Border Collies for a Stampede of Black Elephants
Steering the Unsteerable: How Agentic AI and Living Systems Intelligence Can Guide Us Through Collapse and Regeneration
I. The Collapse of Control-Based Systems
For centuries, our dominant economic, corporate, and political structures have been built for one purpose: control. These systems were engineered to manage complexity through hierarchy, authority, and rigid chains of command. They function like a herd of sheep, centralized and obedient, waiting for direction from a shepherd.
But the world we now inhabit is no longer predictable. The rise of exponential technologies, climate breakdown, financial instability, and geopolitical shocks has created a reality that these old systems were never designed to handle. They were built for control, but they now exist in an environment that defies control.
Thomas Friedman describes this moment not as a stampede of black swans — rare, unpredictable shocks — but as a stampede of black elephants: crises that we knew were coming, yet largely ignored until they became unavoidable. These are not unforeseen disruptions but rather predictable catastrophes, looming in plain sight, yet dismissed or deprioritized until they escalate beyond control. It is this accumulation of neglected systemic failures — climate breakdown, financial instability, technological upheaval — that is now overwhelming institutions built for a far simpler world.
Organizations today — governments, corporations, institutions — are running headfirst into this storm, but they lack the capacity to adapt. Their architectures are brittle, their decision-making is slow, and their logic is outdated.
This is why Horizon One (H1) — the dominant system of today — is self-terminating. The more complexity accelerates, the more these organizations struggle to keep pace. The pressures they face — economic limits, ecological collapse, technological displacement — are not temporary disruptions. They are symptoms of a system that has outlived its own logic.
The mistake many people make is assuming that these systems can be reformed from within. But they cannot. They are too deeply embedded in the logic of control. You cannot turn a centralized, hierarchical institution into an emergent, self-organizing system simply by tweaking it at the edges, nor can it be changed from the inside by acting on the nodes that belong to and reinforce its existing structure.
So, what is the alternative? If we cannot change the sheep, we must focus on training the border collies — the guiding intelligence that can steer these organizations through the transition, without collapsing with them.
II. Steering Complexity: The Role of Border Collies
The old world is breaking down, but the new world is not yet here. We are in Horizon Two (H2), the turbulent middle ground where disruption and emergence coexist.
Horizon Two is a battleground. On one side are forces trying to extend the lifespan of Horizon One, propping up old structures through sustainability efforts that ultimately cannot prevent their demise. On the other side are forces trying to birth Horizon Three (H3), the emergent system that will replace Horizon One.
The challenge is that we do not yet know what Horizon Three looks like. It is in the adjacent possible — not yet fully formed, but rapidly taking shape. The role of those who want to guide this transition is not to impose a rigid blueprint, but to create the conditions for a regenerative system to emerge.
And this is where the border collies come in.
The border collies in this analogy are not individuals. They are systems of intelligence — decentralized, adaptive, and capable of navigating complexity without relying on control. Historically, this intelligence has existed in informal networks, cooperative models, and distributed governance structures.
But today, we have the ability to scale and amplify this intelligence through AI — specifically, through agentic AI.
Agentic AI as the Steering Mechanism
Agentic AI is not traditional AI. It does not merely process data or optimize for efficiency. Instead, it operates within complexity itself.
What makes agentic AI different?
- It does not control; it enables emergence. Traditional AI models are built for prediction and control, but agentic AI is designed to work within the unpredictability of complex systems.
- It creates decentralized coordination. Just as a murmuration of starlings moves in perfect synchrony without a leader, agentic AI enables self-organization without centralized command.
- It acts as an adaptive guide. It does not impose rigid structures, but it nudges, aligns, and steers systems through dynamic feedback loops — like a border collie navigating a herd without force.
Through agentic AI, organizations that were once centralized and rigid begin to take on the properties of living systems. They start adapting. They start self-organizing. They begin to function like ecosystems.
But this raises a critical question: How is agentic AI being coded? What intelligence is it learning from?
If we are not careful, AI will be trained on the same mechanistic logic that created Horizon One in the first place. Instead, we must train it on the intelligence of life itself — so that it can become adaptive, resilient, and thriving, just as life is, by understanding the systemic patterns and principles that enable living systems to regenerate and evolve.
III. AI and the Code of Living Systems
The greatest intelligence on this planet is not artificial — it is natural. It is the intelligence of living systems, which have adapted, regenerated, and thrived through five mass extinctions.
If we are to steer Horizon One through its collapse without falling into total disorder, we must encode AI not with the extractive logic of the past, but with the adaptive intelligence of ecosystems.
Digital Twins of Living Systems
Imagine if AI could learn not from markets, but from forests. If it could study not just supply chains, but mycelial networks. If it could replicate not corporate hierarchies, but ecological resilience.
This is now possible through digital twins of ecosystems — AI models that do not just simulate financial or industrial systems, but actually replicate the adaptability, resiliency, and thriving potential of healthy ecosystems in balance.
These digital twins are not abstract. They are real-time, dynamic models that:
- Map and understand the self-organizing principles of nature.
- Apply those principles to organizations, cities, and economic systems.
- Train AI in the logic of life rather than the logic of extraction.
This is not just about technology. It is about a fundamental shift in how intelligence itself is defined. Instead of treating AI as a tool to reinforce existing institutions, we must see it as a means to unlock the intelligence of living systems and apply it to human structures.
The Future of Organizations: From Herds to Ecosystems
As AI begins to learn from living systems, the organizations it steers also begin to transform.
- Corporations cease to be rigid hierarchies and start functioning as adaptive networks.
- Cities stop being mechanistic grids and start behaving like ecosystems.
- Governance shifts from centralized control to decentralized, self-organizing frameworks.
In other words, the herd of sheep begins to disappear. The more organizations learn to use agentic AI that is trained in the principles of regenerative design, the less they rely on top-down control.
This is the real transition to Horizon Three — not just a shift in energy or economics, but a shift in the fundamental architecture of intelligence itself.
Conclusion: The AI That Steers Us Through
The world is moving too fast for the systems of control to keep up.
Horizon One is self-terminating.
Horizon Three is emerging.
Horizon Two is the space of transition.
We do not need to control this transition. We need to guide it.
The border collies of this shift are not governments, corporations, or even individuals. They are agentic AI systems designed not to impose control, but to enable emergence.
If we succeed in coding AI to learn from living systems, we will create the conditions for organizations to evolve — moving from hierarchies to ecosystems, from control to adaptability, from extraction to regeneration.
We stand at a crossroads. If we fail, we may not only accelerate the collapse but push it beyond a point of no return, reinforcing the logic of Horizon One until it implodes under its own weight — especially given our proximity to the singularity moment, where the accelerating pace of technological and systemic change could either drive regeneration or irreversibly deepen the crisis.
The future will not be controlled. But it can be steered.
The question is:
Who is training the border collies? And what intelligence are they learning from?
Watch this incredible video of border collies at work — precision, agility, and intelligence in motion. A masterclass in decentralized coordination, showing how complex systems can be guided without force.
