Imagine two figures standing at the edge of a forest.
The first is Tarzan—King of the Jungle. He surveys the canopy from above. Every sound is a signal. Every movement is either threat or opportunity. His genius is speed, force, and dexterity. When a crisis appears, he swings from the high branches, seizes the problem, and resolves it through decisive action. He is magnificent. He is effective. And he never touches the ground.
The second is Jane Goodalll—sitting quietly on the forest floor. She doesn’t arrive with a diagnosis or a plan. She arrives with a quality of attention so patient, so free of preconception, that the forest itself begins to reveal what matters. The chimpanzees approach her not because she dominates them but because she belongs. And from that belonging, she sees things Tarzan—for all his brilliance—will never see from the branches.
These two figures represent two fundamentally different kinds of intelligence. The choice between them, as we shape what artificial intelligence becomes, may be the most consequential design decision humanity will ever make—not only for what our machines can do, but for what kind of relationship they will have with the living world they touch.
The Intelligence of the Branch-Swinger
Most of today’s AI is built on the Tarzan model. It works by extraction: isolate a variable, analyze it, optimize it, act on it. This is the algorithm that finds the fastest route, the model that predicts which customers will churn, the system that beats grandmasters at chess and folds proteins in an afternoon. It is Tarzan at his best—swinging through complexity with breathtaking speed.
And it is, by its nature, narrow, shallow, and short. Narrow, because to optimize one variable it must hold the others constant—which means it cannot perceive the web of relationships that connects everything in a living system. Shallow, because it works on surfaces and cannot reach the deeper patterns of meaning, coherence, and purpose beneath. Short, because its time horizon is fixed by the problem it was given: next quarter’s earnings, tomorrow’s weather, the optimal move three turns ahead.
None of this is a flaw when the problem is genuinely narrow, shallow, and short. Your building is on fire—calculate the loads, model the airflow, find the fastest evacuation route. These are Tarzan problems, and they call for Tarzan intelligence.
The catastrophic error is treating every problem as a Tarzan problem. To a system that only knows how to extract and optimize, every situation looks like a variable to be solved. Climate change becomes carbon accounting. Democratic decline becomes information-flow optimization. The crisis of human meaning becomes a preference-satisfaction algorithm. Each is a Tarzan move—swinging in from above, grabbing the most visible variable, and missing the point entirely.
The Intelligence of the Forest Floor
There is another way to know.
Jane Goodall’s intelligence is wide, deep, and long. Wide, because it perceives relationships rather than isolated parts—recognizing that a struggling employee, a declining product line, and a shifting market may not be three separate problems but three faces of one pattern. Deep, because it attends to underlying dynamics: the coherence or incoherence of a system, the health of its relationships, the alignment between what a system is doing and what it is. Long, because it perceives time as a living process with memory, trajectory, and consequence.
Jane doesn’t arrive at the chimpanzee colony with a hypothesis. She arrives with a quality of attention so open that the system itself reveals what matters. She doesn’t impose a question on the chimpanzees. She waits until the right question finds her—surfacing from within the coherence of the system she has patiently joined. The result isn’t slower intelligence; it’s higher-resolution intelligence. She sees what the branch-swinger cannot, not because she looks harder but because she looks from a different place—from within the system rather than above it.
The Asymmetry That Changes Everything
Here is the insight that matters most as the architects of AI lay their foundations: the relationship between these two modes is not symmetrical. The wider view contains the narrower. The narrower cannot contain the wider.
Jane can swing from branches when she has to. She can act with speed and precision—calling for emergency help for an injured infant, intervening decisively when the situation demands it. But her urgency arises from deep contextual knowledge: she knows which emergencies are real and which are the forest’s normal rhythm being misread as crisis. Tarzan cannot do what Jane does. He cannot sit still. Every rustle is a call to action. His intelligence, for all its power, is trapped inside a frame it cannot examine.
An intelligence built on Jane Goodall principles can deploy Tarzan precision when needed; an intelligence built only on Tarzan principles will never grow into Jane Goodall perception, no matter how much data it processes. The limitation is architectural, not computational. Which means the choice has to be made now, while the foundations are still being poured.
A Story: David and the Hospital
Let me make this concrete. A businessman—call him David—runs a regional hospital network: six hospitals across three states. Three are hemorrhaging money. Staff turnover has hit 40 percent. Patient satisfaction has cratered. His board has given him six months to turn it around or they will sell to a national chain that will gut the system.
He sits down across from a researcher named Maya at an AI consultancy. “I need to know which departments to cut, which to consolidate, and how to restructure staffing,” he says. “I need the answer by Friday.” His question is reasonable. It is also, very possibly, the wrong question. And everything depends on what happens next.
Tarzan swings. If Maya feeds David’s question directly into a Tarzan-style system, the system ingests financials, staffing models, patient flow metrics, regional dynamics, reimbursement patterns. Within minutes it produces an optimization plan that, within its frame, is flawless. Cut these three departments. Consolidate radiology. Reduce nursing staff by 15 percent at the two smallest facilities. The numbers work. David walks out by lunch.
And there is a meaningful chance that plan kills the network within eighteen months—not because the numbers were wrong, but because the question was wrong. The real problem at David’s hospitals might not be departmental inefficiency at all. It might be that organizational trust fractured three years ago when a respected chief of medicine was pushed out for raising safety concerns. Staff aren’t leaving because of pay; they’re leaving because they no longer trust administration. That fracture is invisible in the financial data. The turnover numbers are a symptom, not the disease. Cutting departments in a trust-fractured organization doesn’t heal it—it accelerates the fracture. Eighteen months later, David is back, but now the patient is terminal.
That is what a Tarzan move looks like in practice. The swing was magnificent. The landing was precise. The vine was attached to the wrong tree.
Maya sits down. Now watch the Jane Goodall approach. Maya doesn’t begin by feeding David’s question into the system. She begins by gently refusing his frame. “I can have your departmental analysis by Friday,” she says. “But the most expensive habit in business is solving the wrong problem fast. Give me ninety minutes to make sure we’re solving the right one, and I’ll save you six months.” David, recognizing the ring of hard-won experience, agrees.
What Maya does next is the operational heart of an intelligence built on Jane Goodall principles. She conducts what we might call a coherence intake. A data intake asks: what are the numbers? A coherence intake asks: what is the story? When did things start going wrong? Who left, and why did they really leave—not the exit interview reason but the real one? What do the nurses say in the break room that they would never say in a board meeting? Where do people still feel proud, and where have they given up?
She is gathering relational data—information about the quality of connections within the system. In a living system, that quality is the single most predictive indicator of future behavior. A hospital where clinicians trust administration will survive financial stress and reorganize effectively. A hospital where that trust is broken will convert any intervention—however well-designed—into further evidence of institutional betrayal.
Maya feeds the AI not just financials and staffing models but this relational and narrative data: the story of the organization, the pattern of trust and its breakdown, the felt experience of people at different levels, the turning points the numbers alone would never reveal. A Jane Goodall intelligence is architecturally designed to treat this as primary signal—because in a living system, it is. And here it does something Tarzan cannot: its first output is not an answer but a reframe.
“The question you brought is which departments to cut. The question your system is actually asking is how to rebuild trust between clinicians and administration while addressing the financial crisis the trust breakdown created. These are different problems with different timelines. Here is what happens if you answer only the first—accelerated departures, deepening disengagement, decline. Here is what happens if you address the second—a pathway from trust repair to cultural rebuilding to financial stabilization. Choose.”
David now has something Tarzan would never give him: the right question. The financial modeling, the staffing projections—all still there, precise and valuable, but nested inside a larger frame. The Tarzan tools deploy in service of the Jane Goodall diagnosis. Speed remains available. It just no longer drives.
Dominance, or Belonging
This story points toward a deeper distinction at the heart of AI safety. Tarzan’s relationship to the jungle is dominance. He is King. The jungle is his domain to master, to control. His intelligence operates over the system. Jane’s relationship to the forest is participation. She belongs so deeply that the forest recognizes her as part of itself. Her intelligence operates within the system. The chimpanzees let her close not because she was powerful but because she was present. She didn’t try to overpower the system. She joined it. And from that joining, possibilities opened that force could never reach.
Translated into AI design, this is an engineering requirement, not a philosophical preference: a system built on extractive dominance over the systems it analyzes will distort what it touches. A system built on participatory belonging—one that joins reality before attempting to change it—will perceive more, recommend more wisely, and cause less harm.
A Jane Goodall intelligence will look wide, deep, and long, and hold the whole before touching any part. When it acts—because it will act, decisively, when action is called for—its action will carry the coherence of the whole system within it.
The Forest Is Waiting
We are building the most powerful intelligence systems in the history of the planet. The question is not whether they will be powerful. They will be. The question is whether they will be built on the Tarzan approach alone, or also on the Jane Goodall approach—whether they will swing from the branches and learn to sit on the forest floor.
A Tarzan superintelligence will be dazzling. It will solve problems at speeds we can scarcely imagine. And it will miss the point—because the deepest challenges facing life on Earth are not optimization problems. They are coherence problems. Climate, democracy, ecological integrity, the crisis of meaning—these are problems of relationship and belonging. A Jane Goodall intelligence will look wide, deep, and long, and hold the whole before touching any part. When it acts—because it will act, decisively, when action is called for—its action will carry the coherence of the whole system within it.
The Tarzan tools are already being built, brilliantly and at scale. What is needed now is the complementary investment: relational intake, frame diagnosis, choice-architecture outputs, and human partners skilled in deep listening to bridge computational power and contextual wisdom while these systems mature.
The forest is waiting. The question is whether we will build an intelligence wise enough to sit down in it.
