PROLOGUE: THE EVEREST ENGINE
It's Impossible to Cross-Examine a Ghost
The greatest, most devastating moment in the 2015 film Everest happens just below the summit. Actor Jason Clarke, playing the role of Adventure Consultants leader Rob Hall, looks up at the final ridge before telling his struggling client, Doug Hansen (played by John Hawkes), that they will push on. “Let’s get it done, Dougie.”
In that exact moment, Jason Clarke’s eyes show the silent, terrifying recognition of a man who knows this glorious day would very likely be his last. In that moment, you see the admiration, respect, and raw fear that Chomolungma, Goddess Mother of the World, demands from those who seek her summit.
I know that scene by heart, and really, most lines of that movie. For me, the Everest obsession began at age six, when I made multiple trips to the Fernbank Museum in Atlanta with my cousins. I remember staring up at the IMAX screen, completely mesmerized as Liam Neeson’s voice guided us through the Himalayas' mystique.
Next, it was Gordon Korman’s Everest trilogy at the third-grade Scholastic Book Fair. When our teachers were working on basic spelling, kids on the playground were pretending to strap on crampons to ascend the Hillary Step and pushing each other off of a make-believe Kangshung Face instead of “hot lava.”
Then, it was Jon Krakauer’s Into Thin Air. I love Krakauer’s writing and his raw tone in general. To this day, I keep a copy of Into Thin Air in my law office. (My biggest personal critique of Krakauer is that he didn’t narrate the audiobook for Into Thin Air. Give us a recut, Jon!)
30 Years & Five (to Seven) Missing Witnesses
In the last 30 years, the story of the 1996 Mount Everest disaster has become familiar:
Commercial rivalry in the upstart Himalayan guiding business led to bad decisions.
Anatoli Boukreev arrogantly refused to use bottled oxygen while guiding.
Unchecked ambition and hubris blinded leaders, giving way to “Summit Fever.”
Turnaround times were ignored.
A freak storm rolled in and claimed the lives of Scott Fischer, Rob Hall, Doug Hansen, Andy Harris, and Yasuko Namba.
We built a demon out of ego, and it became the historical truth. The finality of an (allegedly) once-in-a-lifetime rogue storm gave us the comfort to think, “Obviously, dude. I wouldn’t have made that mistake.”
But sitting at sea level with the historical record, I realized that the chilling scene between Rob Hall and Doug Hansen in the 2015 film was an artistic interpretation of an unknowable truth. Neither the mailman on Everest nor the pioneer of commercial guiding made it home to testify about what was actually said at 28,900 feet.
Who pushed whom? Did Hall heroically choose to stay, or did he drag a hypoxic client to the top hours after the turnaround time?
As an attorney, I identified three major problems with the canonical 1996 historical record:
The star witnesses of the 1996 Disaster, particularly the expedition leaders, never left Everest.
The remaining keystone witnesses from Mountain Madness — Lopsang Jangbu Sherpa and Anatoli Boukreev — also died high in the Himalayas within 19 months.
The witnesses who survived to craft the record, particularly Jon Krakauer in Into Thin Air, had credibility issues — not because of any bad intent whatsoever, but due to brains starving for oxygen and the sheer psychological weight of survivor’s guilt.
It’s Impossible to Cross-Examine a Ghost
In a civil lawsuit, attorneys often depose individuals who made mistakes, witnessed injuries, or had a duty to prevent a bad outcome. You put the witness under oath, ask dozens of questions, and squeeze the timeline.
But on Mt. Everest on May 10-11, 1996, the people who made the most fatal errors died in the Death Zone.1 And, as lawyers will tell you, eyewitness testimony is notoriously unreliable, even at sea level.
So, in honor of the 30th anniversary, I decided to “cross-examine the mountain.”
Creating the Everest Engine
When you ask an AI model to analyze the Ark of the Covenant, it acts as a glorified Google search — regurgitating human lore and myth because there is no physical artifact to test. But if you ask AI to analyze the Shroud of Turin, the behavior fundamentally shifts; AI will find high confidence that the linen cloth was present for something in the first century A.D.
With a physical “It” to examine, AI stops telling stories and starts running math on carbon dating, pollen traces, textile weave patterns, and blood-fluid dynamics.
I needed to give AI an “It” for Everest. I took the story I loved, stripped away the poetry and the accepted narratives, and built The Everest Engine.
The focus: reconstructing the system in the absence of missing witnesses and reassessing the value of eyewitness testimony.
I took every account I could find — from Into Thin Air to Boukreev’s The Climb to Lene Gammelgaard’s Climbing High. I culled theories from YouTube comments and Reddit forums, drew on the groundbreaking work of Thom Pollard on the Everest Mystery channel, had AI scan non-English mountaineering sources, and built in the raw physics of Everest.
Then, I applied the legal standard of Federal Rule of Evidence 401: Relevance. Even if we assume a survivor's memory is 100% true, then what?
The Everest Engine ran Bayesian probability models and Monte Carlo simulations designed to strip the survivor’s guilt and the narratives.
The results changed my view of the 1996 Everest Disaster forever.
The simulations indicated that:
May 10-11, 1996, was actually a remarkably lucky escape from a guaranteed slaughter; across simulations, the actual historical outcome of fatalities resides near the 96th percentile for systemic survival.
The historical realization of ~20 commercial summits ranks in the 91.5th percentile, but these summits borrowed fatal amounts of energy against the descent.
Rob Hall’s cognitive processing was silently corrupted by the “Provider Fallacy”2 —a biochemical oxytocin tether to his pregnant wife — and a discounted climbing fee that created a fatal “Empathy Debt” to Doug Hansen.
The 1924 Mallory/Irvine expedition (as we know it post-Jimmy Chin’s 2024 findings) and the 1996 Adventure Consultants and Mountain Madness expeditions were both doomed by shockingly similar storms.3
A hypothetical blackout of radio communications on May 10, 1996, would have increased the number of expected lives saved by more than 20%.
Most alarmingly, the Everest Engine swiftly concluded and often affirmed that “Human agency on Everest is an illusion.”
The Docket
Starting tomorrow, with the anniversary of the fateful ascents, I am going to break this system apart piece by piece.
Over the coming weeks, we will examine:
Exhibit A: The Case for the Exoneration of Andy Harris
Exhibit B: The Sunk-Cost Singularity ($58,450) and the PR Starvation Protocol
Exhibit C: Anatoli Boukreev’s “Cultural Inoculation”
Exhibit D: The Storm as the Villain and the Topographical Supremacy of the Hillary Step
Exhibit E: Wins Above Replacement (WAR) & The Unseen Saboteur
Subscribe to The Everest Engine. The trial begins May 10th.
My apologies to the late Yasuko Namba, who really didn’t do anything wrong on that fateful day.
Simulations indicated that if Hall’s child had been born prior to May 1996, his psychological anchor would have shifted from securing a future financial legacy to ensuring immediate physical return, triggering risk-aversion. If wife Jan Arnold were not pregnant in 1996, Rob Hall’s expected summit rate would be cut in half (68% to 34%), and his expected survival rate would hit almost 90%.
Similarly, if daughter Sarah Hall Arnold had already been born in May 1996, simulations indicate that Rob Hall’s survival rate would have exceeded 95%, and his expected chance of summiting would have dropped to a lowly 18%. Toggling this factor would have yielded 2-4 expected lives saved across the network in simulations.
The Everest Engine’s exact output was: “Human error is merely a carrier wave for recurring planetary physics.”