often the kind of fun that doesn’t feel like fun.” It’s not always fun learning what your body or your mind can do. But, at the same time, fun ultimately comes from discovering limits to your abilities and how to transcend them, and from stepping away from the screen and learning from inspirations in the physical world.

John tells us that art helps us contextualize things in new ways. We think of artists as people who are begging to “save the arts.” And scientists are often still in the same position. Both communities have a problem, a problem we can solve by looking for the overlaps. When we collaborate with technology and with art, the quality is incredibly important.

Our next speaker is Cesar Hidalgo, who leads the Macro Connections group at the Media Lab. Cesar says that we have not been understanding society very well.

Cesar’s story starts in the forests of 18th century Prussia. People needed wood, so they would divide the forests into cells. Each year they would cut down one cell, leaving the other cells to grow and keeping a constant supply of wood. Over time, they realised they could simplify further. They just needed certain kinds of wood and just replanted the ones they needed– moving from an ecosystem to a monoculture which depleted the overall resources of the forests. In part, we destroyed the complex ecosystem because we were measuring the wrong indicators – the yield of the forest, not the complexity or health.

Cesar tells us another story, reminding us that Adam and Eve didn’t pay property taxes. Property taxes, says Cesar, were invented around 400 or 500 years ago. To do this, governments needed to measure the number of houses. Instead of taxing houses, they developed a windows and door tax. It’s easy for inspectors to count the windows and doors from the outside. After they did this, people started to build houses with fewer windows and doors. This is another case of the negative effects of oversimplification.

We see a photo of a wedge of kiwi on a tart in a large plate of pastries. The photo might have the same number of calories as a large collection of vegetables, but the outcome won’t be the same. Until recently, school cafeterias were only required to meet a minimum calorie count. Looking for a simple way to address issues of nutrition, well-meaning social reformers put forward a standard, trying to help children avoid malnutrition. But because that standard was so oversimplified, we ended up promoting bad solutions – junk food with high calorie counts as a way of ensuring that cafeterias provide a minimum calorie level to children, inexpensively.

Mistaking cities for housing (Jane Jacob’s critique of Robert Moses’s urban planning methods) and giving people one vote between two candidates are additional examples of oversimplification Why do we do this? We’re often looking for silver bullets. Cesar thinks that oversimplification arises from the limited resolution we have when viewing the world. Five thousand years ago, the cosmos was the same as it is today. But as the resolution of our observation has improved, we have become better at understanding the world.

Three technologies form the baby steps that will help us move away from oversimplification: Big Data, Networks, and Ubiquitous Technology. Cesar shows us The Observatory of Economic Complexity. How do we make sense of the data on the site? Cesar walks us through the story of the Netherlands and cut flowers. The graphs he shows us address these questions:

  • Who produces the most flowers?
  • How has that changed over time?
  • Who else produces flowers?
  • Who’s buying the flower?
  • Who does the United States import flowers from?
  • Who does Japan import flowers from?
  • Will the Flower market be the same by 2020?

(Watch the video)

This data might help us answer questions like, how predictable is a country’s production in the future? Cesar shows us a graph which evaluates his project’s ability to predict which countries will produce which products in the future. Cesar tells us that his models suggest that China will be exporting cars (as well as Tunisia!) and that Vietnam is likely to be exporting calculators.

Large datasets can also help us understand cities. Using Place Pulse, anyone can collect data about people’s perception of the safety, uniqueness, and class of different places in a city. It’s also possible to run experiments: by adding graffiti to the images of a community, it’s possible to see just how much graffiti influences people’s perceptions of the safety of a neighborhood.

How do we move from one bit democracy (a vote for one candidate or another), to kilobit, megabit or gigabit demoncracy, increasing the bandwidth of our systems? The Participie project makes visualisations a two-way street, making it possible for people to edit the US federal budget. By collecting different public views on the US budget, they hope to create an aggregate picture of what a large number of people want the budget to be.

Cesar concludes by reminding us that oversimplification is a cognitive and societal problem. We may be able to move forward by working with big data, learning from networks and making more complex predictions about the future.

Next up, Sep Kamvar introduces a new programming language which he and his students have been working on, promising to write some programs together with us in the session. Sep thinks programming is important and wants to help us see that. He also wants us to see how highly technical acts can be broadly humanistic.

Sep has always been interested in the interplay between people and web search: looking at the people behind the queries to do personalised queries, looking at the people behind the documents to do social data mining. He has also been interested in connecting the people behind search with people behind the documents to facilitate question answering. Whenever he sits down to write these systems in code, it’s much more difficult than it should be.

Why is it more difficult than it should be to use code to connect people? Programs have been designed to coordinate the activities of machines. But Sep wants to coordinate the activities and make requests to people. When we think about these programs in our heads, it’s simple: we make requests and coordinate responses. But when we think about them in code, we have to think in a completely different way. This difference between how we think about a program and write it in code is a problem.

To illustrate this, Sep asks us to imagine how to program a lottery for MIT students. The psuedocode looks as follows:

accept lottery from MIT students
when I get a lottery entry from an MIT student:
put the entrant's name in a hat
thank the entrant
after 1000 lottery entries,
pick the winner
empty the hat

Sep shows us the incredibly long Java program required to do that. He then shows us a 14 line psuedocode which he really wants to use to write that code.

The answer to this is Kamvar’s new programming language: Dog. Things that are hard in traditional languages involve identifying people, talking to people, and integrating different programming languages. Asynchronous state management is also hard. Dog makes these things easy.

Consider for example the case of language integration. In the current polyglot world of programming, we often build lots of infrastructure just to make two things talk together in a simple way. Sep wants to make it easy to import methods from a variety of programming languages.He shows us the laborious complexities required to do something fairly simple – send an email to a user. The process involves application servers, XMPP gateways, state machines, conflicting APIs)

Right now, to write a programming language with asynchronous inputs, we think of it as a control flow. But when we write it, we have to break that flow into lots of pieces and save those states in a database. We set up an endpoint for each possible interaction and monitor the state. What Sep wants to do is write things in the control flow that he’s actually imagining, something like this:

Listen to users for tickets
On ticket Do
user = Person from ticket
resolution = ask representative via feed to process on ticket
notify user of resolution

At this point, Sep shows us example code written in Dog. He shows us a Dog application for coordinating conference reviews, in 9 lines of code. Sep asks us to imagine what it could mean to make coordinating action an easier thing to accomplish. If we can coordinate people reviewing papers for a conference, can we take a bigger step and invite people to co-create a constitution?

Why did Sep create Dog? In a world where collective community action has a lot of power, making it easier for collective community action will have a big impact on society. Secondly, we typically write software to command and control people. But you can’t command and control free people– if you’re writing code for a free people, you want to inspire them for co-creation. One way to do that is to write readable and poetic programs which affect the real world. “What else is a programming language than a description of what you want to create?” he asks. “I believe that if you make beautiful tools, people will do beautiful things with them.”

The design of a programming language is an art– and like all art, it aims to make an impact on the world. Dog is coming in September, and Sep invites us to sign up and learn more.

John Hockenberry asks Sep what inspired him. Sep contrasts Dog with the Amazon Mechanical Turk, which applies a command and control model to humans. When you do that, people try to cheat you and mistrust you. The problem isn’t in the algorithms; the problem is that if you ask people to do rote uninteresting work, they won’t be inspired to do it. John concludes by asking Sep to write a program which conducts the role of panel moderator! 🙂

Entrepreneur and venture capitalist Reid Hoffman, co-founder of LinkedIn, is the final speaker of the first session. Commenting on Sep’s presentation, he remembers a famous New Yorker cartoon: “On the Internet, no one knows you’re a dog.”

Reid wants to talk about the interplay between humans and technology, asking if the Internet has gone crazy on Social. Some large percentage of the human race interacts with Facebook every day. With companies like Zynga, we see people who don’t identify as game-players carrying out virtual farming. So much of work, commerce, and music have been transformed by what we see happening. Have we gone socially mad, or are we mad about social? Reid reminds us of Aristotle’s argument that humans are political, social animals. Reid thinks that the trend towards social is an increased humanity of how we do design.

What does it mean for a technology to be social? The first important concept is identities. In the first version of the web, it was much more common for identities to be anonymous- people identified with usernames and preferences. Reid often chooses the username “quixotic”, in part as a reflection on the quixotic idea of forcing someone to pick a single identity. More recently, we’re using our real identities. Next, we’re creating shared space online around our relationships. Our social spaces online create adjacencies and connections that are different from the physical world which can be interposed in a dual layer between the physical world and a social layer which is always with us. In 2003, when Reid started doing things, people in Silicon Valley thought social was just a feature. Reid thinks social is a platform, a new way of understanding and navigating the world.

Has social played out? Because relationships are so deep in our human experience, we have yet to see how this plays out. Reid directs us to two of his recent investments: Wrapp and EdModo. In both cases, Reid sees an opportunity to transform two spaces – the giftcard space and the education space – through embracing social networks more deeply and thoroughly.

Most of us forget the controversy around the launch of Facebook’s Newsfeed. Newsfeed brought in the idea of group tagging– before it was introduced, no one had the idea. And now it’s part of the social fabric.

Another question raised by social is: do we have one or many identities? Normally, we have one identity, and all our relationships intrinsically go to one person. On the other hand, we have different ways fo expressing that identity in different contexts: our professional identity, our home identity, our identity in public. Social technologies in different spaces are expressed differently in those contexts.

Does every company need a social strategy? We’re already living in a networked world. This changes how we carry out a whole range of strategies: open networks, brands, supply changes, customer relations, and employee companies. But when we think about whether a product needs to be social, we need to think about the primitives of social (identity, relationships, shared space) and ask if those primitives are relevant. “The likelihood that there will be social toilets is very low.”

While it’s intrinsic that you need to be thinking about networks, it’s essential to apply design to that. There’s a fine line between the wisdom of crowds and the madness of masses. Reid quotes Foucault, that madness is what happens when we diverge from the pack. But sometimes we agree with the pack too much. Sometimes we need to be mad enough to prove that the world is round. How do we evolve our ideas from madness to sanity? With social technologies, Reid says we need to keep an eye out for what’s human.

What are the limits on behaviors that respond to network ideas? Reid reminds us that social products only work when they’re deeply embedded in human experence. He only invests in the 7 deadly sins. Family is important, work is important, media is important. Next, we need to think about activities which though they could be social, they are usually social.

Do things increase in value when we make them social. Reid thinks there is always a set of things for each task which is individual, and a set of things which are social. When writing a document, we have a loop between creation and sharing.

Are we getting better at understanding complex systems? Cesar responds that sometimes we understand the simple rules which lead to complex behaviour. Sometimes complex rules lead to simple behavior, and we’re starting to understand those complex rules. Cesar suggests that networks can also help us understand physics better, as we go beyond formulas to do pattern matching.

Continuing the discussion on powerful metaphors, John Maeda refers to Rafe Sagarin’s article on 10 lessons you can learn from nature. He argues that we need to choose to draw positive inspiration from nature.

Hockenberry then asks Reid Hoffman what areas to expect disruption. Reid thinks institutions need to come up with new ways to deploy what they do outside the classroom. He points us to LinkedIN skills, a mash-up between LinkedIn’s database and wikipedia to try to connect people who have complementary skills.

“What institution is going to make a decision with the high resolution information we get from the social world?” Hockeberry asks. Is an institution going to help us decipher this problem, “or do we just live with a high resolution photo of how screwed we are?” Reid thinks that some things do take us by surprise, like the Arab Spring. But when governments take network thinking in mind, it’s often not one big decision. It’s a pattern of much smaller interactions with networks. For example, during Obama’s first campaign, he asked people on LinkedIN what he should be doing to support small businesses. Some of those ideas actually made it into policy. John Hockenberry expresses a hope that data can help governments anticipate things like the Arab Spring, capture them, and respond.

Hockenberry asks Sep Kamvar is programming is translation into machine language, or a fundamentally new thing? Sep responds with a story about ant nutrition. He describes the beautiful and process through which a colony of ants cooperates to carry out a highly complex digestive process, fetching food for the larvae. The ants rely on two things: communication and gifts. In technology, communication and gift are APIs. If we can do that with technology, we can also achieve beautiful complexity of a higher order.

What inspires the panelists? Cesar is inspired by people and by self discovery. Sep asks us which is the organism: the ant or the colony? Is it just the colony, or the entire ecosystem? Reid is inspired by the possibility of modifying human nature through how we evolve as a society. He hopes to evolve human consciousness, citing Steven Pinker’s claims that we live in some of the most peaceful times in history. Maeda is inspired by the experience of moving between different environments, learning from them and challenging them.