My story: tech + collective intelligence for systems change


My story begins in a converted school bus on a residential street in Vancouver, Canada. It weaves through intentional community, meditation, homesteading, and software development, and arrives at Internet of Humans through three volunteer projects that all come from different angles on this question: how can we use technology to build smarter, more collectively intelligent patterns for systemic change?

Through a project I started a couple of years ago, currently called the Collective Intelligence Network (, I’m trying to make sense of the disciplinary silos of research that use the term “collective intelligence”, find common threads, and identify ways that the best of what we’ve learned about cooperation can be embedded in technology.

On a more concrete level, in co-operation with The Open Coop in London, I’m building a protocol and code for organizations to share information about themselves that can be automatically assembled into maps and directories to make movements visible. The aim is to get beyond the multiple, overlapping, incomplete and difficult-to-maintain maps of organizations like co-ops or ecovillages, and create a way for those nodes to very easily host structured data about themselves on their own web sites. Data can be crawled and aggregated, based on filter criteria, so that directories and maps can be built and updated from the authoritative source of the nodes themselves. We’re starting with existing data specifications ( and the most widely-available web technologies, and plan to launch a prototype next month.

The third project is in the research stage: could we build a mutual-credit digital currency that would support economic solidarity among next-economy organizations and individuals, outside the tilted playing field of debt-based money?

All three of these projects are motivated by the sense that the possibility to create a different world – or play a different game – exists in the spaces and connections among us. Many of those connections are mediated by technology. Can we build technology that fits the infinite game of regenerative life?

A hurdle that both motivates and hampers these projects is the nature of our economy. In an economic system that rewards the externalization of costs, that is predicated on competition, and that is destroying ecological and human potential, what tools can we use to leverage the benefits of doing things differently into economic viability and abundance?

And, personally, what tools can I use to be able to work on things that matter and still eat?

Looking forward to the conversations!

Adam | | |


Greetings and welcome. Quite a task you have set out for yourself, mapping human collective intelligence. I was reading your Making Sense essay on your website. Interesting analysis into groups, crowds and swarms. Looking it over I got the impression of it as a kind of journal for a journey you have been on for some time. I recommend others have a look at the Collective Intelligence website.


Hello @amckenty, and welcome. This is really interesting, as it transposes the usual free riding problem of human communities at the organizations’ level. My organization needs to do some work, so that a common good (let’s say, a directory) can result from the aggregation of everybody’s work. This work is typically uninspiring: housekeeping, like renewing certificates for the website or turning in your VAT return. And if my org does not do it (or does not do it well, for example with regular updates of the data), it can still use the common good. What are your thoughts about how to incentivize organizations to do this work, with an acceptable standard of quality?

I have long thought that important data on organizations already exist, stored in very official state databases (Edgeryders, for example, has a legal obligation to keep some information in the Estonian Company Registration Portal. Problem is, those data are not open. Many states have sprouted publicly owned companies that make money selling them (this predates the Internet, by the way). I have long sympathized for the effort of Chris Taggart and his OpenCorporates, who are trying to build a worldwide database of unique identifiers of companies. Could this be an alternative to what you have in mind?


Thank you John! I have not thought of that essay as anything like a journal entry before, but in a sense it is. With your comment in mind, I see it as a static projection of a winding process of puzzling that has occurred over several years, and that remains incomplete.

It has received a bit of attention from people in the academic collective intelligence world, and I will probably develop a more formal version in the future with collaborators.

I’m really interested in the areas where CI research and practice intersect. Innovation in collective process doesn’t (usually seem to) come out of formal research, it comes out of practice. But research can help us understand practice, and forward the process of bringing innovations that work toward the mainstream (highly publicized examples of this include Google’s Project Aristotle and CI studies from CMU and MIT in 2010). In this field, research is a sort of anti-Sisyphus mechanism that builds intellectual flying buttresses supporting the more interesting work of actually cultivating intelligent communities and high functioning groups.

I’m really interested in the feedback loops within the spiral of innovation > research validation > (contextually appropriate) best practices > adoption, and how each of these stages can be accelerated.

Or more simply: how do we update collective guidance systems fast enough to avoid going extinct?

This applies across many levels, from very small groups to planetary governance.

I am very keen on feedback from the Edgeryders community! I hope to include some of you in a series of online collective intelligence salons that I’m launching in a few months.


Excellent questions @alberto!

Regarding incentives, I think there are a couple ways of looking at this. One is a sort of rational economic perspective. In this case, organizations benefit from being shown in maps and directories because it brings publicity and makes them findable for the people they want to be found by. This incentive already exists, which is why many organizations do maintain their data, more or less, in centralized directories. What we’re doing is allowing those organizations to consolidate the effort of data maintenance for many directories into a single point. That point is within the admin interface of the organization’s own website.

The second way of looking at incentives is: how do you actually get people to do things? This is more about psychology and less about economics. Even when there’s an incentive, annoying housekeeping tasks like the ones you mentioned often don’t get done, not because they’re not a rationally good idea, but because nobody gets around to doing them!

Combining these two, our adoption strategy is something like:

  • Partner with existing networks to reach out to nodes at scale. Ideally nodes will receive invitations from multiple networks.

  • Make it super easy to adopt, with initial data automatically populated from the node’s website.

  • Use gamified reminders and notifications to encourage nodes to update or check their data as needed, with the lowest possible barrier to action (key to this is the location of the update process within an admin interface where they already are, not requiring external URL or login credentials).

  • Add features that deliver quantifiable value (syndication of event listings, needs/offers, etc.) to increase the basic incentive to participate.

This strategy remains to be tested. Feedback and suggestions are welcome!

Regarding your second question: you’re right that all the data probably exists elsewhere already! But, after spending many months working with a few specific network organizations towards sharing data between very much overlapping directories, I am not convinced that any kind of centralized authoritative directory is the way forward in the long term.

I am completely in support of the OpenCorporates project. I think it is relevant to but not a substitute for what we’re working on because:

  • It doesn’t include the kinds of human-relevant information we need (things like mission statement or organization description)

  • It’s not organized by values or intent (we want to help movements for a more just and sustainable world see themselves, and OC doesn’t include metadata that captures these intentions)

  • We need to include groups that are not defined by a legal entity

More abstractly, a beautiful thing about moving authoritative data into the hands of the nodes themselves is that can allow the node to decide what about itself is important. With centralized directories the directory decides what about the data subject is important, by defining the data fields. This means that, even if you have shared protocols, it remains difficult to make data usefully interoperable because what different directories think is important doesn’t necessarily match. Even more generally, data out of context loses meaning. Corporate registries are a informational context that is institutionally defined, for purposes and context of the institution, not the organization the data is about. This relation of context and meaning might have something to do with why filling in government forms is an almost universally uninspiring task!

With self-defining data we can turn this around. Hopefully, the better fit the node is for a given directory, the more what’s important will be a match between the two, and the node’s presence within that directory will be appropriately rich.

This approach is experimental, and will remain a balance between defining taxonomies that work for directories, and letting nodes define themselves through a wide array of possible fields.

If any of that is excessively obscure or doesn’t make sense, I am happy to clarify!

Again, very interested in feedback from the Edgeryders community about all of this. Thanks for the conversation.


Do you see machine learning playing a role in this?


Not something I’d thought of before @johncoate. Probably not initially. For now, we’re aiming to keep the technology as simple and boring as possible!

In the future, though, there is one spot where machine learning could potentially play at least an indirect role: enabling a context-aware fuzzy mapping between the words that organizations use to describe themselves and the words that individuals or directories use to search for them. This is a much more general problem that I don’t imagine tackling from scratch for this, but work that’s been done using ML for natural language processing could be useful.


Historic footnote: ML was mainly underwritten at first by huge multinationals with mountains of documents that they didn’t have the people to properly catalog so they would know where everything resides. After they got it working fairly well they started looking around at what else they could train it on.