Organising for neurodivergent collaboration

Collaboration as an evolutionary force


If autistic people can’t always see the depth of the “bigger picture” of the office politics  around us it does not in any way mean that we don’t see the big picture. In fact we are aware of the big picture and often we zoom in from the biggest picture right down to our immediate context and then back out again, stopping at various levels in between that are potentially relevant to our context at hand. Office politics only distract from the genuinely bigger context. Accusing autistic people of not seeing the bigger picture perhaps illustrates the social disease that afflicts our society better than anything else.

Neurodiversity friendly forms of collaboration hold the potential to transform pathologically competitive and toxic teams and cultures into highly collaborative teams and larger cultural units that work together more like an organism rather than like a group of fighters in an arena.

Evolution has mastered a number of similar phase shifts in the past. Consider the evolution of multi-celled life forms. Single-celled micro-organisms have not been replaced, but they have been complemented with a mind-boggling variety of more complex multi-celled life forms. We now know that our bodies harbour of more bacteria than human cells, and the vast majority of these bacteria are in a symbiotic relationship with our human cells. Consider this masterpiece of evolution for a moment. Many billions of collaborating cells and micro-organisms form what you experience as “you”. Statistically speaking our bodies are highly collaborative ecosystems of microscopic entities.

Yes sure, we still need an immune system that protects us from less friendly bacteria, but even such microbial invaders that make “us” sick can also be seen as being part of a bigger collaborative picture. The interactions between “hostile” microbes and our bodies represent a feedback loop of mutual learning, and over longer periods of time, sometimes over many generations, we learn enough about each other to not only coexist, but to even depend on each other in symbiotic ways for ongoing survival.

Evolutionary biologist David Sloan Wilson observes that small groups are the organisms of human societies. This should provide all of us with food for thought and it has massive implications for the gene-culture co-evolution that characterises our species.

Humans are not the first hyper-social species on this planet. Insects such as ants offer great examples of successful collaboration at immense scale over millions of years. Charles Darwin and other early proponents of evolutionary theory appreciated the role of collaboration within species and between species, but many of these early insights including related empirical observations have been suppressed within the hyper-competitive narrative that has come to dominate industrialised civilisation.

Autistic collaboration vs non-autistic collaboration


At a fundamental level, the ability to communicate is limited by the factors outlined in this article. Developing shared understanding is hard work. Always. For all people.

Autistic forms of communication within a neurodiverse team and within a psychologically safe environment actually impart a collaborative advantage to the entire team. The fact that most (all?) autistic people are incapable of holding a hidden agenda and don’t play social games minimises (if not eliminates) large and small sources of deception that afflict all traditional hierarchical organisational structures.

Non-autistic communication protocols make life bearable from a non-autistic perspective by injecting plenty of culturally expected pleasantries (exaggerations and small deceptions) and social cues into conversations, and thereby make it very hard to identify the larger deceptions that a minority of people weave into their social game. Resulting mismatches in expectations are easily explained away as unintended misunderstandings.

Non-autistic people seem ill equipped to recognise how all the little exaggerations they use on a regular basis, such as “you look great” (when you look and feel sick), and small deceptions such as “sorry, I have an important meeting to attend to” (when someone prefers not to help you and hoards information for personal advantage) or “I helped develop a great product” (when the person was not involved and only got to know the finished product and never contributed any feedback to the development team) over time add up to a non-collaborative and potentially toxic culture.

The result is depressing, not only for autistic people. It is time to recognise the key role of autistic and otherwise neurodivergent people in cultural evolution and in recovery from collective insanity. This does not mean that autistic people don’t have communication challenges, Samantha Craft has written an excellent article on this topic. However it pains me to live in a broken “civilisation”. I am working on educating people about the thinking tools at our disposal that can assist in minimising suffering and in paving the path into a more humane social world.

The role of technology


The technologies we develop and use tend to reflect the level of collaboration and competitiveness within our culture. In our role as conscious designers of technology, humans have the potential to influence the level of collaboration in our culture in profound ways, especially in a highly networked digital world. This recent interview between Douglas Rushkoff and Cory Doctorow offers a good overview of our current relationship with technology.

In a world that is dominated by global proprietary platforms the linear language we use for communication is rendered useless by biased algorithms and perversion of potentially valuable concepts. For corporate and government politicians who are drunk on the drug of social power life is a popularity contest. The need for learning and deeper levels of understanding is minimised, and the lives of others become secondary considerations.

Time and trusted collaboration are our scarcest resources. The former is a hard constraint and the latter is the critical cultural variable on which our future depends.

We have reached a point where human societies can choose between a “collapse of human ecological footprint” based on a conscious and significant reduction of cultural and technological complexity or an “ecological collapse, including human population collapse” resulting from a perpetuation of the behaviours that are slowly but surely killing us all. Realistically both kinds of collapse will occur in parallel, and some communities may be able to avoid the latter form of collapse to a larger extent than others.

The only way of avoiding bias and dangerous oversimplification is to perform ecological accounting in terms of relevant physical units. We can continue to live in cities and rely on science and specialisation to develop complementary skills, bodies of knowledge, and technologies, but we will have to rethink how we collaborate and manage genuinely scarce physical resources at a fundamental level.

Regardless of what route we choose, on this planet no one is in control. The force of life is distributed and decentralised, and it might be a good idea to organise accordingly.

Human scale vs super-human scale


One important dimension of human cognitive limits relates to the number of relationships that humans are capable of maintaining,

“Study after study confirms that most people have about five intimate friends, 15 close friends, 50 general friends and 150 acquaintances. This threshold is imposed by brain size and chemistry, as well as the time it takes to maintain meaningful relationships”, Dunbar says.Scientific American, September 2018

These numbers guide my thinking on human scale  and have shaped the NeurodiVenture operating model that limits the size of good company to 50 people.

Unpopular fact: Super-human scale organisations of more than 50 people are collective delusions. In particular larger organisations that contain structures of command and control are not only learning disabled, they are also also detrimental to mental health and trusted collaboration.

When people complain about living in a filter bubble the problem is home-grown, the result of toxic levels of enforced cultural uniformity at super-human scales.

The NeurodiVenture operating model not only raises neurodiversity as a top level concern for good company but by imposing a hard limit on group size (in the case of S23M enforced by our company constitution) it also ensures that every member of the team has spare cognitive capacity for building and maintaining trusted relationships with the outside world, whilst at the same time encouraging creative collaboration for life.

The more we help each other to question in ways we otherwise wouldn’t – and correspondingly discover new insights about the world and ourselves, the more we are able to learn from each other, and the more we start to understand each other. The gift that members of NeurodiVentures bring to the world is the (re)generative potential of all the trusted relationships that they co-create.

Now contrast the NeurodiVenture setup with a traditional hierarchical organisation with several thousand employees.

Firstly the hierarchical social power structure imposes a top-down approach to declaring the organisation’s purpose and organisational values – in ignorance of what the people that make up the organisation are actually interested in and care about, leading to a cultural straightjacket of what is possible – a tiny Overton window that limits the conversations within the organisation.

Secondly, the “career opportunities” offered by the organisation by definition imply a hierarchical career ladder and send a strong signal to all employees that in-group competition is the route to professional success, and since – again by definition of the pyramidal structure – ladder climbing opportunities are limited to a minority, many employees will only stay a limited time and then choose to seek opportunities in other organisations (usually within three years and often within less than two years), with disastrous implications for the organisation’s ability to build up and retain valuable tacit knowledge.

I could compile a long list of advantages of the NeurodiVenture operating model supported by 7 years (and counting) of operating experience (following 10 years of lessons with various other operating models), but many of these advantages are simply corollaries of the cognitive limits highlighted by Dunbar’s research, which by the way are intuitively understood and adhered to by “uncivilised” societies.

Dunbar’s numbers (5 intimate friends, 15 close friends, 50 general friends and 150 acquaintances) focus on the numbers of relationships that an individual can maintain. From this we can deduce that 50 people is an upper limit for a good company. In fact if the members of a company want to maintain strong trusted relationships with the external world, then it is a good idea for every member to maintain some of the 50 general friend relationships with people in other companies.

Let’s assume on average every member has 10 general friend relationships with people in other companies. Then collectively a good company can maintain a very impressive number of strong trusted relationships with other good companies. For example a company of 40 people would have a mind-boggling capacity of up to 40 * 10 = 400 general friend relationships with other companies. Of course the number is lower if people have many shared general friends, but the actual number will still be quite impressive. The important observation here is that we are talking about genuine and trust based relationships between people and not about superficial and untrusted transactional interactions. Collaborating in good company, even across the organisational boundary, is genuinely enjoyable!

A company of up to 50 people contains up to (50 x 49)/2 = 1225 general friend relationships. The possible collaboration patterns within the company are correspondingly complex. In order for individuals to collaborate effectively and in order to effectively coordinate activities across the organisation it makes sense for emergent groups of regular (daily) collaborators to be given recognisable labels – the result is a structure of teams.

A theoretical debate over whether teams should be allowed to overlap completely misses the point. What matters is that high performing collaborative teams tend to have 7 +/- 2 members. A team’s boundary is defined by the existing team members. A new team member joining is an obvious event, prompted by a new need for daily collaboration.

Good companies


Within a good company (smaller than 50 people) and especially within a team, everyone is acutely aware of the competencies of all the other members. Within traditional teams this knowledge about the distribution of available competencies tends to be tacit – locked up in peoples’ heads, it is not available in explicit form. In a NeurodiVenture  all members expose (write down and share) these so-called individual competency networks for the benefit of everyone within the company.

The result is an immensely valuable index of competencies consisting of up to 50 unique perspectives on the company. These perspectives are not merged into some absurd attempt to create a unique source of truth. All perspectives are considered equally valid. Collectively their presence allows the company to rapidly respond intelligently and with courage to all kinds of external events, by drawing on collective intelligence in a very literal sense.

To appreciate the significance, let’s assume that on average for each person in a company of 50 there are 10 to 20 externally or internally triggered categories of events (these events can be thought of as use cases) associated with a demand that relate to the person’s core competencies, and perhaps there are another 10 to 20 events that the person is also well equipped to deal with (beyond the core competencies). This leads to a collective set of 50 x 20 to 50 x 40 = 1,000 to 2,000 competency self assessments, and to a multitude of perspectives from others on a subset of these declared competencies. Having all this information available in explicit form within a company is an extremely valuable tool.

But of course hardly anyone in a traditional organisation with hierarchical power structures would openly share their individual competency network including their perspectives on the core competencies of other members of the organisation. Anyone who thinks about this obvious observation for a couple of minutes has to conclude that traditional organisations represent a form of collective stupidity – the result of inherent lack of mutual trust due to in-group competition.

Individual competency networks are one of the three ingredients of the collaborative (not secret) sauce of good company. The other two essential ingredients that define the NeurodiVenture model (follow the link for details):

  • The 8 trust-reinforcing organisational principles and rituals
  • The more generic tailored 8 pro-social core design principles

The NeurodiVenture model is the result of incremental evolution. The 8 trust-reinforcing principles and rituals are not unique to our approach and have proven their worth in various contexts. At S23M we started with these 8 complementary/orthogonal principles and rituals as an initial minimal viable operating model. Then several painful lessons prompted us to add the prosocial principles.


Our established undocumented practices meant that we  already had implementations for 7 of the 8 prosocial principles identified by Elinor Ostrom, Michael Cox and David Sloan Wilson, but we were missing the 8th principle “Graduated responses to transgressions”. As as our small team grew beyond four people, the idea of explicit individual competency networks took shape and has since become more important and relevant with every new team member.

Learning companies

In a good company coordination and organisational learning happens via a simple advice process (one of our 8 trust-reinforcing rituals), without any need for social power structures. Before making a major decision that affects others in the organisation:

  1. A person has to seek advice from at least one trusted colleague with potentially relevant or complementary knowledge or expertise.
  2. Giving advice is optional. It is okay to admit lack of expertise. This enables the requestor to proceed on the basis of the available evidence.
  3. Following advice is optional. The requestor may ignore advice if she/he believes that all things considered there is a better approach or solution. Not receiving advice in a timely manner is deemed equivalent to no relevant advice being available within the organisation. This allows everyone to balance available wisdom with first hand learning and risk taking.
  4. The 8 prosocial design principles provide guidance for dealing with people who regularly ignore relevant advice (or consistently refuse to seek or give advice) and therefore regularly cause downstream problems for others as a result. Such situations are obvious for all involved. A persistent breakdown of collaboration either results in a significant change in behaviour once the downstream problems are recognised, or in the non-cooperative person leaving the organisation.

According to Frederic Laloux an advice process with the above characteristics is the one noteworthy commonality across all of the non-hierarchical organisations that he has researched. I can confirm that the advice process is an essential corner stone within the NeurodiVenture model.

The outlined NeurodiVenture model is a minimalistic implementation of a non-hierarchical organisation.


Beyond eliminating formal hierarchical structures the NeurodiVenture model also removes all incentives for the emergence of informal “power-over” structures via transparency of all individual competency networks for the benefit of everyone within the company. This is perhaps the most radical idea within the NeurodiVenture model.

Transparency of individual competency networks enables meta knowledge (who has which knowledge and who entrusts whom with questions or needs in relation to specific domains of knowledge) to flow freely within an organisation.

The conceptualisation of meta knowledge flows via individual competency networks assists the coordination of activities via the advice process outlined above and via regular Open Space workshops, and it acts as an effective dampener on the informal hierarchies that can easily come to plague hierarchical and “non-hierarchical” organisations.


Informal hierarchies tend to corrode trust and collaboration, often with particularly toxic effects for neurodivergent people. This is the one notable difference to the forms of self-organisation advocated by Frederic Laloux, which don’t mandate transparency of meta knowledge flows, and which rely on “natural” hierarchies.

Similarly the additional rituals and “circle” structures that are prescribed by recipes for operating non-hierarchical organisations such as sociocracy or holocracy are not needed in the NeurodiVenture model. Organically evolving small – and often overlapping – teams are created and dissolved as needed to achieve specific goals.

The NeurodiVenture model continuously strives to minimise spurious cultural complexity. It has evolved incrementally based on concrete needs, only relying on the timeless wisdom that has been captured in the prosocial principles that predate “civilised” societies.

Further implications of human cognitive limits

  1. People can understand the nuanced dynamics of teams of up to 10 people.
  2. Groups of more than 10 people must be organised in teams of teams for everyone to retain an understanding of the collaboration dynamics.
  3. An individual competency network that references more than 50 people leads to reductions in trusted collaboration and hence a reduction in organisational adaptiveness.
  4. Via peer-to-peer recommendations the potential for trusted collaboration can be expanded to 150 people, however at any point in time an individual can only collaborate effectively with up to 50 people.
  5. Collaborations involving more than 50 people produce emergent results and attempts of alignment are of limited effectiveness.
  6. All forms of social status are legacy technologies that create dangerous illusions of authority or understanding. Decisions made by an individual on behalf of more than 50 people are based on ignorance about the ways in which people will be affected.

Notes on cultural evolution and transformation


When applying multi-level group selection theory of evolution (MLS) to human societies that are driven as much by culture (social norms) as by genetic biological programmes, the most relevant subjects of evolution are social groups at various levels of scale.

Learning how to create collaborative environments for small “human scale” groups (good companies) creates a collaborative edge over other companies as no effort is wasted on in-group competition. This in turn significantly reduces the need to spend time on “winning” direct competitions with other companies. What happens instead is that other companies are increasingly intrigued by the company’s capability.

Education is essential. When beliefs that represent evidence based facts are propagated via a critical self-reflective process of education that is at least one order of magnitude slower than the process of social transmission (imitation/copying without any deeper understanding),  recipients – to a certain degree – are immunised against influence from those with opinions that contradict evidence based understanding.

The journey of transformation of an organisation opens up three broad scenarios:

  1. Only a few individuals within the organisation recognise the full human potential for collaboration. These individuals will leave the toxic organisation and can be supported in forming or joining a good company with a NeurodiVenture compatible operating model. This is a form of palliative care for toxic organisations that allows organisations to die whilst providing an exit strategy for the inhabitants.
  2. A sizeable subgroup of individuals within the organisation recognises the full human potential for collaboration, but the majority of people within the organisation are not yet ready to shift perspectives. Such a subgroup can for example emerge / self-organise as a result of running regular Open Space workshops within the organisation. The subgroup may choose to separate from the organisation and then collaborate with the organisation (and potentially other organisations) from the outside. This is a form of transformational support that allows organisations to incrementally break up into smaller and healthier collaborative parts.
  3. The majority of individuals within the organisation recognise the full human potential for collaboration. Such a result may also emerge / self-organise as a result of running regular Open Space workshops within the organisation. This is a form of phase transition that allows organisations to rapidly shift to a much healthier collaborative operating model.

Note that in all scenarios above the organisational units and their configuration are the subjects of evolution.

Individuals can choose to remain in the original organisational structure / operating model as long as they wish, but as more and more of their colleagues vote with their feet and signal a preference for a collaborative cultural environment, more and more will find the courage to leave what at the beginning may still have been perceived as a somewhat toxic but at least as a familiar and therefore superficially “safe” environment (due to irrational fear of the unknown, i.e. “things are bad but they could be worse”).

The future is neurodiversity friendly 🙂

2 thoughts on “Organising for neurodivergent collaboration

Leave a Reply