Banner general

Diversity trumps ability: why senior civil servants, political leaders and entrepreneurs are teaming up with hackers, squatters, artists and activists in the name of stewardship

Alberto's picture

Left to right: Actor Bembo Davies, lawyer Patrick Andrews and sharing economy serial entrepreneur Robin Chase make pasta together during Living On The Edge 4

The few make decisions that affect the many: this is a common feature of all governance models that work at the national scale, including the democratic ones. Some scholars – myself included – are looking to deploy open decision making processes, so that many people might participate through collective intelligence dynamics; but such processes are nowhere near ready to go. Most people appear to think that the concentration of decision making power is a necessary evil, but it does create a tension between the few in the decision room and the many who stand outside it.

Diversity trumps ability

In 2004 economists Lu Hong and Scott Page came up with a radically new idea: diverse group perform better in solving complex problems than groups composed of the best people available. In a nutshell, diversity trumps ability. It works like this: imagine a group faced with a difficult problem. "Difficult" in this context means it is not even clear what kind of a problem it is. The only way towards solving it appears to be trying to attack it with different approaches, and hope that one will work. Further, imagine that each person in the group has mastered a limited number of approaches to problems, and that each approach solves well a few types of problems, but is useless on all others. An example could be a car whose engine won't start. The problem could be of mechanical nature; or of electrical nature; or in the software of the onboard electronics; it could even be that there is no fuel in the tank. If all you have to throw at the problem is ten mechanics, even very good ones, that might not help you if the problem is in the electric circuitry. A group with diverse skills can try more approaches, and so have a higher probability of coming to a solution. It follows that, when faced with a difficult problem, your best bet is to build a very diverse group to attack it (paper).

This seems to make sense. So how come political and cultural élites tend to be so homogenous? Why do we not see more skaters in ministries, athletes in boards of directors, circus artists heading charities? There are many reasons for this, but one of the most important is that collaborating in diversity is hard. Cultural differences get in the way of communication; ways of working do not match; divergent interests result in conflict and selfish behaviour.

In the global social innovation scene, where I find myself more and more as of the last few years, at the moment we have a really difficult, pressing problem: the state, after centuries of building public assets (from road networks to hospitals, from libraries to mass literacy), seems unable to continue to steward many such assets, and is in retreat. Can citizens come together as communities, step into the breach? And now?

Stewardship by communities

The global Edgeryders community has collected over a hundred case studies of ordinary citizens who decide they care about some asset that belongs to everyone and benefits everyone, step forward and get to work to preserve and enhance it. Their stories are amazingly diverse, from the pensioner running the only botanical garden in Montenegro to the bunch of Californian hackers sitting in an abandoned McDonald's re-engineering the technology to read the digital photos taken by the first Lunar Orbiter in 1966. These stories inspired us to set up Living On The Edge 4, a conference dedicated to citizen-led stewardship.

No point in waiting around, hoping that someone would do something about it. We don't exactly see anyone qualified to do the job. The state? In retreat. Private business? Terrible track record, just look what happened to energy utilities in Russia or railways in the UK when privatization kicked in. The only option left is to rise up to the challenge ourselves. Are we qualified? No way. So, we are coming together to learn how to become better stewards, as citizens. This is really difficult, and to do it we need not only ability, but diversity. In fact we think we weed the most diverse group we can put in the field. Living On The Edge 4 is run by a team of curators the like of which the world has never seen. It includes Fabrizio Barca, former cabinet minister and presently general director at the Italian Treasury; Caroline Paulick-Thiel, German activist who spearheaded the effort to defend the pocket-sized utopia of a href="">Prinzessinnengarten in Berlin from speculation; Swede Amelia Andersdotter, the youngest MEPs in history (elected at 21!), now president of the European Pirate Party; German open source hacker Matthias Ansorg, winner of the first European social innovation competition (a man who bought a 1968 firetruck and is turning it into a fully bespoke open source mobile apartment!); Robin Chase, serial entrepreneur in the sharing economy and Bostonian patrician; and Nadia El-Imam, Afro-Swedish engineer and entrepreneur, currently serving as Edgeryders CEO.

The job of this exceptional team is not that of leading, but rather that to help the Edgeryders community to map out the terrain of stewardship, not very well known as yet. In line with hacker tradition, the agenda of Living On The Edge 4 (and that of the hackathon preceding it) is set by the community; curators are there to connect each talk and workshop to each other, highlight similarities and differences, try to get a sense for what works and in which circumstances.

Trial and error in Matera

Can senior civil servants, politicians and high-profile entrepreneurs really work side by side in relative harmony with hackers, squatters and digital nomads? We are aware of the risks of misunderstandings and even clashes, but on the whole we believe that yes, they can. Certainly we did our best to set the stage for a fruitful, rewarding interaction. First, we carefully handpicked our curators as people who are, certainly, bearers of strong cultures of their own, but are also curious of the cultures and approaches of others. Second, we did not just put them in contact with each other, but put ourselves forward as stewards of those new relationships, offering to explain to the ones what might look weird or unsettling in the behaviour of the others. And third, we chose for Living On The Edge 4 an inspiring physical environment: the unMonastery in Matera, a place where experimentation is normal, and errors are acknowledged and accepted as part of the process, rather than frowned upon.

Ours is just an experiment. It might fail, or be an extraordinary success. Whichever way it goes, we will try to learn from success or failure, to make it a milestone in the journey that leads to better methods to attack, together, the ever more urgent and global problems we face. How the journey goes depends on you, too: if you want to help Fabrizio Barca, Amelia Andersdotter and the others to get there, join the People of the Edge in Matera.

(Translated from the Italian and reposted from:



This is a great post, glad you translated it!

Nadia's picture

I think it is safe to say the experiment worked.


Great Posts, comments from this side of the 'other'

Eimhin's picture

Funnily enough I was just chatting with @alexwhitcroft about this this morning re independent work we are both engaged in:


 I'm finding local integration with civil servants to be a good appraoch to getting things done...any 'site' is informational tooling, but the user community, and the service recipients , clearly defined, are where the tool gets used. Do keep me up to spped with it as it progresses. We have an applicable location here.
I really feel that a community focussed version of this is important... when rome fell the peripheral mud huts fell into the ground with it... not so in the built environment of today, there is a need to rebuild from inside out.
 Tailoring this to community focussed projects around a variety of forms of manifest distributed systems, processes and means of production is key.
When gathering people, remaining cogent of who what and what-variety of skillset is important.
Not everyone in the same place who likes horticulture, and everyone who likes digital fabrication in another place, that won't work...
So, taking early christendom and the rebuilding of some form of balance in early european society as a turn on a the gyre that is now repeating - it is useful to arrange people into groups with enough feedback loopage in given disciplines to remain productive, enough network capacity to ensure this nonlocally, and enough local diversity to be able to apply the rich variety of skillsets necessary in such a way as will be welcomed and adapted by local communities.
The reason the monks were accepted by local chiefs was their capacity to protect water supply, brew beer, make honey, teach metal work via technological distribution re: the forge and mill, teach new agricultural forms,educate your kids, etc etc
So making the spread of expertise wide enough to accommodate a wide variety is key.
We have a software cooperative forming here, if you have need of co-developers, and if this is OS, then we would be onboard and willing to level up re: application of local community development tools
Not to mention a local test pilot site of a certain historic significance in the Irish midlands.
Thats it for now Alex, just checking in.



Oh and the particular Law you are looking for:

Eimhin's picture

Consciousness is not local. It is distributed.

That this is the case can be gleaned from anthropological studies of tribes with no form of art who shared both waking and sleeping consciousness with the animal realm providing the characterisational content that built the early notions of the 'persona'. Those first cave paintings? - Was it any Bull, or was that Great Running Bull, a hero of legend? The latter is more likely the correct interpretation given that  consciousness itself is share in communis.

So, consciousness, in the widest sense, is non-local. Therefore we can drop what I call
'ego-friction'. The small town dynamic that makes a mess of stewardship when it comes to notions of self-importance making Trolls key holders to what should be shared and open resources.

But back to the point, the Law in question is that feedback in human communication systems facilitates the natural function of consciousness which aligns with evolution in the generation of complex order as a means of dissipation. Those who get complexity theory and its beautiful corollaries in thermodynamics please shout! 

This means the more diversity among entities, the more entity-nodes on a system, the faster the rate of latency in communication, all of this amounts to an accelerated rate of order-generation in the subject of focus.

Take for an example the strange result from the Australian medical preschool tests. The ones who answered the most questions were the best at the math section - Everytime! Why?
Because focussed concentration has the effect of the organization of the subject of focus via the generation of complex order in that field.

Diversity Wins.

Why it wins is important, and if people get this, they will tune up their game in terms of active networks and in terms of action on the ground, by organizing around this principle.

Good Luck!


How come political and cultural élites tend to be so homogenous?

danohu's picture

Once upon a time I was involved in a campaign to bring random selection back into politics. The idea was that some representatives should be chosen at random from the population -- like a jury, and similar to what happened in some ancient Greek cities. Citizens' Juries are a similar (and easier to implement) idea.

The problem is that, alongside diversity, you need trust and efficiency and domain knowledge. LOTE4 didn't just go for maximum diversity. It found a mix that combined diversity with enough trust and community to glue everyone together.



Look at the assumptions

Alberto's picture

Wow, thanks, Daniel. You force me to go back to the original paper by Hong and Page, and I understand it better as a result. It is predicated on the following assumptions:

  1. The population is large (predictable behaviour of statistical process).
  2. The problem is modelled as a large number of possible approaches (2000 in the computational experiment). Each approach is mapped to a value between 0 and 100. The value V_i of approach i is drawn from the uniform probability distribution of numbers between 0 and 100. So, the value of approach 187 might be 43, the value of approach 1412 might be 86 and so on (notice this assumption, I'll come back to it). Groups of solvers are trying to find the maximum possible value V that they can find. Approaches are imagined as points in a circle. 
  3. People differ among them because they have different heuristics (they "look in different directions").  A heuristics looks like this: start at the status quo (which is a point on the circle); move clockwise a certain number of steps to another point on the circle and look for the value of the approach at that point; repeat this a certain number of times. Formally, a heuristic is made of three integers. For example, (1, 5, 3) means:
    1. Start at the status quo
    2. Move 1 step clockwise and evaluate the approach you find at that point
    3. Move 5 steps clockwise and evaluate the approach you find at the new point
    4. Repeat step 3 until you have collected 3 values.
  4. Performance of individuals is measured by the maximum value of the approaches they find.
  5. The way groups work together is the following: each individual goes through the process of step 3. They all compare the values of their solutions, and keep the maximum one. so, critically, there is no cost assumed in sharing information.
  6. Diversity is measured by comparing heuristics. It varies between 0 and 1. Diversity 1 means there is no overlap between the approaches tried by one participants and those tried by the other participants. Diversity 0 means all participants take the same approaches.

Like all models, this is an approximation. I trust the proof (and honestly, it makes sense to me, though I have not tried to reproduce it). Mathematically, the result is driven by three assumptions that might not be true for all real-world problems, or at least merit further discussion:

  1. All approaches work to a certain extent. Any approach is assumed to have a "value" between 0 and 100.  Some problems do behave this way: for example, fixing Edgeryders means going to the Task Manager and executing any one of 200+ tasks. An approach might complete only 1, another one might complete 50 etc. On the other hand, some problems have binary outcomes: either you get your car to start or you don't. 
  2. The probability distribution you are drawing the value from is uniform. This looks harmless, but it makes trying approaches at random very attractive. The expected value of an approach randomly drawn from a uniform distribution on the 0-100 interval is 50. On the opposite end of the scale, if only one approach leads to the correct solution (value 100) and all the others have value 0; and if there is a large number of approaches, the expected value of a random approach tends to zero. This is why there is no experimentation in life-and-death situations! 
  3. No cost is assumed for information sharing. This makes large groups attractive (although this is somehow counterbalanced in the computational experiment part of the paper by the opposite effect of fixed and finite set of heuristics).

It is up to each of us to figure out how much we believe all of that for the problems we are facing. :-)



Thanks for that very detailed

danohu's picture

Thanks for that very detailed comment, Alberto. I agree it depends a lot on the assumptions. It also misses out the value of smashing together different conceptual frameworks, coming up with approaches that no one person could have come up with herself.


Recent site activity

Fri, 23/06/2017 - 17:19
Fri, 23/06/2017 - 16:39
Fri, 23/06/2017 - 16:34