The Sense-Making Gap

This month's newsletter is co-written with Grace Matelich. In this unprecedented time of compressed technology advancement, we are thinking through the rules of engagement, that help maintain our humanity.

Adults alive today are the last generation with an embodied memory of analog sensing. We remember what it's like to get lost with a paper map or written directions, and we long for the strange freedom of not-knowing before search engines. We are the bridge generation: old enough to have learned to read the world before algorithms learned to read it for us, and young enough to shape how that affects the generations to come.

Humans love optimization, and we've always had the tendency to trade capacity for convenience. But optimization implies a destination: a linear trajectory toward better, faster, more. What we actually need from our relationship with technology is something closer to what the body already knows how to do: homeostasis. A dynamic, self-correcting balance. The body doesn't optimize toward a fixed point, but rather conducts an orchestra of continuous negotiation. Pain is information, and hunger is recalibration. The body is always sending signals in service of staying whole. That loop (sense, integrate, adjust) is what's now at risk as technology over-involves itself in what we do with what we perceive.

To understand why, it helps to understand what homeostasis actually is. Walter B. Cannon coined the term at the turn of the twentieth century to describe the body's ability to maintain a stable internal environment despite changing external conditions. Broadly speaking, he shifted our understanding of how the body works: an evolving system under constant dynamics in service of a steady state. Our five senses are participants in this dynamic complexity. They are the starting points for the information highways that serve our physiological, emotional, and analytic systems. Critically, these inputs are learned, and the intensity of any signal can be manipulated by use or disuse. A person who loses their sight compensates by sharpening their other senses. Similarly, a person who increasingly outsources perception to digital systems will, just as predictably, lose sensitivity in the channels those systems don't prioritize. The body responds to the perceptual environment it inhabits. In other words, as we know well: use it or lose it.

The technology proliferating around us does more than change and amplify what we can do, it reshapes how we perceive. Unlike previous tools, AI can mediate our reality upstream, before we've decided what to sense or how to make sense of it. Which means the way we integrate these tools will determine not just our own perceptual capacities, but those of everyone who comes after us, who will inherit our choices as their baseline.

Consider an ER doctor with fifteen years of experience examining a patient. The AI triage system has scored them as low priority: stable vitals, minor complaints, can wait. But something makes her pause. Something in the patient's affect, the way they're breathing, a slight grayness around the mouth. She orders bloodwork immediately. She’s right: the patient is in early septic shock. Though their vitals were stable, they were about to fall off a cliff.

The Attending, twenty-nine years old, trained entirely with AI decision support, asks afterward: "How did you know? The numbers were fine. The system put them in the green zone." She tries to explain. Something about the color and the way the patient held their body. The smell, even. Something she couldn't quite name but registered anyway. She realizes she can't fully articulate it because it’s a mix of intuition and pattern recognition built from thousands of patient interactions, resulting in a kind of knowledge that operates outside language and data.

The critical distinction is that the algorithm and the doctor both sensed the patient. The AI had accurate vital signs. The doctor had visual data, auditory cues, olfactory signals, and tactile information. Both were collecting very real data. But they each made sense of that data differently. The algorithm integrated the numbers into a risk score; the doctor integrated multiple sensory streams into a gestalt impression that told her danger, even though no single data point was alarming.

This distinction, between sensing and sense-making, matters across every domain where artificial intelligence is learning to perceive on our behalf.

If sensing is information collection, sense-making is information integration. We are all sensing all day long, in ways we notice and ways we don't: the flicker of a colleague's expression during a meeting, the weight of silence after a text goes unanswered, the shift in a room's energy when someone's mood changes. But sensing alone doesn't tell us what to do. Sense-making is the interpretive layer that weaves raw perception into meaning so we can decide whether silence is anger or distraction, whether that shift in the boardroom signals danger or boredom.

We sense-make constantly, mostly without noticing. A parent reads their child's "I'm fine" and knows it means the opposite. A negotiator catches a micro-hesitation and adjusts their ask. A driver merges into traffic through some alchemy of speed, distance, and the body language of other cars that no algorithm has fully captured. These aren't mystical abilities. They're the product of accumulated experience integrated below the level of conscious thought pattern recognition so deeply learned it feels like instinct.

Neuroscientists call this capacity interoception: the brain's ability to sense internal states and integrate them with external information. It is the substrate of gut feelings and the biological basis of a bad vibe. Research from the University of Sussex has shown that people with higher interoceptive accuracy make better decisions under uncertainty, simply because they integrate information more effectively. But like any capacity, interoception follows a use-it-or-lose-it logic. When we consistently defer to external systems for interpretation, the internal signal weakens like a muscle beginning to atrophy.

This is where the AI moment introduces a specific risk that previous technological shifts did not. Martin Seligman's foundational research on learned helplessness revealed a pattern that applies far beyond its original context: when organisms repeatedly experience outcomes they cannot influence or in this case, interpretations they didn't generate they stop reaching for their own capacity, even when it remains available. The dog that has learned it cannot escape eventually lies down and accepts it, even when the door is open.

When an algorithm consistently delivers a more confident interpretation than your own body does, you gradually stop trusting your own signal. The capacity doesn't disappear, it simply learns to stay quiet. This is not a failure of intelligence. It is a predictable response to a changed environment. The ER doctor's gift is the product of thousands of hours of sensing and sense-making in tandem, building neural architecture that can hold multiple data streams simultaneously and surface a gestalt. Her younger colleagues aren't less intelligent; they're being trained in a different perceptual environment, one where the integration step is increasingly handled upstream.

Ultimately, tools aren't replacing our sensing, they’re replacing our sense-making. And because sense-making is where meaning lives and where perception becomes judgment, that is where the capacity loss becomes a risk to our sovereignty as thinkers, leaders, and practitioners.

Sense-making has always been a collective process as much as an individual one. We inherit frameworks for interpretation: what counts as evidence, which patterns matter, how to weigh competing signals against each other. These shared maps let us coordinate and build on each other's perception. When you say "something feels off" and I nod because I've felt it too, we're working from the same underlying read of the room. But what happens when we diverge? When some of us are sense-making through algorithmic mediation and others through embodied perception, we may no longer even see what the other is seeing. The fragmentation we are already experiencing, in politics, in epistemology, in basic shared reality, is about perception bubbles as much as it is about information bubbles. Increasingly, we are no longer sensing the same world.

This is the hardest kind of gap to argue across. You can show someone different facts, but it is much harder to show them a different perception. When the upstream layer of how we register reality starts to differ between people, the downstream conversation drifts in ways neither party can quite name. The disagreement feels personal, even moral, when it may actually be perceptual. Part of the bridge generation's task, then, is to keep us legible to one another. To make sure that, even as our tools diverge, our capacity to recognize what the other person is seeing holds.

Our choices are quickly becoming the baseline for those without a "before,” and the dance we learn to do, or fail to do, will become the primary dance the next generation learns. Beyond which tools to use, we are encoding the rules of engagement that will shape what they can even perceive as possible. Whether consciously or not, we are quickly deciding which forms of human sensing remain embodied and which get outsourced. And, in doing so, we’re also determining which frameworks for sense-making get transmitted and which atrophy from disuse. This is architect work, shaping the structures that will define what can be sensed and how sense gets made.

The ER doctor above is doing for medicine what many of us are doing across our domains: figuring out what to preserve and what to outsource while training people who will not have her "before." She’s teaching her younger colleagues to notice when their embodied sense-making conflicts with the algorithmic version and to investigate that dissonance rather than defer to the numbers. She’s also aware that she will not be there forever. Eventually, it’s possible that the people making these decisions will be doctors who never practiced without AI support, who never developed that particular form of perception because they never had to.

Here, homeostasis becomes more than metaphor: The body adjusts, listening for signals, weighing them against context, returning to a steady state. What we owe the next generation, alongside the tools, is the discernment to notice when our sense-making is being determined more by technology than by the world we are actually trying to understand. To listen. This will require us to take seriously the gap between what the system says and what the body knows.

AI already mediates human perception. The real question at hand is what we pass on alongside the tools and whether the next generation inherits only the systems, or also the capacity to know when to step around them.