Musing on DAOs, Cybernetic control systems and non-hierarchical structures
Could smart tech and AI simplify the coordination problems that have held back non-hierarchical organisational models?
We are already seeing the decline of hierarchy as an organising principle for leadership in large organisations, and we have increasing evidence, such as this from Leandro Herrero, that positional authority is far less powerful than networked influence as a way to get things done. But beyond leadership, hierarchical models are baked into the operating systems of larger and older organisations, despite the mounting negative effects in terms of speed, agility and responsiveness to change, which means it is a hard problem to solve.
Perhaps AI and smart tech can help us develop viable alternative control systems for organisations, and in so doing, possibly revive some older ideas and organisational models. Among those who seek better, more agile and future-fit alternatives to organisational hierarchy, two rather idealistic models have been popular for a long time: flat organisations that co-ordinate work without the need for costly layers of middle management; and cybernetic systems that seek to use real-time information flows to optimise and guide resource allocation. And, more recently, there is also the idea of the Distributed Autonomous Organisation (DAO), which emerged from the crypto space and has involved a lot of new thinking and experiments relating to distributed governance and ownership systems.
But what is and is not hierarchy?
One of the challenges for all these models is that hierarchy underpins multiple functions in organisational life, which makes it hard to pin down precisely:
it denotes relative status and importance, a function that seems to occur naturally among primates;
it serves as a cascading communications structure;
it serves as a system to co-ordinate work outputs; and,
it anchors the governance system.
So, when people talk about or seek alternatives to hierarchy, it is important to consider which of these roles they have in mind.
Geoff Marlow recently wrote about the loose way some commentators use the term 'flat organisation', referring to the classic text The Tyranny of Structurelessness to suggest that the absence of formal hierarchy often leads to the rise of informal structures that promote unchallenged hegemony and invisible power. He cautioned against conflating hierarchy with decision making, and suggested successful non-hierarchichal organisations tend to be those that avoid bureaucratic decision making structures, and instead focus on creating the cultural conditions for success.
The unnecessary confusion created by describing organisations as flat but not really flat can be easily avoided by focusing instead on creating a future-fit culture.
Then whether or not there are middle managers doesn’t matter — so long as you create conditions where sense making, decision making, and action taking are tightly coupled, rapidly and repeatedly iterated, deeply embedded and widely distributed throughout the organisation.
Even people keen on flat organisations in modern sectors recognise that sometimes we need to protect ourselves from the downsides of flat organisational structures, as this recent article for product managers by Amy Mitchell advises:
Some of the issues that product managers face in a flat organization are:
Unclear scope and expectations of the role of the product manager
Internal resource constraints from conflicting priorities
Communication overload from managing multiple stakeholders
Being a generalist while delivering product commitments
Cybernetic Systems and DAOs
Elsewhere, Dan Davies has been reflecting on the famous “computers with ashtrays” Cybersyn experiment in Chile (most recently here and here) and what we can learn from this about governing and guiding complicated economic systems (or large firms). Since the early days of cybernetic theory popularised by Norbert Wiener, people have dreamed of creating perfect information systems to inform real-time optimisation and decision making, and with recent advances in smart tech and AI, perhaps we can finally achieve it. I think it is possible, but not yet probable, that the application of AI and smart systems could revive or even combine both flat organisations and cybernetic control models in pursuit of leaner, more responsive firms.
We have already seen the rise of the Distributed Autonomous Organisation (DAO) - one of the more useful concepts that emerged from the world of crypto over the past decade - that seek to achieve distributed governance and ownership based on tokens and smart contracts to create automated and autonomous organisations.
Simone and Emanuele at Boundaryless wrote a long post last year considering whether the Haier / DEDA model of platforms supporting an ecosystem of micro-enterprises that collaborate based on mutual agreements might be advanced by using the kind of governance models that DAOs are working on. They seemed to come to the conclusion that DAOs are perhaps better suited to slow-moving protocol or 'industry body'-type groups that sometimes invest in smaller firms or projects that use the protocol or advance the sector they are interested in, rather than substantial firms as we understand them today.
My own scepticism around DAOs stems largely from experience of how human nature plays out in weak tie networks based on transactional relationships, and also the challenges of collective action in non-local voting or coordination systems, plus of course the old 'tragedy of the commons' hole that many well-meaning attempts at collective action fall into.
Cooperation, coordination and governance are different use cases
Decentralisation should be a key design principle for a modern organisational operating systems, but it is best applied to empowering people and groups who share common purpose, not just anons on the internet.
The DEDA model (Digitally Enhanced Directed Autonomy) and Rendanheyi are the most advanced forms of a de-centralised platform organisation today outside of pure tech firms, and they are growing and learning very quickly compared to US or European firms. The idea of algorithmic governance and smart contracts are attractive ideas, and could potentially build on Rendanheyi's idea of mutual agreements and accountability - e.g. P2P service agreements that integrate on the basis of smart contracts. I can also imagine better auditing and financial management, and perhaps also better models of baked-in compliance, rather than just policing guidelines.
But I hope we can leave behind the tokenomics and instead grow our social capital to scale human trust, rather than create so-called trust-free systems with anonymity and 'code is law' purism, which always seem to incentivise bad behaviour. Code is fallible just as law is fallible, and in the case of law, this is a feature not a bug, since human life is messy and norms rarely absolute. We are probably about to learn some of the same lessons with AI as we did with crypto - it should be used to help people do things, not replace them or manage their relationships. It is in fact a social knowledge machine, not something magical and sentient.
If we start small with close-knit teams and use these ideas of peer-to-peer cooperation, cybernetic coordination and networked governance or guiding to create better organisations, with minimum viable bureaucracy and hierarchy to join things together at the highest levels, then perhaps we can finally create workable de-centralised, flat and smart organisations that build on what we have learned so along the way:
flat - or better said, laterally connected, loosely joined sub-organisations working on a common platform are great for peer-to-peer networks of teams doing related work at the lowest level
AI-enhanced cybernetic coordination systems can enhance or replace bureaucratic reporting and water-carrying in the middle
The ideas coming out of distributed governance and DAOs could help with longer-term planning and oversight
As a way to challenge my own thinking and precepts on DAOs, I recently accepted a very kind and gracious offer from Daniel Ospina of RnDAO to join some of his colleagues to discuss these questions, based on our common ground of interest in de-centralisation. Here is the video: