Loading presentation...

Present Remotely

Send the link below via email or IM

Copy

Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.

DeleteCancel

Make your likes visible on Facebook?

Connect your Facebook account to Prezi and let your likes appear on your timeline.
You can change this under Settings & Account at any time.

No, thanks

Collaboration Complexity

Technology platform development for the Hub
by

Mathew Davies

on 11 June 2013

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of Collaboration Complexity

Collaboration is powerful. The Hub's technology platform should facilitate effective collaboration. But collaboration doesn't happen by itself. A central premise of the Hub: While geographic space collapses, the space of possibilities explodes, and information proliferates.

Noise starts to drown out meaning. Collaboration is
collective creation. Fundamentally, collaboration isn't just about sharing ideas or resources. It's about collective creation: people jointly making something that didn't exist before. An idea. A vision. A project. A movement. Even a government. Technology has made it vastly easier to create and recombine information. It's also made some easy things confusing and counter to intuition. For example, email is really easy. But parsing email chains involving a dozen people is not. Making a discussion board where hundreds of people can post is easy. Following hundreds of simultaneous conversations is not. Video chat with a few people is easy. Virtual f2f with thousands is not. The vast scope and scale of interaction enabled by information technologies doesn't necessarily make collaboration easier. It often chains us to complexity of an unprecedented degree, resulting in some tough challenges: Scalability What works for one person isn't as easy for five people, and what works for five may not work for a dozen.

How about a hundred people? A thousand? Different time zones? Different languages? Collective action often remains hard. Fragmentation Knowledge isn't an orderly library; it's a cloud of billions of people and documents from everywhere about everything. Our society isn't a village or a city any longer; it's a shifting network of networks with countless nodes, trillions of bits of data hiding what (and who) is actually relevant. Social Dilemmas Even the best plans can be confounded by innate tension between the interests of an individual and the apparent or perceived interest of a group. These “commons” problems can be extremely hard to solve, because many solutions require coordinated action and universal adoption, a balance easily upset even without considering power dynamics. Friction Interacting with others only through email, video chat and virtual worlds can become alienating. Trivial tasks often become difficult by virtual proxy. In the same way that transaction costs shape the structure of firms, they shape the structure of collaboration. Not always for the better. Commercial distortion Portals like Facebook and Google+ purport to connect us, but the forces driving their development are overwhelmingly commercial in nature: monetization, gamification and revenue generation dominate all other concerns. Many collaboration and crowdsourcing startups have become co-opted by the profitability of web-based marketing for “brand loyalty” and “customer engagement”. The challenges aren't all technical. But underlying them all is complexity: of coordination, of communication, of cognition. To illustrate with a model: Consider a group of five people.

How many two-person channels are there? Ten. How many such channels among ten people? In a linear world,

the answer would be twenty... But it's not a linear world. The answer is not 20, but 45 -- there are 5 new channels for each of the 5 added. ...And it gets worse. Suppose our 5 people have just 2 unique skills each, and their solving a problem together will require a particular, unknown subset of 5 skills. They are looking for 1 out of 252 combinations (10 choose 5).*

10 people? Now it's 1 out of >15,500 combinations (20 choose 5). 60-fold worse!

This is just a model, and very simple at that. Reality tends to be, well, more complicated. Complexity is a natural outcome of the exponential growth of information. But information doesn't always have meaning. A lot of it is just noise, obscuring the signal we seek. Isn't there more to it than just combinatorics? Of course. Yet even in the best case, a mere quadratic explosion (as with the growth of pairwise channels in a network) is more than enough to frustrate progress. And while the complexity of many real-world organizational and technical problems increases exponentially in the size of input, we are frequently stuck with linear tools to solve them. Time complexity of n-collaborator tasks: Email (response-to-responses), discussion board moderation, group document revision, meetings, 360-style evals: effort required grows quadratically - O(n^2)

Subteam selection, path selection, social networking, uncontrolled versioning, unconstrained social search: fundamentally exponential growth - O(exp(n)) More computational power isn't the answer - that's what created all this in the first place! There's no magic bullet, either hardware or software. We need a fundamentally different approach to cope with unbounded complexity: we need methods of organization and orientation that will cut through the noise, help ideas and people cohere, and keep us focused on the human issues that matter, rather than distracted by obnoxious clouds of information. Effective collaboration inverts the logic of complexity. That's the platform structure to seek, to empower users and increase impact. How to make 3125 look like 5 How do you help 125 people contribute ideas to solve a problem without getting overwhelmed by noise, without losing that one quiet voice that has the best answer nobody else thought of? Here's a way using ION, an open-source application inspired by the Delphi method:

Each person submits one idea and then anonymously evaluates 5 other randomly assigned ideas, so that each idea receives 5 double-blind reviews. At the end, there are 625 total reviews of the 125 ideas. Then everybody gets to see the top-ranked ideas, and optionally modify their own. ION: Iteration 2 Now iterate the process. Each person has seen the handful of ideas collectively judged best, and will see 5 randomly assigned new ideas in this round. All who don't agree with the first-round ranking have an equal voice in redetermining the second-round rankings.

By the end of round 2, we've generated 1250 fair evaluations of 125 ideas, which themselves may have been modified and recombined. ION: Iteration 3 In each round, the best ideas gain more and more influence. Apparently different ideas may start to converge, as the best bits recombine. Each round presents the participants just a few simple tasks, yet by round 3 we have 1875 meaningful evals. ION: Completion By the end of the fifth round, 3125 double-blind evals have sorted 125 ideas into meaningful clusters, robustly bubbling the best ideas to the fore -- without the distortions of groupthink or popularity dynamics. Each participant has evaluated just 1/5 of the 125 original ideas over 5 iterations.

This approach gracefully
scales from tens to
tens of thousands of
participants. ION: Disclaimer ION is the brainchild of Michael Johnson, founder of ideologi.org, and myself. It is an example of one of the many new collaboration paradigms in the making. Facilitating collaboration Technology is constantly evolving and changing. Rather than pushing a particular technology stack, I would advocate five basic principles to inform continuing development of collaboration tools for the Hub. 1. Less is More Sometimes, we need to narrow the space of options, not increase it. We can replace the free-for-all discussion with a structured dialog, as with ION; replace long task-oriented email chains with targeted applications, such as Asana. We simplify without losing intrinsic freedom or value. 2. Many Wheels General-purpose applications are giving way to usefully specific and targeted tools. The collaboration space is still thin, but rapidly improving. Always keep an eye out for new tools and methods, and actively seek to improve the state of our art. 3. Be Research A vast literature pertains here: the theory of organizations, industrial psychology, behavioral economics, distributed AI, game theory, interface design, choice architecture, innovation, democratic governance... Let's invite ground-breaking researchers to help create better templates. 4. Open Source I'm not talking just about software, but about the larger freedom movement that's invigorating domains diverse as business, government and education. Open source can be a fountain of innovation and a foundation for new templates for effective collaboration at every scale. 5. People Above all, stay focused on what people need. It's about people, not about apps. *following Scott Page et al. Granted that Granted that While few situations are strictly combinatorial, their underlying dynamics are almost never linear. To say that a project's complexity is only quadratic is a little like saying that flying to the moon only requires a rocket.

The Hub's technology platform must be that rocket. Long to-do lists, overflowing inboxes... this isn't just complexity in theory. I'm going to describe an example of software that actually does this. But software is just a tool. The starting point is to focus on what people need to successfully create things together. {challenges} quadratic pseudolinear exponential linear I'm abusing big-Oh notation for a reason: our brain accomplishes tasks using buggy algorithms running on a one-hertz processor. I hope to discuss these ideas with you in more detail. Thanks for your consideration.
Full transcript