The Last Demonstration:
On Proof, Pride, and the Fatal Theatre of Expertise

On a sweltering July afternoon in 1993, a 38-year-old Toronto lawyer named Garry Hoy walked to the edge of a window on the 24th floor of the Toronto-Dominion Centre, faced a group of articling students, and launched himself into the glass. It held. Buoyed by the confirmation of what he already knew, he did it again. This time the glass didn’t shatter. It popped, clean and whole, from its frame — and Garry Hoy fell 24 stories to his death.

The coroner’s verdict was death by misadventure. The newspapers, inevitably, couldn’t resist. He won a Darwin Award posthumously, that peculiar internet-era garland reserved for people who remove themselves from the gene pool through spectacular self-defeat. Late-night hosts chuckled. Office workers forwarded it around on early email. A man died, and the world found it funny.

I have been thinking about Garry Hoy for a long time. Not about the tragedy of his death — though it was tragic — but about something far more unsettling: the possibility that he was completely right, right up until the moment he wasn’t.

This is the story I want to tell. Not about foolishness, but about the seduction of proof. About what happens when expertise becomes performance. About the razor-thin line between knowing something and needing others to know that you know it. About what philosophers, behavioural scientists, and the great humanists of literature have understood for centuries: that the most dangerous moment in any person’s life is not when they are ignorant — it is when they are certain.

The Architecture of Certainty

To understand what happened on that July afternoon, you have to understand how human beings build the edifice of certainty in the first place. We do not reason from first principles every morning when we walk down stairs or turn on a tap. We rely on pattern. We rely on memory. We rely on the accumulated evidence of ten thousand unremarkable confirmations that the world is the way it appears to be.

The philosopher David Hume called this the problem of induction — the famous chicken-and-sunrise problem. A chicken has seen the sun rise every morning of its life. It expects the sun to rise tomorrow. Then comes the day it doesn’t, or the day the farmer comes instead. The chicken’s confidence was not irrational; it was simply built on an insufficient sample of an unpredictable universe. Garry Hoy was Hume’s chicken in a three-piece suit, on the 24th floor.

The structural engineer Bob Greer, commenting on the incident, reportedly remarked that he knew of no building code in the world that would allow a 160-pound man to run against a glass and withstand it. But here is what that statement obscures: the glass had withstood it. Many times. The failure was not in the glass. The failure was in the frame — in the hidden system that held the glass in place, a system that had been weakening, millimetre by millimetre, with each of Hoy’s previous demonstrations. He wasn’t testing the glass. He was eroding its foundation.

The metaphor is almost too elegant to bear.

When Knowledge Becomes Theatre

There are two kinds of knowing. There is the private kind — the quiet confidence that comes from genuine understanding, held internally, consulted when needed, and never requiring an audience. And there is the performed kind — the knowledge that only becomes real when witnessed, validated, and applauded. The first kind is wisdom. The second kind is a vulnerability dressed up as strength.

Garry Hoy was, by all accounts, a brilliant lawyer. Senior partner at Holden Day Wilson. A man known for his intellectual rigour, his fascination with engineering, his precise and disciplined mind. In a different context, his demonstrations of the window’s strength might have been charming — a party trick that delighted and impressed. But on that particular afternoon, something shifted. He was performing for articling students. Young lawyers at the beginning of their careers, looking up at a man at the top of his. The asymmetry of that audience matters enormously.

The sociologist Erving Goffman, in his landmark 1959 work The Presentation of Self in Everyday Life, described human social interaction as an elaborate theatrical performance — “dramaturgy,” he called it. Every person, in every social encounter, is simultaneously actor and audience, performing a version of themselves calibrated to the expectations of the room. The expert performs expertise. The leader performs authority. The confident person performs certainty — and in performing it, often comes to believe it more completely than the evidence warrants.

This is the mechanism that I believe was operating on the 24th floor of the Toronto-Dominion Centre. Not recklessness. Not stupidity. Dramaturgy. The performance of a self — the self that knows, the self that is not afraid, the self that can be trusted — in front of precisely the audience most likely to amplify the seduction of that performance.

The Hubris Problem: A Very Old Story

The ancient Greeks had a word for it, of course. They always did. Hubris — the overweening pride that invites the punishment of the gods. But the Greeks were too sophisticated to make hubris simply about arrogance. The truly terrifying Greek insight, the one that runs through Oedipus and Agamemnon and Icarus alike, is that hubris is most dangerous in the person who has actually earned the right to be confident. It is not the fool who flies too close to the sun. It is the brilliant craftsman’s son, the one who genuinely has wings, the one who has every reason to trust them.

Icarus did not fall because he was careless. He fell because he had flown so well, for so long, that the experience of competence overwhelmed the instruction of caution. The wax was real. The wings were real. The sun was real. And the gap between those realities — the failure to account for the interaction between a known success and an unexamined variable — is the entire story of human overreach, in every domain, in every age.

In our own time, the annals of professional catastrophe are full of Garry Hoys. Not people who didn’t know what they were doing, but people who knew exactly what they were doing — and therefore stopped asking whether they should be doing it. The engineers at Morton Thiokol who knew the O-rings on the Challenger were vulnerable to cold but struggled to communicate the weight of their knowledge against the momentum of institutional confidence. The traders at Long-Term Capital Management — two Nobel laureates among them — who built models of such breathtaking elegance that they forgot the models were models. The surgeons whose complication rates rise, rather than fall, after a run of successes. Competence can be a drug. And its most addictive form is competence witnessed.

The Audience We Perform For

I want to stay with the question of audience for a moment, because I think it is the part of this story that most of us are most reluctant to examine in ourselves.

When we perform our expertise — when we throw ourselves against the glass — who are we really doing it for? The young articling students standing in that Toronto boardroom had everything to gain from seeing Hoy’s trick succeed. Their admiration was the most obvious reward on offer. But I suspect Hoy was also performing for a more demanding audience: the part of himself that needed the knowledge to be confirmed again, publicly, in front of witnesses. Because private knowledge, unwitnessed knowledge, is somehow never quite enough. We want our knowing to be seen. We want our confidence to be legible. We want the room to understand, beyond all doubt, that we are the person who knows.

There is something deeply human in that want. And something deeply perilous.

The poet Rainer Maria Rilke, in his Letters to a Young Poet, wrote one of literature’s most quietly radical pieces of advice: “I would like to beg you, dear Sir, as well as I can, to have patience with everything unresolved in your heart and to try to love the questions themselves.” Rilke was writing about the interior life, about the spiritual practice of sitting with uncertainty. But his words carry a professional and intellectual weight that I have been returning to for years. The willingness to love the questions — to not rush toward answers, demonstrations, proofs — is one of the rarest and most valuable forms of professional courage.

Garry Hoy did not love the question. He had already answered it. The window was strong. He knew. He had always known. And the knowing had become, over time, something he could not bear to leave private.

What Leaders Get Wrong About Confidence

I work with leaders. I have spent a significant portion of my professional life helping people navigate complexity — inside organisations, inside communities, inside the fog of their own expertise and experience. And the pattern I return to most often, the one that sits at the centre of my work, is this: the leaders who cause the most damage are rarely the ignorant ones. They are the competent ones who have confused their competence for completeness.

True authority — the kind that actually moves people, builds institutions, and earns lasting trust — is not performed. It is practised. It lives in the quiet, unglamorous discipline of continuing to test your assumptions even when you’ve confirmed them a hundred times. It lives in the recognition that confidence shared too loudly becomes a cage rather than a gift. It lives in what the Japanese call shoshin — beginner’s mind — the Zen concept that holds that the expert who approaches their field with the openness of a beginner will always outperform the expert who has closed their mind with the weight of what they already know.

Margaret Heffernan, in her remarkable book Wilful Blindness, argues that the failure to see what is inconvenient — what challenges our existing model of the world — is not a cognitive error so much as a social one. We surround ourselves with people who confirm what we believe. We build careers on the foundations of ideas that once worked. We reward confidence and punish uncertainty. And then we express horror when the confident and unreflective fail spectacularly.

But the horror is misplaced. We built the system that produced the failure. Every time we promoted the certain over the careful, the decisive over the deliberate, the performer over the practitioner, we moved another millimetre of erosion into the frame that holds the glass.

The Frame, Not the Glass

I keep coming back to what the structural engineer said. The glass was not the problem. The frame was the problem. The invisible infrastructure. The system of small, unremarked connections that held everything in place — until it didn’t.

In every domain of human endeavour, there is a version of the frame problem. In medicine, it shows up as the invisible cumulative wear of diagnostic overconfidence — the senior physician whose pattern recognition is so finely honed that it stops generating hypotheses and starts generating verdicts. In finance, it shows up as the slowly eroding risk culture inside institutions that have not had a crisis in a generation. In relationships, it shows up as the assumption that what has held together for years will continue to hold together without attention, without renewal, without the patient work of maintenance.

We test the glass. We forget the frame. We test the part that is visible, dramatic, and confirmable. We ignore the slow accumulation of stress in the connections we cannot see. And when the frame gives way, we call it an accident. A misadventure. A freak event. We do not call it what it actually is: the predictable consequence of confident neglect.

The philosopher Nassim Nicholas Taleb, in The Black Swan, describes the phenomenon of “silent evidence” — the data we never see because the events that would have generated it did not survive. We see the swimming champion; we do not see the thousand equally talented swimmers who trained with equal dedication and simply did not make it. We see Garry Hoy’s last demonstration. We do not see the hundred previous ones, and we have no way to factor their success into an accurate model of what was actually happening to the frame each time.

Every success conceals a hidden cost. The question is whether we are building systems — personal and professional — that can see those costs before they become irreversible.

The Humanity of It All

I want to be careful, here at the turning point of this essay, not to reduce Garry Hoy to a lesson. He was not a lesson. He was a person — a brilliant, passionate, physically and intellectually courageous person who wanted to share something he loved with a group of young people at the beginning of their careers. There is something genuinely beautiful in that impulse, even as there is something genuinely tragic in its consequence.

The humanist in me recoils from the Darwin Award, from the sneering laughter that greeted his death. It is easy, and deeply unsatisfying, to explain human tragedy through the lens of stupidity. It is harder, and far more honest, to look at what Garry Hoy did and recognize in it something we all carry: the need to be believed. The need to prove. The need to perform, for others and for ourselves, the person we have worked so hard to become.

Albert Camus, in The Myth of Sisyphus, argues that the fundamental question of human existence is not whether life is fair, but whether it is worth living in full knowledge of its absurdity — its tendency to offer up the wrong window frame at precisely the wrong moment. His answer, famously, is that we must imagine Sisyphus happy. That the willingness to push the boulder, knowing it will roll back, is itself a form of heroism. That meaning is made in the act, not the arrival.

I am not sure Camus would have approved of my use of his philosophy here. But what strikes me is this: Garry Hoy pushed the boulder. He pushed it because he believed in it. And the tragedy is not that he pushed — the tragedy is that no one in the room, not even he himself, had the presence of mind to ask: but what is happening to the frame?

What We Owe Each Other

The articling students who witnessed Garry Hoy’s death that afternoon were not bystanders. None of us are, when we stand in the presence of performed certainty. When we watch the senior partner throw himself against the glass, we become part of the event. Our silence is a form of participation. Our admiration is a vote for continuation.

This is, I think, the most practically important thread in this entire story. We have built professional cultures — legal, medical, financial, political — in which the expression of doubt is a form of social risk. To say “I’m not sure that’s safe” in the presence of the expert performing their expertise is to commit a kind of social transgression. It reads as insubordination. As naivety. As a failure to understand how things are done around here.

James Reason, the British psychologist who spent his career studying human error, introduced the concept of “organizational accidents” — failures that are not caused by a single bad actor or a single bad decision, but by an accumulation of small, unremarkable choices, each individually defensible, that together create the conditions for catastrophe. Every choice to not ask the question. Every choice to laugh at the trick instead of examine the frame. Every choice to reward the performance instead of the practice.

We produce Garry Hoys in our organizations. We promote them. We celebrate them. And then, when they fall, we write the obituary in the passive voice and call it a misadventure.

Amy Edmondson of Harvard Business School has spent three decades documenting what she calls “psychological safety” — the team culture variable that most consistently predicts organizational performance, resilience, and the capacity to avoid catastrophic failure. It is the freedom to say, without social cost, “I think we should check the frame.” It is the freedom of the articling student to put up a hand and ask: “Garry, have you considered what’s happening to the window mounting over time?”

No one in that room exercised that freedom. And we cannot know whether, if they had, it would have made any difference. But we can build the systems — in our organizations, our relationships, our professional cultures — where that question is not only possible but welcomed. Where the person who asks it is not seen as undercutting the expert, but as doing the expert the highest honour: treating them as someone who can handle the truth.

A Final Word on Peter Pan

The passage that opens the story I began with — “To die will be an awfully big adventure” — comes from J.M. Barrie’s Peter and Wendy, spoken by a boy who cannot die, in a world where consequences are always one step removed from reality. Peter Pan is the great literary monument to the seduction of invincibility. He never grows up not because he refuses to — it is because growing up, in Barrie’s vision, is inseparable from the experience of failure, loss, and the revision of certainty that comes from being genuinely hurt by the world.

Garry Hoy was not Peter Pan. He had grown up. He had built a real career, earned real expertise, accumulated real knowledge. But in the moment of that second demonstration — the one with the students watching, the one with everything already proven once — something in him reached for the Neverland logic: the glass held before, so the glass will hold again, and holding it will feel magnificent, and the students will remember it, and I will be the person who knew.

Growing up — really growing up, not just ageing — means accepting that the glass held last time and may not hold this time. That each demonstration is a new event, not a confirmation of an old one. That the appropriate response to a hundred successes is not a hundred-and-first performance, but a quiet question: what have I not checked?

Peter Pan thought death would be an awfully big adventure. Garry Hoy, I suspect, thought the window would hold. Both were wrong about the physics of the world they lived in. One was a fictional boy in a story about the price of permanence. The other was a real man in a building in Toronto, on an afternoon in July, with students watching.

We remember him with laughter. I think we owe him more than that. I think we owe him the harder tribute: to look at the frame we are all leaning against, and ask, quietly, sincerely, without an audience — how strong, really, is this?