Engineer or Gardener? David Hurst on AI, the Ecocycle and the Rigidity Trap

Mark Blackwell, Arkaro

Senior leaders in mature businesses may know something is wrong. Their organisation is efficient, profitable and increasingly unable to move. What they lack is a language for what to do about it.

David Hurst has spent fifty years developing that language. In this episode of Arkaro Insights, the author of The New Ecology of Leadership and Crisis & Renewal joins Mark Blackwell to revisit his ecocycle framework in the age of AI. The result is a rare kind of conversation: theoretically deep, practically usable and honest about what generative AI can and cannot do for a mature organisation.

Executive summary

The ecocycle is a two-loop model of organisational life. A front loop of birth, growth, maturity and conservation, followed by a back loop of crisis, release and renewal.
Large language model AI belongs on the front loop, where patterns are stable and calculable. Natural human intelligence belongs on the back loop, where the map runs out and only a compass will do.
Many mature organisations sit in what Hurst calls the rigidity trap at the top of the front loop, brilliant at yesterday’s logic and vulnerable to today’s change.
Senior leaders must toggle between two postures: the engineer who optimises a known system and the gardener who creates conditions for an unknown future to emerge.
Odysseus in the Straits of Messina is the picture of adaptive space, steering between the rock of stability and the whirlpool of change, where you get torn apart.
The 4:30am steel delivery ride-along is a concrete instance of ecological rationality, a workable technique for any senior leader who wants to see what head office cannot.

Why the ecocycle matters now

Many leader see the same underlying problem. Their business was designed for one world, and the world has changed. Geopolitics, regulation, technology, climate, all move faster than the organisation can respond. The instinct is to look for a new operating model, a new tool, a new strategy document. Hurst suggests something different: begin by understanding the shape of the trajectory your organisation is on.

The ecocycle came out of his own experience of running a Canadian industrial distributor through the wreckage of a 1979 leveraged buyout, and out of his subsequent reading of the Canadian ecologist Buzz Holling on the adaptive cycle. Holling had studied forests, estuaries and other ecosystems, and had shown that they all follow a two-loop trajectory: a long front loop of birth, growth and maturation, and a shorter, faster back loop of release, reorganisation and renewal.

Applied to organisations, the front loop is the story of Western business as most of our audience knows it. Entrepreneurship, growth, scaling, structure, optimisation. The back loop is the story of what happens when that structure runs into a world it was not designed for, and either finds a way to renew itself or does not. It is, in Hurst’s phrase, a Daoist framework of two processes, each of which contains within it the seeds of the other.

Nothing lasts unless it is incessantly renewed – David Hurst

The two loops: front and back

The front loop begins with a small community of practice and trust, held together by shared passion. It is a place of improvisation, of what Henry Mintzberg calls bricolage: make a little, sell a little, make a little more. Steve Blank calls this a temporary organisation searching for a repeatable, scalable business model. The vast majority of ventures fail here. Those that succeed graduate into deliberate strategy, adopt Michael Porter’s value chain, scale up, and enter conservation.

At the top of the front loop sits the mature organisation. It is efficient, profitable, structured. It is also designed not to change, because its whole reason for existing is to preserve success. That is a rational design when tomorrow looks like today. It becomes brittle the moment the world begins to move.

The back loop begins with crisis. Hurst borrows his metaphors from the Hebrew Testament: wilderness, mountain, promised land. Wilderness is a place of confusion where it is easy to get lost and where many organisations spend years. Hurst notes wryly that General Motors spent forty years wandering in the wilderness before it went bankrupt in 2009. To get out of the wilderness, leaders must climb the mountain, which is the process of adaptive leadership: finding a view of what could come next, and building a shared sense of purpose that lets people act. Mission and purpose, in his experience, emerge from struggle and setback, not from external supply by a leader.

The critical point, and one Hurst returns to throughout the conversation, is that these are not sequential stages of a career. A working manager needs both loops in mind at the same time, and needs to know which loop the specific situation in front of them belongs to.

Where AI fits, and where it fails: the map and the compass

The single cleanest image from the conversation is the contrast between the map and the compass. The map is the front loop. It is the domain of stable patterns, known routes, calculable probabilities. Large language model AI is fabulously good at maps. It has been trained on a vast collection of yesterdays and can predict, with remarkable fidelity, what a stable pattern is likely to look like tomorrow.

The compass is the back loop. It is what you use when the map runs out, when the terrain is unfamiliar, when the pattern itself is changing. This is the domain of what Hurst calls ecological rationality, the human capacity to read a specific situation and find the affordances (action possibilities) that it offers. The philosopher J.J. Gibson coined the term for the way an animal reads its environment. Hurst applies it to the way an experienced manager reads a particular customer, a particular team, a particular moment.

Hurst is careful about his terminology. He talks about LLM AI, not AI in general. Other forms of AI, notably the active inference work associated with Karl Friston and Andy Clark, are attempting to build systems that work more like the human mind. That work has not yet gained traction, but it is where any serious progress towards genuine artificial general intelligence is likely to come from. Today’s generative tools are not headed there.

It’s forced us to reconsider what it means to be human. And it’s quite obvious that being human is much more than LLM AI can do – David Hurst

The rigidity trap: why mature organisations get stuck

The senior leader in the audience of this podcast is most likely running a business in conservation. Efficient, profitable, structured. Head office is a well-appointed place. When you were an entrepreneur, if you wanted resources you had to find a customer, because customers were the only people who had them. Now, if you want resources, you have to work through internal courtiers who guard access to a prince who allocates them. Your back is to the customer. Your face is to the boss.

That is the rigidity trap. It is not a failure of individual leaders. It is a normal, predictable trajectory that follows from success, exactly the same trajectory that mature forests follow when they become monocultures and can no longer regenerate. Hurst is emphatic that provided tomorrow is just like today, this is a fabulous place to be. And you will do fabulously well, for a while.

The signals that the world has stopped being stable are usually clear enough, if leaders look for them. A change of a customer, a change of a regulator, a change in the climate, a change in the political weather. What tends to be missing is not the signal, but a language for what to do about it. That is what the ecocycle provides.

Scylla and Charybdis: the adaptive space

For the manager caught between preserving success and preparing for renewal, Hurst reaches for Homer. Odysseus, sailing the Straits of Messina, has to navigate between two dangers: Scylla, the rock of stability where everyone gets eaten in restructurings and force reductions, and Charybdis, the whirlpool of change where organisations get torn apart. Steer too close to Scylla and the business dies of consolidation. Steer too close to Charybdis and it dies of chaos. The task is to hold the boat in the adaptive space between them.

This is a picture, not a solution. But it is a picture that senior leaders in mature businesses recognise instantly. It captures the daily judgement call about how much continuity to preserve and how much disruption to encourage, and it does so without collapsing into either the machine metaphor of pure engineering or the utopia metaphor of pure agility. The reference lands harder because of Christopher Nolan’s just-released Odyssey adaptation, which puts the same picture in front of a much wider audience.

From engineer to gardener

The core operating instruction that comes out of the framework is the toggle between two postures. The engineer designs and builds. Given a known problem in a known context, the engineer’s approach is the right one, and LLM AI is a powerful addition to the engineer’s toolkit. The gardener cultivates the conditions for growth without controlling the outcome. Given an unknown problem in an unstable context, the gardener’s approach is the right one, and this is a place where LLM AI cannot help.

Neither posture is better than the other. Neither is more sophisticated. Both are legitimate and both are necessary. The failure mode of mainstream Anglo-American management over the last sixty years, in Hurst’s view, has been to elevate the engineer and disregard the gardener. The result is organisations that are excellent at optimisation and hopeless at renewal.

The manager’s real job is to know, situation by situation, which posture the moment demands and to switch smoothly between them. Sometimes the toggle is slow, at the level of a year or a strategic cycle. Sometimes it is fast, at the level of a single meeting where you might discuss a structured operational issue in engineer mode and then move to an emergent question in gardener mode. Dr Iain McGilchrist‘s work on the divided brain, which Hurst has drawn on for decades, gives a neurological foundation for why humans are equipped to do this in the first place. Hurst himself has been living with the framework since 1984, when a Harvard Business Review article he wrote about the leveraged buyout turnaround, “Of Boxes, Bubbles and Effective Management”, laid out the same yin-yang tension in different vocabulary.

Ecological rationality in practice: the steel delivery story

The most concrete practitioner story of the conversation comes from Hurst’s time running a Canadian steel service centre. Once in a while, he would set his alarm for 4:30am, meet a driver at the yard, and go out on a steel delivery to a customer at six in the morning. Standing in the customer’s back yard at that hour, he could see what no head office report could ever show him. Whose competitors were queued up delivering to the same customer, and what they were delivering. Whether the customer’s production line was full or empty. Whether the workforce looked as if they might be about to strike. Whether the yard was well-run or in disarray. What was arriving, what was leaving, what was piling up.

This is ecological rationality made concrete. The affordances of the situation, the action possibilities, become visible only from a specific place at a specific time. No abstract dashboard can generate them, because they do not exist as data in any system. They exist in the world, and they only become available to a leader who deliberately puts themselves in the world where they live.

This is the fine-grained awareness of what’s actually happening, which allows you to adjust your top-down picture, which allows you to predict the world. – David Hurst

Every senior leader has some version of this available. A morning at a farm-gate. A shift on a production line. An afternoon in a distributor’s warehouse. A day in a formulator’s application lab. The specific place matters less than the discipline of going to it, on foot, at the hour when the work is actually happening.

What this means for leaders on Monday morning

Asked what senior leaders in mature businesses should do next week with their teams, Hurst declined to give a general answer. Any generic recommendation, he pointed out, is going to be true but useless. Cultivate psychological safety. The opposite is clearly nonsense, so everyone will nod, and nothing will change. The advice that actually helps is particular to this organisation, with these people, right here, right now.

The takeaway, then, is not a checklist. It is a discipline. The discipline is sense-making. Put effort into diagnosing where on the ecocycle your organisation actually sits, and where different parts of it sit, because a large business will have some units clearly on the front loop and others clearly on the back loop. Notice which posture, engineer or gardener, the situation in front of you demands. And when you are about to reach for a piece of context-free advice from a management text or from a generative AI tool, ask whether the advice is a map fit for the terrain, or whether the terrain requires a compass.

That is the work. It is the work that AI cannot do for you. And it is the work that, done well, makes AI genuinely useful in the parts of the business where it earns its place.

Listen to the full conversation

The full episode with David Hurst is available on YouTube (www.youtube.com/@arkaro) and on the Arkaro Insights podcast (arkaroinsights.buzzsprout.com), along with Apple Podcasts, Spotify and all major platforms.

About the speakers

David Hurst

David K. Hurst is a reflective practitioner, speaker, educator and writer on management. He spent twenty-five years as a senior operating executive, including a decade as Executive Vice-President of Federal Industries’ Metals Group in Canada, following the rescue of Hugh Russel Inc. after a leveraged buyout. He is the author of Crisis & Renewal (Harvard Business School Press, 1995) and The New Ecology of Leadership (Columbia Business School Publishing, 2012). He writes regularly at davidkhurst.com and on Medium.

Mark Blackwell

Mark Blackwell is the founder of Arkaro Ltd, a boutique B2B consultancy specialising in change management, innovation, integrated business planning and commercial excellence. Arkaro works with clients in agriculture, food and chemicals through a ‘do it with you’ philosophy that builds lasting capability alongside measurable performance improvement. Mark hosts the Arkaro Insights podcast.

Working through a rigidity trap of your own?

Arkaro’s four-step approach, understand, co-create, enable, sustain, is built for exactly the kind of adaptive work David Hurst describes. If the shape of the conversation resonates with what you are seeing in your own business, we would be glad to explore whether a first conversation might be useful.

www.arkaro.com | mark@arkaro.com

External resources and further reading

By David Hurst

The New Ecology of Leadership: Business Mastery in a Chaotic World (Columbia Business School Publishing, 2012)
Crisis & Renewal: Meeting the Challenge of Organizational Change (Harvard Business School Press, 1995)
Of Boxes, Bubbles and Effective Management (Harvard Business Review, 1984)
davidkhurst.com, David’s blog and complete essay archive
davidkhurst.medium.com, recent essays including Toggling Between Two Worlds

Foundational reading he references

C.S. (Buzz) Holling and Lance Gunderson, Panarchy: Understanding Transformations in Human and Natural Systems
Iain McGilchrist, The Master and his Emissary: The Divided Brain and the Making of the Western World
Richard Nisbett, The Geography of Thought: How Asians and Westerners Think Differently
Gerd Gigerenzer, on fast and frugal heuristics and ecological rationality
Gary Klein, Sources of Power: How People Make Decisions (Naturalistic Decision Making)
Karl Friston and Andy Clark, on active inference and predictive processing as alternative foundations for AI

Subscribe to Arkaro Insights

For long-form conversations with the practitioners and researchers helping senior leaders in agriculture, food and chemicals think more clearly about change, complexity, AI and commercial performance.