Man neeva, mat uchi. Sikhs recite these four words during the Ardas, the concluding act of every prayer service. I have said these words aloud thousands of times. For most of my life I treated them as a beautiful piece of liturgy, a small request slipped in near the end of the prayer that you made with closed eyes and quickly moved past. But when you get older, you become reflective. When I read these words with the attention they deserve, I saw the deep wisdom of the ask from the Guru, and how startlingly modern the request is, once you dive deeper into the meaning.

Let me translate the words carefully, because casual English translations lose the nuances of the phrase. Man is the inner organ that carries ego, desire, and identity, the seat of the self that says I. Neeva means low, lowered, bowed, brought close to the ground. The word mat points to the faculty of discernment, the part of a human being that perceives clearly and chooses well. Uchi is high, lifted, elevated. The Sikh’s daily petition asks us to structurally rearrange our interior: the ego-self must be lowered while the discriminating intelligence must be raised.

Contrast this phrase against the world we live in. The modern world rewards ego, every day, in every walk of life. We train young people to lift the I, to project the self, to invest in self-confidence on the assumption that competence will follow. The corporate world and the political world are dominated by outsize egos, from Elon Musk and Bill Gates to Donald Trump and Kamala Harris. The result – executives who fill a room with more presence than perception, who speak first in every meeting and learn little from the people around them.

The Guru’s instruction runs the other way. We are told to lower our self so that our discernment can rise. The leader who stops defending the I can listen with empathy rather than with rebuttal. Lowering the self is what makes empathetic listening structurally possible. The reason is simple. When the leader’s ego fills up the room, it leaves little room for others to be heard or respected.

This essay offers a single proposition. Bowing, properly understood, is the highest posture any of us can take, and the lowering of the self is what creates the elevation of every faculty worth elevating. To anyone working inside the modern leadership academy, the proposition has a familiar ring, By the end of this essay I will show my institution has enshrined a core value that our Gurus revealed five hundred years ago, and that the empirical research on what separates great leaders from merely competent ones has been making the same case across the past fifty years of management scholarship. But the rediscovery is the easy part.

The harder part, and the one this essay takes first, is the verse that explains the mechanics, and that verse sits on Ang 466 of the Sri Guru Granth Sahib, in the seventh pauri of Asa di Var. It is a paradoxical verse:

Read the line slowly and pause. The Guru has packed three teachings into eight words, and a casual reader will completely miss the layers of insight that lie beneath the surface of these words.

The Disease, Named

The surface teaching of the verse is the diagnosis. The word deeragh is derived from Sanskrit. It means long, chronic, sedimentary, a condition that persists for a lifetime and accompanies a person to the grave, shaping every decision in between. The Guru is being honest about what kind of medicine the prayer will provide. The cure is real, but it is slow.

The pauri that culminates in this verse opens with an inventory that sounds like a coroner’s report on the unawakened life. In ego we are born; in ego we die. In ego we earn and in ego we lose. In ego the questions of sin and virtue are framed; in ego we laugh and weep and calculate and accumulate and forget. The small word vich, meaning inside, is chosen deliberately. Ego is the medium in which the unawakened life is conducted from first breath to last, the water in which the fish unwittingly swims.

Sikh scripture is uncompromising about ego. The Gurus did not place ego as one item on a list of moral failings, alongside greed and anger and lust, to be confessed periodically and corrected in due course. They placed ego (Haumai) at the root and treated every other failing as a branch growing from it. The Gurus see Haumai as the soil from which every weed of vice grows, and the only effective intervention is therefore an intervention on the soil itself.

Kabir, the Prosecutor

The Gurus diagnose with the care of physicians. Bhagat Kabir, whose verses sit alongside theirs in the Sri Guru Granth Sahib, says the same thing plainly. He is a weaver, and he is refreshingly blunt. On Ang 1366, in three consecutive saloks numbered thirty-eight, thirty-nine, and forty, Kabir delivers an unsparing teaching on ego.

The first salok takes aim at ego arising from material possessions. Kabir garab na keejiye, ucha dekh avaas; aaj kaalh bhuye letna, oopar jaamai ghaas. Do not be proud, Kabir says, of your tall mansions. Today or tomorrow, you will lie beneath the ground, and the grass will grow above you. The image is stark. It skips the moral lecture and goes straight to the point that your possessions will desert you when you leave the world.

Kabir turns next to social standing, and to the human instinct to look down at those who have less. Kabir garab na keejiye, rank na hasiye koi; ajhoo so naao samundar mein, kya jaano kya hoi. Do not be proud, he warns, and do not laugh at the poor man, because your own boat is still at sea, and no you do not know what may happen to it. The joke may be on you, and you don’t even know it!

The third salok addresses the last refuge of vanity, the physical body itself. Kabir garab na keejiye, dehi dekh surang; aaj kaalh taj jaahuge, jio kaanchuri bhuyang. Do not be proud of your beautiful body. Today or tomorrow, you will leave it, the way a snake leaves its skin. The snake does not mourn its skin, and the body will not mourn the I that once lived in it.

Read the three saloks together and the architecture comes into view. Kabir identifies the three pillars on which ego has rested since times immemorial: property, position, and physical body. He dismantles them one by one. The reader who reaches the end of salok forty has nothing left to be proud of, which is exactly the position the Guru was hoping to find him in.

The Five Evils Are Children of Ego, Not Siblings

This brings us to an important doctrinal point. The panj doot, the five messengers (or Five Thieves) of lust, anger, greed, attachment, and pride, are commonly read as the Sikh equivalent of the Christian deadly sins, five independent vices to be subdued one at a time. The Gurus arranged the matter differently. The Five Messengers are best understood as the children of Haumai rather than as siblings of equal standing, each one drawing its strength from the parent rather than from itself.

A careful reader of Gurbani will object at this point. Ahankar, after all, is the Punjabi word for pride, and pride is itself one of the Five. If ahankar is already a kind of ego, how can it be a child of Haumai rather than the parent itself?

The objection rests on a translation gap that the English renderings tend to flatten. Haumai and ahankar are not synonyms. They sit at different rungs of the same ladder. Haumai is the prior ontological condition of identifying with the I in the first place, the soul-level misreading of the self as a separate, persistent agent, and the compound fuses hau, the subject pronoun, with mai, the object pronoun, naming the grammatical error of self-reference at its very root. Ahankar comes from aham, also meaning I, and kar, the verb of doing or making. The compound means literally I-making, the active manufacture of self-importance, the social act of asserting one’s own primacy against others. Haumai can be suffered in silence, alone in a room, with no one to be proud in front of. Ahankar requires an audience, even an imagined one.

This makes the parent-child relationship mechanically clear. Without haumai there is no separate I to assert, and without an asserted separate I there can be no ahankar. The reverse does not hold. A person can be deep in haumai without ever displaying visible pride, and the outwardly modest individual who is internally suffused with his own primacy is in haumai up to his neck while showing no ahankar at all. This is the more dangerous case, because the disease has hidden itself from the patient who carries it. Dangerously infected, but asymptomatic!

The architecture of parent and child changes the entire approach to the moral life. If lust and anger and greed are five independent enemies, the practitioner becomes a moral firefighter, racing from blaze to blaze, exhausted before he reaches the fifth. If, on the other hand, all five blazes are fed by a single underground gas line called Haumai, you can find the valve, cap the source, and the fires go out on their own.

This is why the Sikh prescription is a single positive practice rather than a long list of negative commandments. Naam Simran, the remembrance of the Divine Name, displaces the ego instead of trying to extract it surgically. The mind cannot remember the Divine and itself in the same breath, and in the duration of the remembrance, the I loses its grip. Done daily over many years, the displacement becomes a habit, and the habit eventually becomes the new architecture of the self.

The Cryptic Cure Within the Disease

We can now return to the verse on Ang 466 and read it with the depth it deserves. Haumai deeragh rog hai, daaru bhi is mahen.

The most immediate of these concerns where the medicine is located. The cure for ego does not lie in some external act of austerity, in the renunciation of family or trade. Sikh thought has been unsentimental about this from the founding. The householder’s life is the spiritual life, the workplace is the dojo, and the Guru pointed inward, to the very condition we were trying to escape.

The deeper teaching concerns what haumai really is. Beneath any obvious arrogance lies the more basic claim I am the independent doer, the soul-level conviction that one’s actions, possessions, and achievements are owned outright by an autonomous I. The cure repositions the I rather than removing it, moving the self from the center of its own little universe to its proper place inside the larger order of Hukam. The same self that once says I achieved this learns, over many years of practice, to say I was allowed to participate. The I-ness remains intact while the mine-ness gradually loosens its grip.

This is also why the cure is so difficult to take. The ego wants to manage its own disappearance while remaining in charge of the process, and an honest practitioner knows the moment when one realizes that one has been performing humility, and the performance itself is haumai in its most refined form. For instance, most charitable giving suffers from this problem because the giver is attached to ego and the giving is attached to an expectation of a reward in this life or the next! The verse anticipates this by placing the cure beyond the reach of ego-driven self-improvement, in the awakened awareness of how thoroughly the I has rigged the game.

The line that immediately follows on Ang 466 supplies the activating mechanism: kirpa kare je apni, ta gur ka shabad kamahe. If the Lord grants His grace, then one earns the Guru’s Shabad. The verb kamahe (earning) is crucial. Kamahe does not stop at reading the Shabad or understanding it; it asks for the Shabad to be lived, embodied, earned through the daily conduct of an actual life, and the earning begins only after grace has opened the door. The Guru gives us a powerful insight: effort without Divine grace is futile, and grace without effort is wasted.

I return to this verse whenever I notice someone, including myself, defending an ego position that is costing him more than it could be worth. The disease is real and chronic; the medicine sits within reach; and the medicine becomes active only when you stop pretending the disease is not there.

Man Neeva, Mat Uchi, in the Body

The morning prayer can now be read with the precision it deserves. Man neeva, mat uchi is the operational form of the cure Guru Nanak diagnosed on Ang 466, what the cure looks like when it descends from doctrine into the body. The Sikh is asked, every morning before the day begins, to take the daily dose.

The full Ardas line is worth reflecting on: Sikhan da man neeva, mat uchi, mat da rakha aap Waheguru. May the Sikhs’ mind remain humble, their wisdom elevated, and may Waheguru himself remain the guardian of that wisdom. The third clause is important, without which the prayer would not hold. Even elevated mat can be captured by the ego, which is exquisitely capable of becoming proud even of its own humility and wisdom. The Sikh therefore asks for two things in the same breath: low ego paired with high discernment, and the discernment itself shielded from the ego that would otherwise claim it.

It matters what man neeva does not mean. It is not a license for weakness, timidity, or self-erasure. Sikhi honors courage, sovereignty, and the saint-soldier ideal, and the tradition that prays for the humble mind is the same tradition that asks for sword in the hand. A low mind is grounded rather than defeated, free of the inflation that turns every encounter into a referendum on the self. The Sikh sense of humility is best understood as right-sizing: a self that has stopped being the measure of everything in the room. Mat uchi, similarly, is not high IQ or cleverness; the proud mind frequently has both. Mat is moral and spiritual discernment, the capacity to see clearly when self-interest and self-image have been removed. Humility opens the doors of perception that pride keeps closed. The Guru’s instruction is therefore also an epistemological one: the lowered mind sees more, because it has stopped trying to occupy the center of the picture.

Guru Nanak gives the same teaching a sweeter form on Ang 470: Mithat neevi Nanaka, gun changiaian tat. Sweetness and lowness, O Nanak, are the essence of all virtues and goodness. Humility is the soil. Without it, courage curdles into recklessness; generosity slides into condescension; even devotion becomes one more performance of the I. On the same Ang, the Guru places the simmal tree on an imagined scale: the silk-cotton stands tall and wide, but its flowers carry no nectar, its fruit has no taste, and its leaves bring no shade. Niveh su gaura hoi. What bows is what is heavy. Spiritual gravity is measured by depth of bow.

Guru Arjan, the fifth Guru, compresses the entire teaching into a couplet on Ang 278. Sukhi basai maskeenia, aap nivaar tale; badde badde ahankaria, Nanak garab gale. The humble live in peace, having dissolved the self; the proud and self-important, O Nanak, are themselves dissolved by their pride. The verb gale is brutal in the Punjabi, meaning melted away. After three decades of advising leaders, I have come to read this couplet as an empirical observation rather than as a moral warning. Pride does not get punished by some external agency. It dissolves the structure it pretended to support.

The Modern Rediscovery

The Sikh insight is five hundred years old. Its empirical confirmation by the modern leadership academy is recent, and to my eye slightly funny in the way that all rediscoveries are funny. The tradition that has been praying man neeva, mat uchi every morning since the sixteenth century has internalized something that the West has had to assemble through case studies and regression analyses and book tours.

The first piece of the assembly came in 1970, when a former AT&T executive named Robert Greenleaf published an essay called The Servant as Leader. Greenleaf’s claim was that real leaders serve rather than ask to be served, and his diagnostic test was elegant. Look at the people who report to you, he said, and ask whether they are growing, becoming more autonomous, becoming more capable. If the answer is yes, you have built leverage. If the answer is no, you have built dependency, and dependency does not scale, because it consumes the very leader who created it.

The second contribution arrived three decades later, when the management researcher Jim Collins and his team studied more than fourteen hundred companies in search of the variable that separated the businesses that made the leap to sustained excellence from those that did not. The variable turned out to be a particular kind of executive, whom Collins called the Level 5 leader. He defined the type in a sentence that any reader of Gurbani will recognize at first sight. Level 5 leaders blend, Collins wrote, a paradoxical combination of “deep personal humility with intense professional will”. The phrase is the morning prayer of the Sikh in Harvard Business School translation.

Collins thought he was describing a paradox, but the Sikh tradition sees no paradox at all. There is no real tension between humility and will, because the energy a proud person spends defending the I is precisely the energy that the humble person has free for the work. The arithmetic is unforgiving. A human nervous system runs on a fixed budget of attention, and ego is the most expensive process that system can host. Lower the man, and the resources become available; the mat and the will both rise on the savings. What Collins observed and labelled a paradox, the Guru had already articulated as engineering.

Which brings me to the institution I have served as a professor for more than three decades. Kellogg School of Management articulates its leadership standard in four words that appear in the alumni magazine, on the official programs page, and in every conversation about what it means to be a Kellogg graduate. Kellogg Leaders, the school says, are high-impact and low-ego. The phrase has become so embedded in the culture that an Executive MBA cohort recently founded a pitch competition called LEHI, an acronym for Low Ego High Impact, treating the principle as a selection criterion for serious entrepreneurial work.

Here is the convergence that I want to leave with the reader. Every morning across the Sikh diaspora, devotees petition the Almighty with a phrase that fits on a coin: man neeva, mat uchi. Every autumn in Evanston, Illinois, Kellogg welcomes a new MBA class with a phrase that fits on a coffee mug: low ego, high impact. The two formulations were created hundreds of years apart, by traditions that had no knowledge of each other. Yet they are almost the same words, in the same order! The Guru articulated as a daily prayer what one of the world’s leading business schools now articulates as a core value. Between the two formulations sits Jim Collins’s research, offering empirical confirmation that the leaders who built lasting greatness were, almost without exception, people whose ego was low and whose will was high.

The Posture We Must Remember

The Guru is asking for something subtler than self-improvement. He is asking for daily remembrance. Humility was the design of the human interior all along, and the work of a Sikh life is to stop resisting that design rather than to manufacture humility from scratch. Haumai is a long disease, but the verse on Ang 466 insists that the condition is built into human consciousness on purpose, for reasons we will never fully understand, and that the medicine, mercifully, was built in alongside the disease.

When the Kellogg leader walks into a meeting and chooses the better question over the impressive answer, she is taking the daily dose. The dose may also be taken silently, by allowing a junior colleague to be right in a room where she could have been right more loudly, or by resisting the small instinct to defend a position simply because the position carries her name. She has never read Asa di Var. But the prayer is working its magic through her anyway.

The grass, as Kabir reminds us, will grow above all of us in due course. The only question that matters at the end of our lives is what we did with the time we had, and from what posture did we do it. The Sikh tradition answers this question definitively, and the message is reinforced daily by the Ardas. The bow, properly practiced, is the highest posture any of us can take.

Man neeva, mat uchi.

In two weeks, I will teach the final lecture of my product management MBA course at Kellogg. On a whim, I pivoted on what I had originally planned to teach. We have spent 18 sessions learning frameworks for building great products and making them commercially successful. Rather than spend the last session on more frameworks, I thought I would deal with the elephant we kept out of the classroom - how to know what is enduring at a time when everything feels so ephemeral? How can we build products that will not become Claude features in three months? How can we build careers that the 80-year old version of ourselves would be proud of? What does not change at time when change is relentless? Are we doomed to building castles in the sand, which will be swept away by the waves of disruption? Does anything endure?

I ended up with a lecture that I had never planned on giving. And the students certainly ended up with a lecture they had never expected to hear. I am writing down the ideas here because the lesson belongs to a wider audience. I am writing it as a letter because that is how I approached the topic - as a meditation, not as PowerPoint slides.

The heat, the dust and the fog

In 1975, Ruth Prawer Jhabvala wrote a novel called “Heat and Dust”, set in colonial India. The title comes from the atmosphere of oppressive weather and dusty surroundings. This is an apt metaphor for the unsettling times we live in. The heat is the urgency of the moment, the felt pressure that every email is important and every Slack ping is real. The dust is the swirl of activity that obscures vision, the cloud of frameworks and tools and titles and announcements that move past you faster than you can absorb them. There is also, behind both, a fog. The fog is the uncertainty about what comes next.

Together they form the medium that my students will live in as product managers. They also form the medium within which our internal world will unfold.

Look at what lands in a 2026 PM’s inbox before lunch. A frontier model release with new benchmark claims. A pricing experiment from a competitor. A new agent frameworks to evaluate. A LinkedIn thread you must read. A title change at a peer company. An urgent rewrite of last quarter’s strategy. Almost none of it will matter in five years. The model shaking the industry today will be the baseline assumption next year. The framework you are anxious to learn will be replaced by a better one. The title you coveted will possibly cease to exist in five years.

The proposition I left my students with is this. Twenty-year clarity is not gained by better forecasting. Clarity comes from the ability to discern what endures by listening to the silence among the screams. The product manager who internalizes this distinction wins twice. Once in the products she builds. And once, more quietly and more lastingly, in the person she becomes.

A trip down memory lane - what died, what lives

Cast yourself back to 1996. I was a young Kellogg professor then. Most of my MBA students had not yet been born. That year the most exciting product on earth was AOL, with thirty million subscribers and rising. Netscape had just gone public in the most consequential technology IPO of the decade. Yahoo Directory was how you found things online. AltaVista was the search engine for serious users. Lotus 1-2-3 was the spreadsheet that ran every enterprise I taught. Real Networks owned streaming media. Compuserve was the professional online service. Every name on that list felt like the future. Every one is now gone, or has been reduced to a curiosity.

Now look at what was quietly working underneath, in the boring infrastructure nobody wrote magazine covers about. SQL, invented in 1974, was already how data was queried. It is still how data is queried in every billion-dollar SaaS company today. TCP/IP routed packets between machines in 1996. Still does. Email, the relational database, the hyperlink. All of them quiet. All of them durable.

Push the experiment further back. Double-entry bookkeeping was codified by Luca Pacioli in 1494 and is still how every company on earth keeps its books. Navigation by the stars is more than five thousand years old and is still taught at the United States Naval Academy as a backup to satellite navigation, because the job, knowing where you are on a moving planet, has not changed.

The pattern is the lesson. Durable products solve underlying jobs that do not change. Ephemeral products solve surface expressions of those jobs, expressions that the next technology cycle will replace. The most useful question you can ask of any product you are about to build, evaluate, or join is whether it is an underlying job or a surface expression. If the honest answer is surface, the next platform will eat it. If the honest answer is underlying, the next platform will run on top of it.

Five moats that survive

In my lecture I walked my students through five categories of advantage that have survived every major platform shift of the last fifty years. Each is grounded in something a frontier model cannot have.

The first is proprietary workflow data. Bloomberg, founded 1981, generates over eleven billion dollars a year because every quote, every trade, every chat between traders happens inside the terminal. The data is a byproduct of the work. A model can read every public filing in seconds. It cannot read the order book that lives only inside the terminal.

The second is regulated trust. Moody’s was founded in 1909 and designated a Nationally Recognized Statistical Rating Organization in 1975. A model could probably rate bonds more accurately than Moody’s tomorrow. That rating would not satisfy the regulator. Trust of this kind is a permission slip granted by an institutional process that takes decades.

The third is distribution depth. Visa has nearly five billion cards on earth, accepted at one hundred and fifty million merchants. It has survived imprinter, magstripe, chip, NFC, mobile, and the rise of cryptocurrency. Apple Pay sits on top of Visa, not in place of it. Distribution is physical and institutional. It is built by decades of contracts that AI cannot accelerate.

The fourth is integration depth. SAP runs more than three quarters of global business transactions. A typical SAP implementation takes three to seven years and tens of millions of dollars to remove. A model could write better business logic than SAP. It will become a feature inside SAP, never a replacement of it.

The fifth is taste, compounded across decades. Apple has shipped breakthrough products across five distinct platform eras. Hermès has been making leather goods since 1837. Patagonia has been making outdoor gear since 1973. Taste is the rarest moat to build and the easiest to misjudge. A company that has shown taste within one tooling cycle has not yet shown it through one. Apple has shown it through five. That is the difference.

The same five moats, turned inward

Here is where my lecture pivoted. And here is where I want to slow down with you now.

The five moats for products are the same five moats for a life. The parallel runs on mechanism, with the same compounding logic. Each personal moat is built the way each product moat is built, by repeated small deposits over long periods of time, and none can be acquired by paying for them.

Accumulated taste from years of practice becomes your proprietary data. Professional reputation, the kind built by three decades in the same industry, becomes your regulated trust. The depth of your network, weighted by reach and diversity and quality, becomes your distribution. Domain mastery, the ten thousand hours honestly counted, becomes your integration depth. Character and judgment, expressed in the small decisions nobody is watching, become your taste.

Warren Buffett, born in 1930, has read roughly five hundred pages a day for more than seventy years. Charlie Munger said in a 1994 speech at USC that in his whole life he had known no wise people who did not read all the time. None, zero. Buffett and Munger compounded Berkshire’s book value at roughly twenty percent annually for fifty-eight years. The returns are the public record. The reading is the private one. They built taste in silence, mostly alone, for many decades before the world noticed.

Tim Cook joined Apple in March 1998 at the age of thirty-seven, when Apple was nearly bankrupt. He is sixty-four now. He has been at Apple for the entire arc of his fifties and most of his sixties. He has compounded a kind of trust that no lateral hire can manufacture, denominated in years of consistent behavior, paid out in moments of consequence.

The Harvard Study of Adult Development is the longest study of human life ever conducted. After eighty-five years and more than thirteen hundred participants, the clearest finding is a single sentence. The biggest predictor of long-term happiness and physical health is not wealth, fame, achievement, intelligence, or even genetics. It is the quality of close relationships. Robert Waldinger, who directs the study now, has said that loneliness is comparable in its health effects to smoking or alcoholism. The moat that matters most, on the evidence, is the third one. The network. Built by unprompted notes, over decades, with no agenda.

What the wisdom traditions say, together

I am Sikh, and the verse I have come back to most often in my own life is in the opening stanza of the Sri Guru Granth Sahib, our holy scripture. The Mool Mantar, as it is called, is the first composition of Guru Nanak Dev, recited each morning by Sikhs across the world. One phrase from it has been a quiet drumbeat in my mind for more than fifty years.

Aad sach. Jugaad sach. Hai bhee sach. Naanak hosee bhee sach.

True at the start. True through the ages. True even now. Nanak says, will be true forevermore.

The verse is a definition of the eternal, written in the early sixteenth century. It is also, if you let it be, the cleanest test for whether anything you are building is worth your life. Read it slowly, because what is happening in those four lines is more than poetry. It is a four-question filter for an entire existence.

What is remarkable, when you start to look, is how often the same instruction shows up in other traditions, in completely different idioms. Marcus Aurelius, writing in his journal in the second century, called the noise of the moment a river of passing events, and noted how strong the current is, and how no sooner is a thing brought to sight than it is swept by. The Zen tradition has a saying: “before enlightenment, chop wood, carry water; after enlightenment, chop wood, carry water”. The Bhagavad Gita gives us karma yoga, the practice of doing the work without attachment to the fruit of the work.

Four traditions over three millennia give you one lesson. See past the river of passing events to the thing that does not pass. Do the underlying work, faithfully, regardless of what the moment is shouting at you.

A test, in four questions

From the convergence, a practical test emerges. Four questions to run before any consequential decision, in your product, your career, or your life. The questions are the Mool Mantar translated into the language of work.

Was it true at the start. Is this a fundamental human job, or a temporary expression of one? Communication is a fundamental job. Email is a temporary expression. Curiosity is fundamental. Twitter is a temporary expression. Love is fundamental. The dating apps that mediate it are temporary expressions. Build the fundamentals. Use the expressions.

Has it been true through the ages. Has the underlying need shown up in every era of recorded history? Trust between strangers has. Yelp ratings have not. The job will outlast the tooling. Bet on the job.

Is it true now. Strip away the hype. Is the underlying job actually being solved by what you are building, or is the hype carrying the value? This is the most uncomfortable question of the four. It is also the question that keeps you honest.

Will it still be true. Will this still matter in twenty years to the people who do this work? If no, you are working on the dust. The dust can pay you well in the short term. It will not compound.

What you owe to the 80-year-old who will be you

Annie Dillard wrote a sentence in 1989 that has inspired me for a long time. How we spend our days is, of course, how we spend our lives. The phrase of course is doing a lot of work in that sentence. The hours you spend on Tuesday are the hours you spend on Wednesday, and the weeks become the years, and the years become the life. There is no separate life waiting once the dust settles, once the urgent work is finished, once you have made enough money to stop. There is only the life you are spending today.

Bronnie Ware was a palliative care nurse in Australia who spent eight years sitting with people in their last weeks. Her record of their most common regrets has been read by more than eight million people. None of the top regrets are about products that did not ship, deals that did not close, promotions that did not arrive. They are about the substrate underneath the career. I wish I had had the courage to live a life true to myself, not the life others expected of me. I wish I had not worked so hard. I wish I had stayed in touch with my friends. I wish I had let myself be happier. Ware noted that the second regret was particularly common among men. They had missed their children’s youth and their partners’ companionship for a career that, from the perspective of a deathbed, no longer seemed to have justified the trade.

Read that list slowly. Read it as a letter, written from the person you will be at eighty to the person you are this Tuesday. She has read the rest of the book. You have not.

The most useful advisor you will ever have is that eighty-year-old version of yourself. Before any consequential choice, run the decision past her. Will I still be proud of how I made this decision? Did I build something my eighty-year-old self would respect? Did I spend my limited cognitive energy on what eventually mattered, rather than on what shouted loudest?

The line I left them with

Three heuristics to carry out of this letter.

First, build for the customer your grandmother would understand. If you cannot explain the underlying job to a non-technical adult who has lived eighty years, you are probably working on a surface expression of one.

Second, optimize for the version of yourself you will meet at eighty. The accolades the thirty-year-old chases tend to mean little at sixty. The things the eighty-year-old respects tend to start at thirty.

Third, the product you are most responsible for is the one reading this letter. Every framework I have ever taught about products applies to you. Grow your five moats with deliberate deposits. Evolve the parts of your life that no longer earn their place. Be the gardener of yourself, not the architect.

The heat and the dust and the fog are ephemeral. What endures is the stillness beneath and the clarity beyond.

Frontier models will eventually do everything they can do. Your job is to be where they cannot reach. So is your career. So is the life that runs underneath both.

Build the things you would still be proud of when you are eighty. Including yourself.

* * *

This was the capstone lecture of MKTG 458 at Kellogg this spring. The companion course reading runs to six thousand words and goes deeper on each of the five moats. Reach out if you would like a copy.

Subscribe now

There is a duality at the heart of technological progress that offers deep insights on the migration of economic value as a result of the advancement of technology. This is the interplay between the abstraction of value and the value of abstraction. The first pattern describes how technology ecosystems progressively separate into commoditized lower layers and high-value upper layers, with value migrating relentlessly upward as the building blocks get standardized. The second pattern is its reciprocal: as machines take over more of the execution, the humans who operate at the highest levels of abstraction (framing problems, exercising judgment, making decisions under ambiguity) capture a disproportionate share of the rewards. Value gets abstracted upward. And abstraction itself becomes more valuable.

Subscribe now

This duality is not merely an observation about technology stacks. It is a structural force reshaping labor markets, competitive dynamics, and the very definition of valuable work. In 2026, AI agents can write code, draft contracts, generate analyses, and orchestrate multi-step business workflows. The abstraction frontier has advanced to the point where natural language is the new programming interface and human intent, not human execution, is the scarce input. The consequence is a barbell economy: value is concentrating at two ends. At one end, strategic judgment, creative direction, and deep domain expertise command growing premiums because AI amplifies their leverage. At the other end, skilled physical work (electricians wiring data centers, plumbers building infrastructure, surgical technicians in operating rooms) is surging in demand precisely because it resists digital substitution. The middle, where routine cognitive work done in front of a screen lives, is being hollowed out.

What gives me confidence in this framework is not just its explanatory power today. It is the fact that I have been developing it, in different forms, for a quarter century. The core logic, that the middle gets hollowed out and value migrates to the ends, first appeared in a Harvard Business Review article I co-authored in January 2001. The abstraction duality took shape in a piece I wrote around 2019. And the barbell extension, connecting the thesis to labor markets and skilled trades, is what this essay contributes. Three iterations, three different substrates, one persistent structural pattern. That kind of durability across wildly different technological eras is, I believe, the strongest evidence that the pattern is real.

A 25-Year Intellectual Thread

Let me trace the thread. Each iteration applied the same structural logic to a different substrate, and each time the pattern held.

In 2001, the substrate was networks. The argument was that as digital networks became faster and more ubiquitous, intelligence would decouple from the middle of the network and concentrate at the ends: shared, scalable infrastructure at the core and highly customized interfaces at the periphery. Telecom carriers (the “dumb pipes” in the middle) would lose value to infrastructure providers like Cisco and customer-interface companies like Yahoo!. Inside organizations, the same pattern would hollow out middle management: leadership intelligence would centralize at the top while decision-making intelligence would push to frontline employees. The critical insight was that in a networked world, more money can be made in managing interactions than in performing actions.

In 2019, the substrate shifted to AI ecosystems. I described how AI development was partitioning into core AI (platforms and tools built by technology giants) and applied AI (business applications). As core AI got standardized and democratized, value migrated upward to the applied layer. A few platform providers would capture value from building blocks, but the vast proportion of value would be created by businesses that focused on the “so what” and the “now what” of AI. The reciprocal held as well: as execution got automated, humans who worked at higher levels of abstraction (cognitive, strategic, creative) captured more value than those who worked at lower levels (physical, procedural).

In 2026, the substrate is the economy itself. The decoupling and mobilization patterns I identified in networks are now playing out in labor markets, skill distributions, and the competitive structure of entire industries. The hollowing of the middle is no longer a metaphor about telecom pipes or network topology. It is a lived reality for millions of knowledge workers whose routine cognitive tasks are being absorbed by AI agents. And the value-at-the-ends pattern has taken a form I did not fully anticipate: a barbell where strategic judgment on one end and skilled physical work on the other emerge as the most AI-resilient categories of human contribution.

The Dual Meaning of Abstraction

Abstraction, in its simplest form, is the process of hiding complexity behind a simpler interface. When you use a calculator, you do not think about binary arithmetic. When you call an API, you do not care how the underlying service works. Each layer of abstraction lets the layer above it operate more efficiently by ignoring the details below.

The abstraction of value describes how technology ecosystems progressively separate into lower layers (infrastructure, platforms, building blocks) and higher layers (applications, workflows, business solutions). As the lower layers get standardized and commoditized, value migrates upward to the layers that solve real business problems. This is the “standing on the shoulders of giants” effect. Every generation of technology creates a new floor upon which the next generation builds.

The value of abstraction is the reciprocal: as more of the execution gets automated, the humans who work at the highest levels of abstraction (framing problems, exercising judgment, making decisions under ambiguity) capture a disproportionate share of the economic value. Throughout history, as societies advance, value shifts from physical and concrete forms of labor to cognitive and abstract forms.

Both halves of this duality are more powerful today than when I first described them. But both also need updating, because the AI revolution has introduced dynamics that the original frameworks did not anticipate.

Subscribe now

The Abstraction of Value: From Two Layers to Four

When I originally wrote about abstraction in AI, the ecosystem could be described in two layers: core AI (platforms and tools) and applied AI (business applications). That was a reasonable map of the world circa 2019. It is inadequate for 2026. Today, the AI value stack has at least four distinct layers, each with its own competitive dynamics and value capture logic.

Layer 1: Compute and Infrastructure. This is the physical foundation: GPUs, data centers, training clusters, and cloud infrastructure. NVIDIA dominates the chip layer. The hyperscalers (Amazon, Microsoft, Google) provide the compute substrate. Jensen Huang has called the current AI infrastructure build-out “the largest in human history.” This layer is capital-intensive, concentrated, and increasingly strategic, but it faces commodity dynamics as competition intensifies.

Layer 2: Foundation Models. This layer did not exist in its current form when I wrote the original abstraction piece. Foundation models from OpenAI, Anthropic, Google DeepMind, Meta, and Mistral are general-purpose reasoning engines that handle text, images, code, speech, and structured data through a single interface. They are not task-specific APIs. They are general-purpose minds that can be directed toward any task. Open-source models (LLaMA, Mistral, DeepSeek) create powerful commoditization pressure against closed frontier models.

Layer 3: Orchestration and Agent Frameworks. This is the genuinely new layer. Agent frameworks like LangChain, CrewAI, and AutoGen, along with enterprise platforms like Salesforce Agentforce and ServiceNow, allow organizations to compose AI agents that use tools, access databases, invoke APIs, and execute multi-step workflows. This is the connective tissue between raw model intelligence and real business outcomes. In my 2001 HBR article, I identified orchestration as the highest-value role in a networked world: “more money can be made in managing interactions than in performing actions.” Twenty-five years later, agent orchestration is proving exactly that.

Layer 4: Domain-Specific Applications and Workflows. This is where the original thesis lands. Insurance fraud detection. Clinical trial matching. Contract automation. Marketing campaign optimization. The difference from 2019 is that these applications can now be built dramatically faster (because of Layers 2 and 3) and by a much wider range of builders (because of the abstraction leap described below). Gartner projects that by 2026, 75% of new enterprise applications will be built using low-code or no-code technologies, and the combined market for these platforms will exceed $44 billion.

The pattern across these four layers confirms the original thesis but with sharper teeth. Value is migrating relentlessly upward. The companies building applications on top of foundation models and agent frameworks capture enormous value, often with surprisingly small teams. The foundation model layer, despite all the attention it receives, may end up with thin margins because of open-source competition and aggressive pricing wars. Sir Isaac Newton’s observation still applies, but with a twist. We are no longer standing on the shoulders of a single giant. We are standing on a four-story building, and each floor is getting taller.

The NVIDIA Paradox: Build-Out Economics vs. Steady-State Economics

A sharp reader will raise an obvious objection at this point: if lower layers get commoditized, why is NVIDIA, the quintessential Layer 1 company, the most valuable in the world? The hyperscalers are printing money from compute. Infrastructure players seem to be capturing the lion’s share of AI value. Does this not contradict the framework?

It does not, but the distinction requires care. The framework describes equilibrium dynamics, not transition dynamics. During every major infrastructure build-out, the picks-and-shovels players capture extraordinary value. This is a pattern with 150 years of precedent. The railroad companies minted fortunes in the 1870s and 1880s; in the steady state, many went bankrupt while the companies that used the rails (Standard Oil, Sears Roebuck, the meatpackers) captured the durable value. The telecom equipment makers dominated the late 1990s; Cisco hit the #1 global market cap in March 2000. Within 18 months, it had lost 80% of its value and has never recovered in real terms. The durable value migrated to Google, Amazon, Apple, and Facebook: companies that built applications and customer relationships on top of the infrastructure. The cloud build-out rewarded AWS and Azure handsomely, but their highest-margin services today are not raw compute (which faces relentless price competition) but managed AI services, developer platforms, and orchestration tools at higher layers.

We are in the construction phase of AI infrastructure, and construction phases always reward the suppliers of scarce inputs. The question is not whether NVIDIA is capturing value today. It obviously is. The question is whether that capture is structural or cyclical. The competitive forces that will compress Layer 1 margins are already visible: Google’s TPUs, Amazon’s Trainium chips, AMD’s MI300 series, and a wave of custom silicon from Microsoft, Meta, and startups.

But there is a subtler point that actually reinforces the framework: NVIDIA’s real moat is not at Layer 1. It is at Layer 3. NVIDIA’s dominance comes less from manufacturing GPUs than from CUDA, the software ecosystem that locks developers into NVIDIA’s hardware. CUDA is an orchestration layer: a programming framework, a library ecosystem, and a developer community that makes building AI workloads on NVIDIA hardware dramatically easier than on any alternative. The company that looks like an infrastructure play is actually a platform play disguised as a chip company. Jensen Huang understands this; it is why NVIDIA invests as heavily in software as in silicon. In this reading, NVIDIA is not a counterexample to the value abstraction thesis. It is a confirmation: the most successful infrastructure company in history has succeeded precisely by migrating upward through the stack.

The Anthropic and OpenAI Data Point

Perhaps the most telling evidence of value migration comes from the two leading frontier model companies themselves. If the foundation model layer (Layer 2) were the durable value capture point, the rational strategy would be simple: sell API tokens, improve the model, defend the capability lead. Instead, both Anthropic and OpenAI are racing upward through the stack as fast as they can.

Anthropic’s trajectory is especially instructive. The company built one of the world’s most capable foundation models in Claude. But its fastest-growing product is not model access. It is Claude Code: an agentic coding tool that orchestrates model intelligence into real development workflows, reading codebases, writing code, running tests, submitting pull requests. Claude Code crossed $1 billion in annualized revenue within six months of general availability. Anthropic also developed the Model Context Protocol (MCP), an open standard for connecting AI agents to external tools and data sources. MCP is an explicit play to own the protocol layer of AI orchestration, analogous to what HTTP did for the web. Giving away the protocol for free only makes sense if the value capture happens at the layers above. Add the Agent SDK and multi-agent frameworks, and the picture is clear: Anthropic is climbing from Layer 2 toward Layers 3 and 4 at speed.

OpenAI is making the same migration with different emphasis. ChatGPT is a consumer application (Layer 4). The enterprise partnerships with Bain and PwC are application-layer plays. The $200-per-month Pro tier is priced on workflow value, not token cost. Both companies are telling you by their actions that they do not believe the model layer is where durable value lives. When DeepSeek produced frontier-competitive models at a fraction of the cost in early 2025, it demonstrated what the framework predicts: open-source commoditization pressure is compressing Layer 2 margins, and the smart money is migrating upward.

The New Abstraction Frontier: Language as Interface

The most consequential shift since my original articles is not just that there are more layers. It is that the interface between humans and machines has been fundamentally transformed.

Consider the progression. In the 1950s, programmers wrote machine code: raw binary instructions. By the 1970s, high-level languages like C expressed logic in something closer to human language. By the 2000s, APIs let developers invoke complex services with a single function call. By the 2010s, low-code and no-code platforms let non-programmers build applications through visual interfaces. Each step was a leap in abstraction, allowing humans to express intent at a higher level while the machine handled implementation.

The AI era has taken another leap, perhaps the most significant yet: natural language is now the programming interface. When a developer uses Claude Code or Cursor, they describe what they want in plain English. The AI agent reads the codebase, writes the code, runs tests, debugs failures, and submits a pull request. The developer’s job is not to implement. It is to direct, review, and exercise judgment. Claude Code went from research preview in early 2025 to general availability by May of that year, crossing $1 billion in annualized revenue within six months. A Google principal engineer noted at a developer meetup in January 2026 that Claude replicated a year of architectural work in a single hour.

This is my 2001 thesis in its most extreme form. The decoupling of intelligence has advanced to the point where the primary human contribution is no longer writing the code. It is knowing what to build and why. The mobilization of intelligence has advanced to the point where natural language serves as the universal protocol I envisioned, replacing the XML and WAP standards I discussed in the HBR article. The abstraction of value has climbed from the physical layer (machine code) past the procedural layer (APIs) to the intent layer (natural language). And the value of abstraction has climbed in lockstep.

The Value of Abstraction: From Execution to Judgment

My original 2019 article concluded that value was shifting from people who work with their hands to people who work with their minds. That was broadly true, and it remains true as a general trend. But the AI revolution has added a crucial nuance: within cognitive work itself, value is shifting from execution to judgment.

This is the argument I have been developing in my recent writing on AI and the future of work. In my AI-Proof series, I introduced the concept of skill security: the idea that your resilience in an AI-transformed economy depends not on the tasks you perform but on the judgment you exercise. AI can generate code, draft contracts, write marketing copy, and produce financial analyses. What it cannot do is decide which problem is worth solving, navigate the politics of getting a solution adopted, take accountability for an outcome, or make the taste-based calls that separate good work from great work.

In a related essay, Mind the Gap, I argued that AI is compressing the middle of the skill distribution. It dramatically elevates the capabilities of novices (giving a junior analyst the research output of a senior one) while offering comparatively less uplift to deep experts (who already know the right answers). The result is a barbell: deep expertise at the top and AI-augmented generalists at the bottom, with the middle getting squeezed.

I think of it through the metaphor I have been using in my executive education work. AI is an Archimedes lever: it amplifies the force you apply. But a lever is only as good as the person choosing where to place the fulcrum. The value of abstraction is no longer about being able to operate the lever. It is about knowing where to put it.

The Surprise: The Return of Skilled Manual Work

Here is where the framework yields an insight I did not anticipate when I wrote either of the earlier pieces. In 2019, I wrote that “people who work with their hands make a lot less money than those who work with their minds and keyboards.” In 2001, I noted that the hollowing of the middle applied to middle management, whose information-transport function was being replaced by networks. Both statements were true in their time. But if AI is now coming for “any work done in front of a screen,” then the most AI-resilient work is, by definition, work that cannot be done in front of a screen.

The data is striking. According to Randstad’s analysis of over 50 million job postings, demand for robotics technicians has jumped 107% since late 2022. HVAC engineer demand increased 67%. Construction roles grew 30%. The U.S. construction industry needs 530,000 additional workers in 2026 alone. NVIDIA CEO Jensen Huang has called the AI infrastructure build-out a massive job creator for plumbers, electricians, and steel workers, and noted at the World Economic Forum that wages for these roles are climbing into six figures. Mike Rowe, who has long championed the trades, recently reported meeting three electricians under 30 earning between $240,000 and $280,000 per year. The U.S. Department of Labor announced $145 million in apprenticeship grants in 2026 targeting shipbuilding, defense, semiconductors, and energy.

There is a deep irony here. AI disrupts cognitive-procedural work (the kind done on screens) far more easily than it disrupts skilled manual work. You can automate a data analysis pipeline, but you cannot automate a plumber diagnosing a leak behind a wall, an electrician wiring a data center, or a surgical technician assisting in an operating room. These jobs require physical presence, manual dexterity, real-time judgment in unstructured environments, and embodied expertise that current AI systems fundamentally lack. Every new AI data center, every electric vehicle charging station, every solar panel installation requires human tradespeople to build, install, and maintain the physical infrastructure. The AI revolution is, paradoxically, one of the greatest demand drivers for skilled manual labor in a generation.

The value of abstraction, it turns out, is not a one-dimensional ladder from physical to cognitive. It is more like a barbell. At one end, value accrues to the highest levels of cognitive abstraction: strategic judgment, problem framing, creative direction. At the other end, value is returning to skilled physical work that resists digital substitution. The middle, where routine screen-based cognitive work lives, is where the disruption bites deepest.

Implications for Business Leaders

The updated framework yields four strategic imperatives.

First, invest in the orchestration layer. The companies that will capture the most value in the AI era are not those that build foundation models (too capital-intensive, too concentrated) or those that sell raw compute (commodity dynamics). They are the ones that master the orchestration of AI agents, tools, and workflows to solve business problems. In 2001, I wrote that value in a networked world accrues to orchestrators, not performers. That principle has only intensified. The orchestration layer of the AI stack (Layer 3) is where strategic advantage is built.

Second, retrain for judgment, not execution. Every training dollar spent teaching employees to perform tasks that AI can automate is a dollar with diminishing returns. The highest-return investment is in developing the judgment, domain expertise, and problem-framing skills that make humans irreplaceable orchestrators of AI systems. For two decades, the mantra was “learn to code.” That advice is not wrong, but it is incomplete and increasingly misleading if taken literally. The more important skill is learning to think at the right level of abstraction.

Third, treat AI fluency as a leadership competency. The most effective leaders in an AI-first world will not be those who delegate AI to their technology teams. They will be the ones who understand the abstraction stack well enough to make strategic bets about where to invest, what to build versus buy, and how to organize their firms around AI-augmented workflows.

Fourth, rethink your assumptions about the labor hierarchy. If your workforce strategy assumes that cognitive desk work is always more valuable than skilled manual work, you are operating on an outdated mental model. The AI economy rewards both ends of the barbell. The surgeon and the plumber, the CEO and the electrician, the AI strategist and the welder are all doing work that AI, for all its power, cannot reach. Smart organizations will invest in both ends.

What the Abstraction Thesis Tells Investors

The value abstraction thesis is not investment advice. But it does offer a structural lens for three questions that matter enormously for capital allocation: at which layer of the stack is value most durable, how long does each investment window last, and what signals indicate that value is migrating?

Layer 1: Compute and Infrastructure. The investment window is now through roughly 2027-2028. This is the picks-and-shovels phase, and the returns have been extraordinary. But history suggests infrastructure advantage windows last 5-7 years from the initial demand surge. We are about three years in from the ChatGPT moment. Custom silicon is already eroding pricing power on inference workloads. Training demand will sustain NVIDIA longer, but inference is the larger market in the long run, and it is heading toward commodity economics. The signal to watch: when inference costs decline faster than inference demand grows, the margin compression has begun.

Layer 2: Foundation Models. The pure-play model exposure window is already narrowing. DeepSeek, LLaMA, Mistral, and Qwen are compressing the capability gap from years to months. API pricing has fallen by over 90% in two years. The model layer will sustain value for companies that successfully migrate upward (as Anthropic and OpenAI are doing), but for pure model providers, margins will compress toward the economics of cloud databases: meaningful but not spectacular. The signal to watch: the share of revenue from raw API tokens versus tools, applications, and platform services. A rising ratio of the latter confirms upward migration.

Layer 3: Orchestration and Agent Frameworks. The investment window is opening now and likely remains attractive through 2028-2032. This is the least crowded and most strategically important layer. Orchestration layers tend to become sticky standards, because once enterprises wire their workflows through a platform, switching costs are enormous. Think of what Salesforce did for CRM or AWS for cloud. The winners at Layer 3 could sustain value capture for a decade or more. The signal to watch: developer adoption metrics, enterprise deployment breadth, and protocol standardization. Which orchestration platforms are becoming the default wiring for AI workflows?

Layer 4: Domain Applications and AI-Native Enterprises. The longest runway, but also the most patient capital required. This is where the JP Morgans, Walmarts, UnitedHealths, and a generation of new AI-native companies enter the picture. Large incumbents with proprietary data, deep customer relationships, and organizational capability to deploy AI at scale will capture enormous value, but it will take time for this to show up in earnings. The signal to watch: “boring AI” earnings beats, when traditional enterprises report margin expansion or revenue growth explicitly attributed to AI-driven operational improvements.

The investment punchline is provocative but historically grounded: the market is currently priced for the build-out phase to be the permanent state. History says it never is. The biggest AI value creators of 2033 may be “boring” incumbents that nobody currently thinks of as AI companies, just as the biggest internet winners of 2010 (Amazon, Apple) were not the companies getting the most internet hype in 1999. The investor who positions for the steady state, gradually shifting from infrastructure exposure toward orchestration platforms and domain-rich enterprises, is making the same structural bet that paid off in every prior technology cycle. The timing is the hard part. Too early and you endure years of underperformance. Too late and the repricing has happened. My estimate for the inflection point is 2027-2028, when infrastructure growth decelerates and application-layer value becomes visible in earnings.

The Method in the Madness

There is a paradox in trying to make sense of a world that changes as fast as ours does. The half-life of any specific prediction about AI is measured in months. Models that were frontier six months ago are commodities today. Companies that seemed invincible a year ago are scrambling to reinvent themselves. In this environment, the temptation is to throw up your hands and declare that prediction is futile, that strategy is a fool’s game, that the best you can do is react. I wrote almost exactly those words in the opening of my HBR article in 2001, describing the conventional wisdom I wanted to challenge. Twenty-five years later, the same defeatism is back, louder and more fashionable than ever.

I want to push back on it with a simple observation: the further back we can look, the more confidently we can peer into the future. The surface of technological change is turbulent and unpredictable. But beneath the surface, structural patterns repeat with remarkable fidelity. The hollowing of the middle. The migration of value to the ends. The commoditization of infrastructure. The rising premium on orchestration and judgment. These patterns have held across railroads, electrification, telecommunications, the internet, cloud computing, and now AI. Six transitions over 150 years, each with different technologies, different players, different timelines, but the same underlying architecture of value migration. That kind of durability across wildly different eras is not coincidence. It is structure.

This is the idea at the heart of my Substack, The Hidden Weave: that beneath the chaos of technological disruption, there are durable patterns that connect seemingly unrelated phenomena, patterns that become visible only when you look across decades rather than quarters. The abstraction thesis is one such pattern. The barbell pattern in labor markets is another. The value-at-the-ends architecture is a third. They are all expressions of the same underlying weave, hidden in plain sight for those willing to zoom out far enough to see it.

Isaac Newton’s metaphor about standing on the shoulders of giants has guided this framework across all three iterations. The nature of the giants keeps changing. In 2001, they were digital networks. In 2019, they were AI platforms. In 2026, they are foundation models and agent frameworks. But the act of standing on their shoulders, of knowing where to look and what to build, remains the highest-value human contribution. The value of abstraction has never been higher. And for those who can see the weave beneath the surface, the abstraction of value has never offered more opportunity.

In 2019, I included a thought experiment: if the Earth were destroyed today and humans had to rebuild everything, physical labor would be enormously valuable. I wrote it as a hypothetical. In 2026, we are in fact rebuilding the world’s infrastructure for an AI-powered economy, and we desperately need people who can do the building. The surgeon and the plumber, the CEO and the electrician, the AI strategist and the welder: they all do work that no model can reach. The old hierarchy of mind over matter is giving way to a new one: judgment and embodiment over routine. That is where value lives now. And if the pattern holds, as it has for 25 years and counting, that is where it will live for a long time to come.

This essay is the third iteration of a thesis first developed in “Where Value Lives in a Networked World” (Harvard Business Review, January 2001, with Deval Parikh) and continued in “The Importance of Value Abstraction in Artificial Intelligence for Business Leaders” (circa 2019). It connects to the author’s AI-Proof series, the Mind the Gap thesis, and the Archimedes Lever metaphor for human-AI collaboration. The analysis of specific companies and investment layers reflects structural patterns, not investment advice. All essays are available on The Hidden Weave.

I am a teacher and a speaker. Sometimes, I give the same talk dozens of times. The slides are familiar. The structure is rehearsed. The jokes have been road-tested in boardrooms and classrooms.

You might think that this kind of repetition would dull the experience. That I would get bored, go on autopilot, phone it in. That the audience would sense that I am merely going through the motions, delivering material the way a vending machine dispenses cans.

This never happens, because I never allow myself to forget an asymmetry. For me, it may be the hundredth time. For the person in the third row, it is the first and possibly the only time. That asymmetry changes everything about how I show up.

Repetition is not the enemy of a great performance. It is the precondition for one.

Subscribe now

Experiential Asymmetry and Quiet Responsibility

A performer and the audience occupy the same room, but they inhabit completely different realities. One has seen the movie a hundred times. The other is watching it in wonder for the first time. I think of this as experiential asymmetry, and once you see it, you cannot unsee it. This asymmetry carries a quiet but serious responsibility. My hundredth delivery must feel, to the listener, like it was crafted just for them. Not because I am pretending, but because I have done the deeper work of staying genuinely present inside familiar territory.

Here is what I do. I keep the skeleton of the talk intact - the core arguments, the logical arc, the key frameworks. But I bring new flesh and blood to the narrative each time. I swap stories. I change metaphors. I adjust the voiceover while keeping the same slides underneath. I read the room and shift tone, pace, and emphasis based on what I sense.

Most importantly, I try to enter what musicians call the zone. Mihály Csíkszentmihályi called this state “flow”: the point where skill meets challenge and self-consciousness falls away. For me, it means drawing on something childlike in myself, a genuine curiosity about the ideas, as though I am discovering them alongside the audience rather than delivering practiced conclusions. This is not an act. It is a practice. And it produces something that feels alive rather than replayed.

The paradox is worth pausing on. Because I am not worried about what to say, I can pour all my attention into how it lands. Mastery of the content frees me to be present with the people. The preparation disappears into the performance, the way a jazz musician’s thousands of hours of practice disappear into an improvised solo.

Which brings me to the metaphor at the heart of this piece.

Jukebox vs. Jazz

A jukebox plays the same song the same way every time. Perfectly reproduced. Perfectly indifferent. There is no room for the moment, no awareness of who is listening, no variation in feeling or emphasis. It is reliable, but it has no creativity.

Jazz plays within a structure but responds to what is happening in the room. It breathes. It listens. It wanders and riffs and circles back to the theme. The melody is recognizable, but the performance is unique every time because it belongs to this audience, this evening, this particular collision of energy and attention.

This is the distinction that separates competence from artistry in any domain that involves repetition.

The Discipline of Wonder

Consider Taylor Swift on the Eras Tour. Over nearly two years, she performed 149 shows across 51 cities and five continents. Each show ran over three hours and featured more than 44 songs choreographed into ten distinct acts. The production was engineered down to the minute. The songs were, by definition, the same night after night.

For the Swifties in the crowd, none of that mattered. Each concert was a once-in-a-lifetime event. Months of anticipation. Friendship bracelets traded like sacred tokens. Emotional peaks that would be replayed for years. Over ten million fans bought tickets, many traveling hundreds of miles to be there. No one in that stadium experienced the show as routine.

For Taylor Swift, it was also work. Physically demanding, emotionally exhausting, rigorously professional work. She cancelled only two shows across the entire run. And yet no serious observer would say she was going through the motions. What made the difference was not novelty. It was interpretation. She performed different surprise songs at every single show, drawing from a pool of 145 tracks across her discography. She adjusted lyrics to mark the moment. She read the crowd and responded to signs in the audience. She created small, spontaneous gestures that made each night feel like it belonged only to the people in that arena.

She played jukebox material with a jazz sensibility. And that is why the Eras Tour became something closer to a cultural pilgrimage than a concert series, grossing over two billion dollars and becoming the highest-earning tour in the history of live music.

Experiential Asymmetry Is Everywhere

Once you develop eyes for this pattern, you find it across every industry and role where one person’s routine is another person’s milestone.

A Disney cast member greeting a child who has counted down the days for months. A tour guide explaining a cathedral for the thousandth time to someone seeing it for the first. A nurse delivering test results that are, for her, a Tuesday afternoon briefing and, for the patient, a life-altering moment. A TSA agent repeating the same instruction hundreds of times a day to travelers who are anxious, disoriented, or simply having a terrible morning.

In every case, the interaction is deeply asymmetric. For one party, it is routine. For the other, it is anything but. And the professional who recognizes this asymmetry, who refuses to let their own familiarity flatten someone else’s experience, is the one who turns a transaction into a connection.

Professional excellence is not about pretending every moment is magical. It is about recognizing that someone else’s wonder often lives inside your routine and choosing to honor it.

A Practice, Not a Performance

I want to resist turning this into a tidy five-step framework, because the heart of the idea is more like a discipline than a technique. But there are practices that help, and they are worth naming.

Master the structure so it disappears. Know your material well enough that it no longer consumes your attention. Expertise is not about showing off how much you know. It is about freeing yourself to be human in the moment. A jazz musician who is still thinking about chord changes cannot listen to the room.

Vary the expression, not the essence. Change the stories, the metaphors, the examples, the pacing. Variation keeps you engaged with your own material, and engagement is contagious. Audiences do not catch your ideas. They catch your energy.

Play for the room you are in. Every audience has a rhythm. Some rooms are skeptical and need to be won over slowly. Some are eager and want you to match their pace. Energy is not volume. It is the ability to listen while you speak.

Perform for one. Before you begin, imagine one person in the audience for whom this moment truly matters. Perhaps it is the young professional who will rethink their career because of something you say. Perhaps it is the student in the back row who almost did not come today. That person becomes your anchor. You play for them.

Protect presence, not enthusiasm. You do not need to be “on” every second. You need to be genuinely present at the moments that count. One authentic pause, one honest aside, one real connection can outweigh ten minutes of polished delivery. Presence is not a performance. It is a choice to show up fully, even when the content is familiar.

The Musician’s Return

There is a concept in Indian classical music called riyaz, the daily practice through which a musician internalizes ragas so completely that performance becomes an act of expression rather than execution. The notes are known. The structure is given. But within that structure, the musician finds freedom, because mastery has made the scaffolding invisible.

I think about riyaz every time I step onto a stage. I know the slides. I know the arc. I know where the laughter tends to come and where the silence deepens. The audience does not. And that difference is not a burden. It is a gift. Because it means I get to watch someone encounter an idea for the first time, and if I have done my job, I get to feel the electricity of that encounter as though it were my first time too.

So I show up as a musician, not a machine. I play the same tune, but I listen for new notes. I wander, I riff, and I return to the theme. And in doing so, I rediscover why I love this work.

Because for someone in the room, this is not repetition.

It is a moment.

And moments deserve jazz.

I took a sip of my morning chai today, when the brainwave hit. I was reading overnight coverage of the 2026 Midterm elections in the US: a Chatham House brief on Hormuz, a New York Times piece on Trump’s latest Iran address, a Polymarket forecast showing Democrats at 51.5% for a sweep. The pundits were confident. The forecasters were confident. The markets were confident. And I realized that none of them were asking the question that actually mattered.

How might we forecast these midterms in a way that was genuinely unbiased and corrected for errors that polls and prediction markets keep making? And might we factor in events with no precedent, like the Iran conflict?

Subscribe now

The question had been nagging at me for weeks. Every polling average I’d seen felt like it was treating the world as if 2016, 2020, and 2024 hadn’t happened. Every prediction market I’d watched felt like it was hostage to whichever whale had the biggest position that afternoon. Every forecaster was reaching for historical analogs to a wartime midterm that has no clean analog. The Strait of Hormuz has been closed or restricted for nearly sixty days. Brent is at ninety-six. Gas is above four dollars. Yet, pundits are writing as if 1994 or 2006 tells us much.

The instruments felt wrong for the moment.

So at around 7 AM, masala tea in hand, I opened a conversation with Claude and asked a question I’d never asked an AI before. Not “predict the midterms for me.” Not “give me a summary of what the forecasters say.” But something closer to what I’d ask a brilliant graduate student who was about to become my research partner for the morning.

“Can we build a better one?”

The curious mind meets a brilliant collaborator

Let me say something upfront about what happened next, because I think it matters more than the model itself.

I have been a professor for thirty three years. I have mentored thousands of students, collaborated with dozens of co-authors, taught tens of thousands of executives. I know what intellectual partnership feels like, the texture of it, the push and pull, the moment when someone catches an error in your reasoning before you do. I used to think AI was tools. Useful tools, faster-than-Google tools, but tools.

What happened this morning was different. It was a genuine partnership. Claude pushed back when my first weighting scheme double-counted signals (I had approval and generic ballot both sitting at 30% without realizing how collinear they are). It volunteered an orthogonalization approach I hadn’t specified. It proposed a stress test when I hadn’t asked for one, and when that stress test broke the model, it offered a revision. I asked for a back test to recalibrate. I ruminated that the Strait of Hormuz situation had no precedent, so we needed a real-time Bayesian updating approach with 60 days of data, and a sophisticated causal chain from Presidential social posts to Iran’s reaction to WTI price to gas prices to consumer sentiment to poll implications. Claude built it in a flash. When I introduced a devil’s advocate challenge about prediction markets, Claude marshalled a defense rooted in the academic literature on market bias and then conceded, accurately, the three cases where markets genuinely dominate.

That isn’t a tool. That’s a colleague.

I want to describe what we built, because the model itself is interesting. But I want you to hold in mind what the process looked like, because this matters more.

Why polls and prediction markets fail

Before we get to what we built, let me outline the problems with conventional instruments.

Polls are systematically biased, and we know it. The American Association of Public Opinion Research has documented, in three separate post-election studies, that Republican and independent voters are less likely to respond to surveys than Democrats, and the ones who do respond are less likely to support Trump than the ones who refuse. Across 2016, 2020, and 2024, national polls underestimated Trump’s support by an average of 2.3 percentage points. This isn’t noise. It’s a structural error that pollsters have tried, and largely failed, to correct.

Prediction markets aggregate beliefs, not truth. I want to be fair to markets, because they have genuinely beaten polls in recent cycles. But markets have documented biases of their own. The favorite-longshot bias means unlikely outcomes are systematically overpriced. Concentration dynamics mean a single trader with thirty million dollars can move the price several points, which happened on Polymarket in October 2024. And markets are reactive. They price information after it arrives. They cannot tell you in advance which three variables would reshape a race.

Historical analogs fail in novel regimes. This is the deepest problem. The 2026 midterm is the first modern cycle to feature an active war, a closed strategic waterway, wartime stagflation, and a second-term president whose polling error pattern is both famous and partially attenuated (because he’s not on the ballot). There is no 1994 here. There is no 2006. The structural features of this moment are new enough that reaching for precedent is not rigor. It is laziness dressed as rigor.

The question I brought to Claude was whether we could build something that corrected for each of these failures explicitly. Not a better poll aggregator. Not a more sophisticated market-watcher. A meta-model that consumed the conventional instruments, acknowledged their biases, and added the one thing they cannot add: live causal structure.

Can AI build a meta-model?

That was the sharper version of the question, and the one that made my chai sit half-drunk.

A meta-model must synthesize signals that disagree with each other. It has to weight them by their actual predictive power, not their cultural prominence. It has to know which inputs are collinear and therefore shouldn’t be double-counted. It has to apply documented corrections to known biases. And it has to do all of this while remaining falsifiable, meaning every assumption has to be visible and revisable.

Claude and I spent the first fifteen minutes of the session on exactly this question. Not on the 2026 numbers. On the architecture. What are the true independent signal classes? What is each one’s historical predictive power? Where do they overlap, and how do we strip the overlap without losing information? What corrections do we apply to each, and why?

The answer we arrived at, which I named POLARIS-26, has four architectural layers. Let me walk through each, because the layering is where the sophistication lives.

Layer 1: Five orthogonalized signal pillars

The core insight is that generic ballot, presidential approval, and the economy are not independent. Approval absorbs the economy. Generic ballot absorbs both. Naive weighting double-counts roughly thirty-five percent of the underlying information. POLARIS treats generic ballot as the primary signal, weighted at forty percent, and weights approval and economy on their residuals only, meaning the portion of variation not already explained by generic ballot movement. This is the same logic Nate Silver uses. It matters more than most aggregators admit.

The final weights, after orthogonalization, are: generic ballot 40%, approval residual 15%, economic residual 10%, prediction markets 20%, geopolitical shock index 10%, fundraising momentum 5%. Every weight had to defend itself against a historical backtest.

Layer 2: The Polling Integrity Adjustment

This is the filter that corrects for the documented Trump-voter undercount. The critical calibration question is how much correction to apply. Too little and you miss 2024-style systematic error. Too much and you over-correct for a midterm where Trump isn’t on the ballot.

The literature gives us an anchor. The average Trump-era presidential polling bias is 2.3 points. The 2022 midterm bias, with Trump off the ballot, was closer to 1.0 point. POLARIS splits the difference with a 1.5-point Republican shift, carried as a distribution with an uncertainty band of 1.0 to 2.0. That uncertainty band is itself consequential. If the true midterm bias turns out to be 2.5 points, POLARIS’s Senate point estimate shifts from D 49 to D 48 and the probability of Democratic Senate control drops from 22% to 14%. The model expresses this uncertainty honestly rather than pretending it doesn’t exist.

Layer 3: The Swing Seat Gate

Most forecasters waste compute on races that aren’t races. A district with an 85% incumbent retention probability is not informative. POLARIS excludes every seat with baseline win probability above 80% or below 20% and runs full simulation only on the genuinely competitive seats: roughly forty-five House districts and ten Senate races. This is how you focus compute where it actually matters. The Senate gate currently includes Maine, North Carolina, the Ohio special, Iowa, Alaska, Georgia, Michigan, Minnesota, New Hampshire, and Kentucky. Every other seat is locked. Every gated seat gets the full probability treatment.

Layer 4: NovaWatch, the live candidate feed

The final architectural element is a live monitoring layer for candidate-specific events: indictments, sexual misconduct revelations, retirements, primary upsets, viral gaffes, fundraising shocks, district-specific economic shocks. Each category has a calibrated race-level impact range and a decay function. A Roy Moore–scale scandal moves a race by 8 to 18 points with a 45-day half-life. A retirement announcement moves it by 3 to 8 points, permanently. These events feed directly into race-level probabilities, bypassing the national environment index, because candidate events rarely move the national wave but routinely flip individual races.

Together, those four layers are the static architecture. But the most interesting thing we built that morning was dynamic.

The Hormuz Cascade: why causal beats analog

The Iran situation is the swing factor in this cycle, and it is the hardest thing to model because it has no precedent.

When I asked Claude how we should handle it, the first instinct was to reach for analogs. Iraq 2006. The Gulf of Tonkin. The 1973 oil embargo. Each has some structural feature in common with 2026, but none shares enough to be useful. The 1973 embargo wasn’t wartime. Iraq 2006 was a four-year-old war, already metabolized by voters. Hormuz has been closed for sixty days in a month when gas prices are setting records. We are in new territory.

So we did something I find genuinely interesting. Rather than force the present into the shape of the past, we built a live causal model that could estimate its own coefficients in real time.

The Hormuz Cascade is a six-link chain:

Iran events → oil prices → gasoline prices → consumer sentiment → presidential approval → generic ballot → seat outcomes.

Each link has a coefficient. The trick is that we populate those coefficients not from historical data but from the sixty days of observations we now have since the war began. We have seen major escalations move Brent by $3 to $8 intraday. We have seen ceasefire signals move it down by $5 to $12. We have seen gas prices pass through oil at a rate of $0.028 to $0.035 per gallon per dollar of Brent, which is fifteen to twenty percent higher than the historical baseline, probably because Gulf infrastructure damage has constrained refining capacity. We have seen Michigan sentiment drop six points for every fifty cents of sustained gas price increase, which is fifty percent steeper than the historical pattern.

These are not assumed numbers. These are observed numbers, fit to sixty days of post-war data and updated weekly as new observations arrive. The causal chain tightens as the data accumulates. Bayesian updating in real time.

And because we built it causally rather than analogically, we can do something genuinely useful: we can specify tripwires.

A tripwire is a pre-specified threshold that forces an automatic model re-run because it represents a potential change in the underlying structure, not a marginal move. POLARIS has three:

1. Brent above $110 sustained for ten trading days

2. Cumulative Hormuz disruption exceeding 90 days

3. A single-event US casualty count above 30

Each tripwire has a pre-calculated impact vector. We don’t wait to see what analysts say. The model updates automatically when the trigger fires. This is the discipline causal modeling enforces that analog-based forecasting cannot.

The stress tests that almost broke the model

Building the architecture took about twenty minutes. The next ten were the ones that made me respect the collaboration.

Claude proposed six stress tests, unprompted. Three of them broke the initial model.

The first was a collinearity check. Generic ballot, approval, and the economy were all weighted at 25% or higher, which double-counted their shared variance. The fix was orthogonalization, which I described above.

The second was a backtest against 2010, 2018, and 2022. The model performed well on 2010 and 2018, where the generic ballot was running above five points. It performed poorly on 2022, where the generic ballot was close to zero. This surfaced an important truth: the generic ballot is most predictive when it’s clearly above or below the noise band, and less reliable inside it. Today’s D+5.6 reading is comfortably outside the noise band, which gives us some confidence. But we flagged this as a condition to monitor.

The third was a sensitivity analysis on the Polling Integrity Adjustment. At a PIA of 0.5, the model predicted a Democratic sweep. At 2.5, it predicted a Republican hold of both chambers. The model is more sensitive to that single parameter than to any other. The fix was to carry the PIA as a distribution rather than a point estimate, and to be transparent about how much the prediction depends on it.

By the end of the stress-test round, the model looked different from the first draft. That is what testing is supposed to do.

The prediction that came out the other side

After all of this, with the environment orthogonalized and the corrections applied and the causal cascade populated, POLARIS-26 produced its first run. Here is what it says as of April 20, 2026:

The House flips to the Democrats. Point estimate: D 225, R 210. Eighty percent confidence interval: D 217 to 233. Democratic control probability: 73%.

The Senate holds for the Republicans, narrowly. Point estimate: R 51, D 49. Democratic control probability: 22%. Probability of a 50-50 tie: 19%.

The most likely joint outcome is divided government, at 51% probability. Democratic sweep at 22%. Republican status quo at 23%. All other scenarios at 4%.

For context, Polymarket is pricing a Democratic sweep at 51.5% today. POLARIS is two points lower than the market on the House and eight points lower on the Senate. That divergence is the model’s contribution. It comes almost entirely from the PIA correction, which markets structurally cannot apply because they don’t decompose polling error by cycle type.

If the Senate result lands inside POLARIS’s D 46 to 52 range on November 3, the model will have earned its keep. If not, markets were right and I owe a public revision.

The dashboard, and the 45-minute miracle

Here is the part I’m still struggling to process.

7:00 AM — I ask the question. We sketch the architecture on the fly.

7:15 AM — Signal pillars defined, weights derived, orthogonalization applied.

7:20 AM — Polling Integrity Adjustment calibrated against 2016-2024 literature.

7:25 AM — Hormuz Cascade structured as a six-link causal chain with live coefficients.

7:30 AM — Stress tests run. Three revisions made.

7:35 AM — First prediction produced with point estimates, confidence intervals, and joint probability table.

7:40 AM — Committed the methodology to a reusable skill, so future runs follow the exact same protocol.

7:45 AM — Interactive React dashboard built. Sliders for every signal. Live Brent cascade. Scenario buttons. Tripwire indicators. Fonts chosen, colors set, responsive layout tested.

8:30 AM — This article was written and posted!

A working midterm forecasting model. Architecturally sophisticated, methodologically honest, empirically testable, with its own interactive dashboard. One hour from idea to publication.

I want to be careful not to overclaim. This is a model, not an oracle. It could be wrong. I’ve been explicit about where it could be wrong and how I’d know. But the fact that we went from question to falsifiable production artifact in an hour is, I think, genuinely a new thing in the world.

I’ll tell you what I find most moving about the experience. Not the speed, although the speed is astonishing. Not the output, although the output is good. What I find moving is the quality of the collaboration. Claude didn’t just execute my instructions. It proposed architecture I wouldn’t have thought of. It caught errors I would have missed. It pushed back on weak reasoning. It held the thread across ninety minutes of technical conversation without losing the narrative. It committed the methodology to a reusable skill so I can run this model again next month with a single command.

That is not tool use. That is colleague-scale partnership, available to anyone with a laptop and a question.

The interactive artifact

By tomorrow (and another cuppa tea), I will build an interactive dashboard so you can play with POLARIS-26 live. You will be able to:

• Move any signal slider and watch the seat counts update in real time.

• Toggle the Polling Integrity Adjustment on or off to see how much it moves the Senate probability. (Spoiler: a lot.)

• Push the Brent oil slider past $110 and watch the Hormuz tripwire fire red.

• Click any of the four preset scenarios (Baseline, Democratic Wave, Republican Recovery, Hormuz Break) and see the model jump to that world.

• Watch the joint probability matrix recalculate.

The code is open. The math is visible. Every assumption is revisable.

This is what I want more of in the world. Models that are interactive rather than opaque. Methodologies that are visible rather than hidden. Collaborations with AI that feel like colleagues rather than tools.

Reflections on what just happened

I started this Monday morning with a question and a cooling cup of tea. By 7:45 AM I had an architecturally sophisticated, empirically testable, interactively explorable political forecasting model, a committed reusable skill for future runs, a static dashboard, and a clear set of falsifiable predictions with a November resolution date.

This is not normal productivity. That is a new kind of intellectual leverage. A curious mind plus a capable AI collaborator, asking good questions together, iterating in real time, stress-testing each other’s thinking, and producing something that neither could have produced alone at anything close to that speed.

My seventh decade is going to be more interesting than I expected.

Gotta love Claude.

Mohanbir Sawhney is the McCormick Foundation Professor of Technology at the Kellogg School of Management. He writes about AI, strategy, and the interior life of modern work

I often get asked a question about AI - isn’t it the great equalizer for competence? Everyone now has access to the same frontier models. A first-generation college student in rural India can use the same Claude or GPT that a McKinsey partner uses. The playing field, the argument goes, has been leveled. This argument seems logical. But it is wrong. And believing it is one of the most dangerous mistakes a knowledge worker can make today.

Every wave of technology disruption generates the same leveling narrative. The internet was supposed to democratize commerce. Social media was supposed to democratize influence. MOOCs were supposed to democratize education. In every case, the tools became universal while the advantages did not. What happened instead was a pattern: the floor rose for everyone, the ceiling rose faster for the few, and the middle got compressed.

AI is Microsoft Word at civilizational scale.

When word processors arrived, everyone had the same tool. Having access to Word did not make anyone Shakespeare. The tool was universal. The talent was not. What happened was predictable: the minimum acceptable quality of a written document rose sharply, while the distance between competent and exceptional grew wider, not narrower. AI is doing the same thing, but across every domain of knowledge work, simultaneously, and at a pace that leaves no time for gradual adjustment.

The floor has been raised dramatically, suddenly, and for almost everyone. The ceiling has also risen, but only for those who already had the height to reach it. The distance between the floor and the ceiling has not shrunk. It has expanded.

The Hollowing of the Middle

For decades, organizations built massive throngs of credentialed middle knowledge workers: analysts, associates, coordinators, junior managers, and specialists of every variety. These people turned senior judgment into structured output. They synthesized research, built models, prepared presentations, drafted communications, and managed the operational metabolism of large enterprises. They were the connective tissue of the knowledge economy.

AI narrows that gap aggressively. The tasks that defined middle-tier knowledge work, synthesis, summarization, structured analysis, first-draft production, research compilation, presentation building, are precisely the tasks that AI now performs at or above the level of a competent junior professional. Not in the future. Today.

I see this in my own classroom. Five years ago, the students who excelled at case analysis were the ones who could grind through data, build clean spreadsheets, and produce well-structured slide decks. Those students had an edge because execution was hard. Today, execution is table stakes. AI handles it. The students who stand out now are those who ask better questions, reframe problems in surprising ways, and exercise judgment that the model cannot. The bar for differentiation has migrated upward, and it did so in about eighteen months.

Consider the product manager. Five years ago, a solid PM could differentiate herself by writing crisp user requirements, synthesizing user research into clear insight summaries, building competitive landscapes, and structuring sprint backlogs with care. That was the job. Those were the skills that built a career. Today, any PM with a Claude subscription can produce a first draft of a requirements document in ten minutes that would have taken two days to write manually. Competitive landscapes, feature comparison matrices, user journey maps: AI generates all of these at a quality level that meets or exceeds what most mid-career PMs produce on their own.

So what separates the exceptional PM from the rest? Not the ability to produce artifacts. The ability to decide which product to build and why. The ability to see a market signal that the data does not yet confirm. The ability to say no to the feature that customers are asking for because you understand the job they are actually hiring the product to do, and that job points somewhere else entirely. The ability to hold a room of engineers and executives in a prioritization debate and make a call that balances technical debt, business model economics, competitive timing, and customer psychology, all at once, in real time, with incomplete information. AI cannot do that. AI will not do that for a very long time. But the PM who cannot do that either is now competing against a machine for the tasks she used to own.

The middle is caught in a compression. The floor has risen to meet them from below. The ceiling has pulled away from them above. The comfortable plateau of “competent and credentialed” is disappearing.

The Wedge That Must Not Close

To understand what separates those who will thrive from those who will merely survive, consider a simple mental model: the wedge.

Imagine two lines on a graph. One line represents the advancing capability of AI: its ability to produce high-quality knowledge work output. That line rises steeply and without foreseeable limit. The second line represents the unique capability of a given human professional: the judgment, taste, creativity, contextual awareness, and synthesis ability that the person brings above and beyond what AI can produce. The vertical distance between these two lines at any given moment is the wedge. It is the person’s margin of relevance.

For most people in most jobs, that wedge is narrowing. AI’s capability line is ascending faster than human capability lines. The wedge closes from above.

The only sustainable response is to move the second line upward faster than the first line rises. Not by doing the same things better, but by consistently relocating to the frontier: to the tasks, questions, and forms of judgment that AI cannot yet perform. This is not a one-time migration. It is a permanent posture. The frontier is not a destination. It is a direction.

What does the frontier look like? It is not a fixed set of tasks. It is a set of characteristics. Frontier work is ambiguous: the problem is not well-defined, and framing it correctly is itself the value. Frontier work is integrative: it requires combining insights across domains that AI treats separately. Frontier work is high-stakes: the consequences of getting it wrong are significant, and no one is willing to delegate the decision to a machine. Frontier work is relational: it depends on trust, persuasion, negotiation, and the kind of contextual reading that emerges only from human interaction.

The knowledge workers who thrive will be those who stay perpetually ahead of the closing wedge, who treat AI not as a tool to do their current job faster, but as a displacement force that continuously redefines what their job must become.

What Raising the Ceiling Actually Looks Like

Let me make this concrete with an example from my own work, because abstraction is the enemy of action here.

I write business case studies. I have been doing it for thirty-five years, and the methodology I have developed is specific, opinionated, and built on thousands of hours of classroom testing. Could I ask Claude to “write me a business case study”? Of course. And it would produce something that looks like a case study. It would have a protagonist, a company context, some decision points. It would be competent. It would also be mediocre: generic structure, predictable analysis, no pedagogical design, no narrative tension.

So instead of using AI as a replacement for my judgment, I did something different. I carefully encoded my entire case writing methodology into a structured set of instructions for Claude: what makes a good case protagonist, how to create decision tension, how to structure exhibits, how to calibrate complexity for different classroom contexts. And I gave Claude examples of award-winning cases, and the ones that didn’t sell as well. I taught Claude my style, my tone, my brand guidelines, and my voice. And I encoded all this in a Claude Skill.

With that in place, Claude does not produce generic case studies anymore. It produces case studies that embody my methodology. The floor for a first draft rose dramatically. But here is the critical point: the ceiling rose even more, because I now spend my time on the work that only I can do: selecting the right company, identifying the non-obvious strategic tension, shaping the narrative arc, pressure-testing the teaching plan. The AI handles the execution. I invest in the judgment. I can now produce a polished and insightful case study in two days. Down from six months.

A great chef follows consistent technique: proper mise en place, correct knife cuts, precise heat control. They follow a structured method. The structure makes the dish consistent. But it also frees them up to infuse creativity . The paradox - structure enables creativity. It does not constrain it. The same is true for me as a “case chef”. Before I build my system, I spent so much cognitive energy on execution that I had less bandwidth for originality. Now I have more.

The AI did not replace my expertise. It amplified it. And the amplification is proportional to the expertise I brought to the table in the first place.

Two Gaps, One Fate

The capacity to migrate upward is not evenly distributed, and this is the uncomfortable truth.

There are two distinct gaps operating simultaneously in the AI economy. The first is a capability gap: the difference in what people are able to do with AI based on what they already know. The second is a fluency gap: the difference in what people are able to do with AI based on how well they know how to work with it. These gaps are different in kind, different in consequence, and different in what it takes to close them.

The capability gap is perhaps the less discussed but more consequential of the two. AI is a multiplier. It amplifies whatever stock of judgment, taste, domain expertise, and synthesis ability a person already has. Give the same AI tool to a thirty-year veteran of product strategy and to a freshly minted MBA, and the outputs will differ enormously, not because of the tool, but because of what each person brings to the collaboration. The veteran knows which questions to ask, which outputs to reject, which subtle signals in the data matter and which are noise. The rookie MBA does not. Not yet. The AI amplifies the gap between them rather than closing it.

If I hand my case writing methodology to a student who has never written a case study, they will get a dramatically better first draft than they would have gotten without it. The floor rises. But the distance between their output and mine will increase, because I am working with the same amplifier on top of a much deeper base of knowledge and pattern recognition.

The fluency gap is different. It is not about depth of domain knowledge. It is about a cognitive posture shift that some people make and others do not. The knowledge workers who genuinely understand AI treat it as collaborative intelligence. They iterate with it. They challenge its output. They give it context, constraints, and examples. They think of prompting as a form of creative direction, not a form of search. They build on AI’s output rather than accepting or rejecting it wholesale.

The people who have not made this shift interact with AI the way they interact with Google: type a question, get an answer, done. They prompt AI like they would prompt a junior analyst: give it a task, receive a deliverable, move on. They do not iterate, refine, co-create, or push back. The result is that two people with the same domain expertise can get wildly different outputs from the same AI, simply because one has learned to collaborate with the machine and the other has not.

These are not technical skills. They are mindset shifts, and they are teachable. The people who have not made them are not less intelligent. They are often simply less willing to abandon a mode of working that served them well for decades. That reluctance is human. It is also increasingly expensive.

The Strategic Imperative

For the individual knowledge worker, the implications are clear but not comfortable.

Survival requires closing the fluency gap immediately and without equivocation. AI avoidance is no longer a viable professional strategy. Those who refuse to engage with AI are not making a principled stand. They are falling behind in a race they have not yet realized they are running.

Bridging the capability gap is more difficult, and it demands self-awareness. The relevant question is not “am I using AI?” but “what is the quality of judgment I am bringing above AI’s output?” and “is that judgment appreciating or depreciating in value as AI improves?” If your primary value-add is synthesis that AI can now do, you do not have a moat. You have a memory.

The specific discipline of thriving is frontier migration. It means deliberately and continuously moving toward ambiguous problems, novel synthesis, and high-stakes judgment. It means building attribution: a body of work, a methodology, a point of view, a network of trust that is identifiably yours and cannot be replicated by a model trained on the collective average.

For organizations, the imperative is to resist the obvious but limited play of deploying AI purely for cost reduction, in favor of the more difficult but durable play: using AI to shift the composition of work toward higher-value activities. The companies that use AI only to eliminate headcount will save money in the short term and lose capability in the long term. The companies that use AI to move their people uphill, to free human judgment for the work that actually creates differentiation, will build organizations that are genuinely difficult to compete with.

The Tide is Rising - Move Uphill

The rise of AI capability is a tide. It is rising for everyone simultaneously, and it lifts certain boats spectacularly. But a rising tide also submerges everything that is not elevated enough to stay above the waterline. The tasks, roles, and competencies that sit just above the current water level will be underwater within months, not years.

Shakespeare did not become irrelevant when the printing press democratized access to text. He became more valuable, because the multiplication of words made the rarity of genuine literary judgment and creative brilliance more visible, not less. The printing press did not level the playing field between Shakespeare and the average pamphleteer. It widened the gap between them permanently.

AI will do the same to every domain of knowledge work. The question is not whether you have access to the tool. Everyone does. The question is what you bring to the tool that it cannot bring to itself.

The tide does not wait for you to learn to swim. It does not care about your credentials, your title, or your years of experience. It rises. The only question is whether you are moving uphill faster than the water.

This is the final installment in a three-part series on skills, jobs, and learning in the age of AI. Part 1, “AI Isn’t Taking Jobs. It’s Taking the Ability to Learn,” explored how AI disrupts not just work but the learning process that builds expertise. Part 2, “Philosophy, Plumbing, and Where the Jobs Will Be,” mapped the skills and fields that endure. This piece is for parents. It is the most personal of the three, and in some ways, the most important.

I am a parent. Like you, I want two things for my children. I want them to succeed now: good grades, strong test scores, admission to a respected university, the credentials that open doors. And I want them to be capable for life: resilient, adaptive, able to think independently and navigate an uncertain world.

These two goals have been somewhat conflicting. But AI has deepened the tension between these two goals. Today, your child can use AI to produce a flawless college essay, ace a homework assignment, generate a research paper with perfect citations, and assemble a portfolio that looks like the work of a gifted student. The transcript will be stellar. The admissions committee will be impressed. And underneath it all, the cognitive muscles that your child really needs to build do not get used and will slowly atrophy.

This is the parental dilemma of our times. You need to decide which game you want to play. Are you optimizing for your child’s GPA this semester, or for their cognitive capability over a lifetime? These two objectives are increasingly in conflict, and AI is driving a wedge between them.

The Efficiency Trap

We live in a culture that worships efficiency and treats struggle as a malfunction. When a student finds a homework assignment difficult, the instinct, for both the student and the parent, is to remove the difficulty. Google it. Watch a YouTube tutorial. Ask ChatGPT. Get the answer and move on.

This instinct feels rational. Why should my child spend two hours struggling with a calculus problem when AI can explain the solution in thirty seconds? Why should they agonize over an essay when a tool can produce a polished draft instantly? The logic of efficiency says: eliminate the friction, get the result, move to the next thing.

But here is what the logic of efficiency misses: the friction is the learning. The two hours of struggle with the calculus problem is where mathematical intuition gets built. The agony of the essay is where your child discovers what they actually think. The confusion, the dead ends, the moments of being stuck: these are not obstacles to learning. They are the mechanism of learning. Remove them, and you have removed the thing that produces a capable mind.

The analogy I keep returning to is exercise. “Why should I walk when I can drive?” feels perfectly rational, until you realize the walking was building your cardiovascular health and the driving is slowly weakening it. The effort was not the cost of getting somewhere. The effort was the benefit. AI, used carelessly, is the intellectual equivalent of driving everywhere. The destination looks the same. The body underneath is atrophying.

Parents must realize that they may be unconsciously encouraging the atrophy. Every time you say “just ask ChatGPT,” every time you smooth a path that was supposed to be rough, every time you prioritize the grade over the growth, you are trading long-term capability for short-term comfort. Your child will pay the price, not now, but in ten years, when they are sitting across from a problem that requires real judgment and they have never built the muscle to exercise it.

Subscribe now

What Parents Must Do

I have given you laudatory principles, but principles are useless without practical guidance. Here is my specific advice on “protecting the struggle”:

Protect the struggle window. There is a critical developmental period, roughly ages 12 to 22, where the cognitive foundations of expertise are being built. During this window, productive difficulty is the raw material for learning. This does not mean banning AI. It means sequencing it wisely. There are three stages in using AI wisely: Build, Spar, Orchestrate. Build means doing the foundational cognitive work yourself. Spar means using AI to challenge and pressure-test your thinking. Orchestrate means delegating with earned authority. The first stage must come first. There are no shortcuts, and parents are the ones who enforce the sequence when every other force in a teenager’s life is pushing them to skip ahead.

Change what you celebrate. If you praise your kids’ grades, you are rewarding the output. Your child hears: the result is what matters, by any means necessary. And AI is now the most efficient means available. Instead, try asking: “What stumped you this week? What did you struggle with? What did you get wrong and what did you learn from it?” Children internalize what their parents signal as valuable. If the implicit message is that struggle is failure, they will avoid it. If the message is that struggle is where growth happens, they will seek it out. This shift may seem subtle, but it may be the most consequential shift you can make as a parent.

Model learning yourself. Children learn more from watching than from listening. If your own relationship with AI is purely delegatory, if you ask it for every answer and never wrestle with a hard problem yourself, your child absorbs that pattern. But if you visibly engage in hard learning, reading a difficult book and talking about it, taking on an unfamiliar challenge, admitting what you do not know and then working to figure it out, you create a culture where intellectual effort is respected. The most powerful curriculum is not what you assign your child. It is what they see you do.

Emphasize capability, not credentials. Unfortunately, the college admissions system still rewards polished outputs, and AI makes polished outputs trivially easy to produce. Your child knows this. Their friends are using AI for applications, essays, and projects. Pretending otherwise is naive. The conversation you need to have is not “don’t use AI” but “what are you actually building?” Help them see the difference between a credential and a capability. Yes, the transcript matters. No, it is not the point. The point is to become someone who can do hard things, and the transcript should be evidence of that, not a substitute for it. This is a values conversation, not a rules conversation. Rules get circumvented. Values get internalized.

Rethink the status hierarchy. If you read Part 2 of this series, you know I argued that skilled trades are a smart career bet in an AI world. But here is the uncomfortable truth for many parents: you may agree with that argument intellectually while sending very different signals to your child. Children detect status signaling instantly. If you say, “trades are respectable” but your body language says “I would be disappointed if you became a plumber or a chef,” they will read the body language. If you want your child to pursue skill security, you need to respect the paths that offer it. That means talking about electricians and nurses with the same admiration you give consultants and software engineers. Not as a performance. As a conviction.

Teach them to be editors, not consumers. This may be the single most practical habit you can cultivate. Whenever your child uses AI for anything, ask them to critique the output. What did it get wrong? What would you change? What is missing? What does it assume that you would not? This one practice, treating AI output as a first draft to be improved rather than a final answer to be accepted, builds the critical judgment that separates the people who lead AI from the people who are led by it. It takes thirty seconds. It transforms the relationship with technology from passive consumption to active engagement.

Give them real responsibility with real consequences. Chores, part-time jobs, projects where their decisions matter and the outcomes are not simulated. The teenager who manages a small budget, runs a lawn-mowing operation, or volunteers where people depend on them is building judgment in a way that no classroom exercise and no AI tool can replicate. What matters here is not the specific activity. It is the experience of ownership: making decisions, living with the consequences, and learning from what went wrong. AI cannot give your child this experience. Only life can. And only you can make sure they encounter it.

The Hardest Part

I have given you a list of practical things to do. But I want to be honest about something: the hardest part of this is not knowing what to do. It is having the emotional fortitude to do it.

Watching your child struggle is painful. Every parental instinct says: help them. Fix it. Make it easier. And AI has made “making it easier” frictionless. The answer is always one prompt away. The polished output is always available. Saying no to that, or more precisely, saying “not yet,” requires a kind of faith that is genuinely difficult to sustain.

It is the faith that difficulty is not cruelty. That the answer you refuse to give them is a gift. That the frustration they feel tonight is building something inside them that will matter for the rest of their lives. Every wisdom tradition understands this. In Sikh philosophy, the path of discipline and devoted practice (“Naam Japna”) is not a punishment; it is the mechanism through which character is forged. Zen training places the student before the koan, a riddle that defeats conventional logic, not because the teacher enjoys watching the student suffer, but because the struggle itself is the curriculum. The koan cannot be outsourced. The transformation happens only through the direct encounter with difficulty.

Parenting in an AI age requires the same faith. Not blind faith. Informed faith. You are not withholding help to be cruel. You are protecting the process that builds a capable, independent, thinking human being. You are ensuring that when your child eventually picks up the most powerful cognitive lever in human history, they have something solid to stand on.

The Ground Beneath Their Feet

Yes, you must be the wind beneath the wings of your children. But more importantly, you must be the ground beneath their feet. Throughout this series, I have returned to Archimedes: “Give me a place to stand and a lever long enough, and I will move the world.” The lever is magnificent. It is more powerful than anything I have seen in my long career.

But the lever is not your job. The world will hand your child the lever. It is already handing it to them. Every app, every tool, every classroom is integrating AI at an accelerating pace. The lever will take care of itself.

Your job is the ground.

Your job is to help them build a place to stand: the judgment, the resilience, the pattern recognition, the first-principles thinking, the ethical clarity that comes only from years of doing hard things. Your job is to protect the struggle that builds that ground, even when it would be easier for both of you to skip it. Your job is to believe, with conviction, that the frustration your child feels today is not a problem to be solved but a foundation being laid.

Give them a place to stand. The lever will take care of itself.

* * *

Here is a truth that took me decades to fully appreciate: the quality of your life depends on many things, but few of them matter as much as the quality of the people in it. The right people expand you. They challenge your thinking, lift your energy, and bring out a version of you that you like. The wrong people do the opposite. They slowly erode you in ways you do not notice until you step back and wonder why you feel so drained.

As I get older, I’ve become far more intentional about who I give my time to. Not out of arrogance, but out of awareness. Life is finite, and energy is finite, and every hour spent with someone who leaves you depleted is an hour you could have given to someone who leaves you inspired.

The problem is that most of us make these decisions on vibes and intuition. We tolerate relationships that quietly cost us because they seem “fine.” We keep investing in people who have been running a deficit with us for years, confusing familiarity with value and proximity with friendship.

I want to offer a framework that you can use to choose who gets to be in the front row of your life. I propose three ratios that, taken together, give you a clear-eyed way to evaluate any relationship in your life.

Ratio 1: Saying to Doing Ratio

I work with a lot of entrepreneurs, and over the years I’ve noticed a recurring pattern. There is a type of person who speaks fluently in strategy, vision, and big ideas, the kind of person who is captivating in a room and can paint a picture of the future that makes you want to lean in. But when you circle back six months later and ask what got done, the cupboard is bare. The grand vision hasn’t moved an inch.

I like to say that strategy is 10% vision and 90% execution, and the ratio of what someone says they’ll do and what they do tells you a lot about their character. Think about your last several interactions with someone who matters to you. Were there positive surprises, moments where they went beyond what was expected, took initiative without being asked, showed real ownership over a shared problem? Or was there a pattern of commitments that evaporated, promises that were made with conviction and forgotten with ease?

A small story. I once asked my housekeeper to organize my closet. When I came home that evening, she had sorted my shirts by color, arranged my jackets by how frequently I wear them, and grouped my suits by season. I didn’t ask for any of that. She didn’t just follow instructions; she took ownership of what “organized” meant for me and delivered something I hadn’t even thought to request.

That is what a high saying-to-doing ratio looks like in practice. She walks the talk. Most people, I’m afraid, just stumble the mumble.

Ratio 2: Taking to Giving Ratio

Every relationship has a ledger, whether we acknowledge it or not. I encourage you to try a simple accounting exercise. Think of someone important in your life, a friend, a sibling, a colleague, a partner, and replay the last ten interactions you’ve had with them. Of those ten, how many were gives and how many were asks? How often did they reach out because they needed something from you, and how often did they reach out simply to give, to check in, to be present?

I have a colleague whose last words on every phone call are the same: “Is there anything I can do for you?” It’s not a throwaway line and it’s not performative. He means it every single time. That one small habit tells me everything about his orientation toward the world. He leads with contribution rather than extraction and being around him makes you want to do the same.

The deeper point is that giving and taking is a mindset, not a transaction. It’s not about who picks up the check or who sends a nicer birthday gift. It’s the difference between someone who is genuinely other-centered and someone who is fundamentally self-centered, between someone who shows up when you need them and someone who is conveniently unavailable, between someone who remembers your struggles and someone who only remembers your usefulness. Once you start paying attention to this ratio, you see it everywhere.

Ratio 3: Negativity to Positivity Ratio

You already know the difference between these two kinds of people, even if you’ve never had a name for it. I think of them as radiators and drains.

Radiators bring warmth, empathy, humor, and a glass-half-full energy to every room they enter. After spending time with them, you feel lifted, as though someone quietly recharged a battery you didn’t realize was low. Drains bring gossip, criticism, and a persistent undertone of pessimism. After spending time with them, you feel smaller, as though something has been subtracted from you that you can’t quite name.

Now, this is not about toxic positivity or expecting everyone to be cheerful all the time. Life is hard, and real friends sit with you in the hard parts without flinching. What I’m talking about is net energy over the long run. Does this person, on balance, add to your emotional, cognitive, and spiritual well-being? Or do they quietly deplete all three?

If you consistently feel drained after spending time with someone, that is data. Treat it with kindness but treat it seriously.

Putting It Together

When you stack these three ratios on top of each other, you get a simple but powerful diagnostic for any relationship in your life. Saying-to-doing tells you whether someone delivers or just declares. Taking-to-giving tells you whether they contribute or just consume. Negativity-to-positivity tells you whether they radiate or drain.

Someone who scores well on all three belongs in your front row, close to you, worth investing in deeply. Someone who consistently falls short on all three may deserve your compassion but probably needs to sit further back. Most people, of course, fall somewhere in between, which is exactly why the framework is useful. It replaces a vague sense of unease with a specific and honest assessment.

I want to be clear: this isn’t about cutting people out of your life with a cold calculus. It’s about being thoughtful with the most precious resource you have, which is your presence. Some relationships deserve deep investment. Others deserve warmth and well-wishing, but from a greater distance.

The Mirror Test

And here is the part that most people skip, the part that matters most.

This framework applies to everyone, including you. In fact, you are the most important person to run through these three ratios. Before you start evaluating the people around you, take an honest look in the mirror and ask yourself the same questions. Are you delivering more than you promise? Are you giving more than you take? Are you adding warmth to the rooms you walk into, or are you the one doing the draining?

The best way to attract good people into your life has always been the simplest: become someone worth being around. As the old saying goes, the best way to have a friend is to be one.

Improve your own ratios first, and the rest will follow.

This is the second in a three-part series on skills, jobs, and learning in the age of AI. Part 1, “AI Isn’t Taking Jobs. It’s Taking the Ability to Learn,” explored how AI disrupts not just work but the learning process that builds expertise. If you haven’t read it, I’d encourage you to start there. This piece gets concrete: what should students actually study, and where will the jobs be? Part 3, “Advice for Parents: Protect the Struggle,” will address how to raise capable humans when the easy path is always available.

* * *

Students heading to college ask me an urgent question: What should I major in?

It is the wrong question. Majors are institutional categories. They describe how universities organize departments, not how the world organizes value. Nobody hires a “political science major.” They hire someone who can analyze complex systems, write with clarity, and make a persuasive case under pressure. The major is just the container. The capabilities are the content.

In an AI world, this distinction matters more than ever. The right question is not “what should I major in?” but “what capabilities will remain valuable and become more valuable as AI gets better?” Job titles come and go. Skill portfolios are timeless. The student who builds the right portfolio will have options that no single major can guarantee. The student who picks a “safe” major without building underlying capabilities will discover that no major is safe.

In Part 1, I argued that AI is the most powerful cognitive lever in human history, but a lever without a fulcrum is just a stick. The fulcrum is human judgment, pattern recognition, and first-principles thinking, built through years of struggle and practice. This piece is about what that fulcrum is made of. What, specifically, should you study and build?

The answer involves more philosophy and plumbing than you might expect, and less coding than conventional wisdom suggests.

Think Skill Security, Not Job Security

Before I get to specific fields, let me introduce a concept that should reframe how you think about career preparation: skill security.

Job security means you hold the same position for a long time. It is a relic of an era when industries moved slowly and institutional loyalty ran in both directions. That era is over, and AI is accelerating its end. Entire job categories will be created and destroyed within your career span.

Skill security is different. It means you have mastered capabilities that will be valued regardless of which jobs exist. The specific role changes. The underlying abilities transfer. A person with strong analytical reasoning, clear communication, and the ability to orchestrate complex projects will find work in industries that do not yet exist, doing jobs that have not yet been named. That is skill security. It is the only kind worth pursuing.

So what are the skills that endure? I see five meta-capabilities that every student should be building, regardless of what major they choose.

1. Pattern recognition. The ability to see structural similarities between seemingly unrelated problems. This is what makes a great strategist, diagnostician, or investor. It is trained by exposure to breadth, not just depth. The student who studies history, economics, and biology will see patterns that the student locked into a single discipline will miss.

2. Judgment under uncertainty. Knowing when the data is sufficient, when to act despite ambiguity, when to trust the model and when to overrule it. AI can process information. It cannot take responsibility for a decision when the information is incomplete. And when decisions have serious consequences. That requires a human who has been wrong enough times to develop calibrated intuition.

3. Orchestration. Designing and managing systems where humans, AI agents, and processes work together. This is the skill of the conductor, not the violinist virtuoso. Whether you are orchestrating a marketing campaign, a construction project, or a clinical trial, the ability to see the whole system and coordinate its parts is increasingly the most valuable thing a professional does.

4. Communication and persuasion. The ability to translate complexity into clarity, to move people to action, to build consensus across conflicting interests. AI can draft a memo. It cannot own a relationship, stand behind a recommendation with its reputation on the line, or read a room full of skeptical executives. These are irreducibly human skills.

5. Ethical reasoning. As AI systems gain autonomy, someone has to set boundaries, define what “good” looks like, and take responsibility when things go wrong. This is governance, safety, and values-driven leadership. These are skills that philosophy and liberal arts teach you.

Fields That Will Win

With those meta-capabilities as the frame, let me walk through specific fields. I am going to be opinionated here, because my readers expect me to shoot straight, and not hedge.

Liberal arts. I will say it plainly: the liberal arts are the single best training ground for the cognitive skills AI cannot replicate. This is not the defensive “liberal arts still matter” argument that embattled humanities departments trot out at fundraising dinners. This is the offensive argument. Philosophy trains you to detect bad logic, and in a world flooded with AI-generated plausibility, detecting bad logic is a survival skill. History trains you to recognize patterns across contexts, which is the foundation of strategic thinking. Literature builds empathy and narrative skill, which are the foundations of leadership and persuasion. These are not soft skills. They are the hardest skills to automate and the hardest to acquire.

In a world drowning in AI-generated content, the scarce resource is not production. It is taste, editorial judgment, and the ability to say something worth saying. Liberal arts build these muscles.

Economics and mathematics. Economics is the grammar of incentives, tradeoffs, and systems. It teaches you to think about second-order effects, unintended consequences, and equilibrium dynamics. Every AI deployment decision is fundamentally an economics problem: cost-benefit under uncertainty, principal-agent tensions, market design. Economics also bridges quantitative and qualitative reasoning in a way few disciplines do.

Mathematics, particularly statistics, probability, and optimization, gives you the conceptual foundation to be a credible participant in any technical conversation without necessarily being the one writing the code. You do not need to build the model. You need to know how the model works, and when the model is lying to you. Mathematical fluency is the difference between being a consumer of AI and being its architect.

Computer science, redefined. AI writes code now, and it will write it better next year. Coding may be dying, but computer science lives. Computer science as a discipline of systems thinking, abstraction, and architecture remains powerful. The question is whether you are learning CS to be a coder or to be a systems architect who understands how software, data, and AI compose into larger systems. The latter is deeply durable. What is dying is the middle tier of implementation work. What is thriving is the ability to design, specify, evaluate, and govern complex technical systems. CS as vocational coding training is finished. CS as computational thinking is alive and well.

Design thinking. Design frames problems before solving them, by understanding context and constraints, prototyping and iterating. AI can generate a thousand options in seconds. A human has to decide which ones are worth pursuing. Whether it is product design, service design, organizational design, or policy design, this is the skill of the architect: setting the criteria, evaluating the options, and taking responsibility for the choices. AI is a spectacular option generator. It is a terrible judge. Judgment is where humans earn their keep.

Cognitive science. This is the sleeper field that few people talk about. As AI systems become more capable, the people who understand how humans think, perceive, decide, and err will be disproportionately valuable. Human-AI teaming, AI safety, user experience design, behavioral product design: all of these require deep understanding of cognition. If you want to build AI systems that actually work for humans, you need to understand how humans work.

Plumbing, Welding, and the Smartest Career Bet Nobody Is Making

Now let me make an argument that will surprise some readers and feel obvious to others. One of the smartest career moves a young person can make today is to learn a skilled trade.

The economic logic is solid. Start with what robotics researchers call Moravec’s Paradox: tasks that are easy for humans, like navigating a cluttered basement, diagnosing a strange rattle in an HVAC system, or running plumbing through a 90-year-old building with no two walls the same, are extraordinarily difficult for machines. High variability physical work in unpredictable environments is the last frontier of automation, not the first. Your plumber’s job is safer from AI than your financial analyst’s.

Next, consider supply. In the United States, we have spent decades steering every capable student toward four-year university degrees, creating a massive skilled-trades shortage. Electricians, welders, plumbers, and HVAC technicians have pricing power that many white-collar knowledge workers would envy. An experienced electrician in a major metro can earn $120,000 to $150,000 a year with no student debt and high job security. Try saying that about a freshly minted communications major.

Even better - AI blends beautifully with these trades. An electrician who uses AI to optimize energy systems, diagnose problems faster, and manage a crew with AI-powered scheduling becomes dramatically more productive. AI is a complement to physical-world expertise, not a substitute. The trades are not a fallback. They are a strategic choice.

The obstacle is cultural, not economic. We have created a status hierarchy where a philosophy major working at a coffee shop has more social prestige than an electrician earning six figures. That is a market inefficiency driven by signaling norms. I am calling it out plainly. If we are serious about preparing young people for an AI world, we need to talk about the trades with the same respect we give investment banking and consulting. The economics demand it, even if the social signals have not caught up.

Healthcare: Durable but Transformed

Nursing and medicine are structurally safe. Aging populations, chronic disease burden, and the fundamental human need for care from other humans ensure that healthcare demand is not going away. But the nature of the work will shift enormously.

AI will handle diagnosis support, treatment planning, administrative burden, and routine monitoring. Much of what medical students spend years memorizing will be handled by systems that are faster and more accurate than any human memory. Fewer radiologists will be needed to analyze medical images. What remains uniquely human: intricate surgery, clinical judgment under ambiguity, the ability to integrate AI recommendations with a specific patient’s context, emotional presence, the conversation that helps a scared patient make a difficult decision, and the ethical weight of choosing when to override the algorithm.

The advice for a student interested in healthcare: pursue it with conviction. But build your professional identity around physical skill, judgment and human connection, not around information mastery. The doctor who thrives in 2035 is not the one who memorized the most. It is the one who knows what to do when the AI’s recommendation does not match what they see in the patient’s eyes.

Where the Jobs Will Be

Let me move from fields of study to the actual landscape of work, because students rightly want to know: where do I end up?

The jobs that are growing sit at the intersection of human judgment and AI capability. They are roles where a human sets the direction, defines the quality standards, manages the exceptions, and takes accountability for outcomes, while AI handles speed, scale, and pattern-matching.

Every AI system needs someone to design it, someone to evaluate whether it is working, someone to handle the cases it gets wrong, someone to explain its outputs to stakeholders, and someone to decide when to override it. Those are all human roles, and they require the meta-capabilities I described earlier. They also require domain expertise: you cannot govern an AI system in healthcare if you do not understand clinical practice, and you cannot orchestrate AI in marketing if you do not understand customer behavior.

The jobs that are shrinking are the ones in the middle: roles defined by processing information, following established procedures, and producing routine outputs. These include much of traditional financial analysis, standard legal research, basic software development, routine content creation, and administrative coordination. Not because these tasks are unimportant, but because AI can now do them faster, cheaper, and often better.

This is why I keep coming back to skill security. The specific job titles of 2035 are unpredictable. But the capabilities that will be rewarded are not: judgment, orchestration, communication, ethical reasoning, and deep domain expertise that gives AI something meaningful to amplify.

Start Building Now

I want to close with something actionable, because advice without practice is just pontification. If you are a student, or someone advising a student, here are habits worth starting this week. Not because they guarantee a specific career, but because they build the kind of mind that AI amplifies rather than replaces.

Write by hand regularly. Not because handwriting is sacred, but because the slowness forces you to think before you write, to choose words deliberately, and to develop an internal voice that is yours. In a world where AI can generate fluent prose on any topic, having a distinctive voice is a competitive advantage.

Read long-form material without summarization tools. Build the stamina to sit with a 300-page book and extract meaning through your own effort. This is cognitive endurance, and it is disappearing. The ability to hold a complex argument in your head, follow its logic, and form your own view is exactly the capability that AI threatens to atrophy.

Argue positions you disagree with. Force yourself to argue for the other side. This builds intellectual flexibility and guards against the confirmation bias that AI can amplify when it tells you what you want to hear.

Build something physical. Woodworking, cooking, gardening, wiring a circuit, fixing an engine. Engage with the material world where feedback is immediate, honest, and cannot be prompt-engineered away. There is a reason that every wisdom tradition values craft: it teaches you that reality does not negotiate.

Practice being wrong. Keep a journal of predictions and beliefs. Review them. Notice where you were wrong. Update. This is the fundamental loop of learning, and it requires the intellectual honesty to confront your own errors rather than letting AI shield you from them.

* * *

The question I hear most often, “what should my kid study?,” assumes that the answer is a field. It is not. The answer is a set of capabilities that no field owns and no AI can replicate: the ability to think from first principles, to see patterns others miss, to communicate with precision and empathy, to make decisions when the data is incomplete, and to take responsibility for the outcome.

Build those capabilities, and every field is open to you. Skip them, and no degree will save you.

In Part 3, I will speak directly to parents. Because knowing what to build is only half the battle. The harder half is creating the conditions where your children actually build it. That means protecting the one thing every instinct tells you to eliminate: the struggle.

* * *

This is the first in a three-part series on skills, jobs, and learning in the age of AI. Part 1 maps the terrain: what AI is doing to how we learn, work, and build expertise. Part 2, “Philosophy, Plumbing, and Where the Jobs Will Be,” offers concrete guidance on which skills and fields survive. Part 3, “Advice for Parents: Protect the Struggle,” addresses the hardest question of all: how to raise a capable human when the easy path is always available.

* * *

I am asked this question almost every week.

It comes from parents at dinner parties. From executives in my programs who are thinking about their children, not their companies. The question takes different forms, but the anxiety underneath is always the same: What should my child study? Where are the jobs? How can kids learn to learn when AI can give them instant answers?

I have been a professor of marketing and technology at Northwestern for almost 35 years. I have taught tens of thousands of students, and I have advised senior leaders at some of the world’s largest technology companies. I will be honest with you: these questions have never been harder to answer than they are right now. The speed at which AI is reshaping the landscape of work and learning has no precedent in my career, and I have lived through the internet, mobile, cloud, and social revolutions.

I will try to answer these questions. I will not speak with the hubris of a futurist. I will speak with the perspective I have gained over decades building pattern recognition skills across technology and business. And from the point of view of someone who has spent the last seven years going deeper into AI than anything else in my professional life. This series is my honest attempt to share what I see.

Let me begin with a story about a lever. This story will anchor the three-part series.

The Precondition Everyone Forgets

“Give me a place to stand and a lever long enough, and I will move the world.”

Archimedes said this over two thousand years ago. Everyone remembers the lever. Almost nobody remembers the precondition. He did not say “give me a lever.” He said “give me a place to stand and a lever.” The fulcrum comes first. Without solid ground beneath your feet, the lever is just a stick.

This analogy is very apt for where we are with artificial intelligence. We are in a moment of collective infatuation with the lever. Every conference, every headline, every investor pitch is about the power of the tools. Don’t get me wrong. The tools are extraordinary. I have experienced the power, and I am in awe. The lever is real. And it is getting more powerful by the day.

But few people talk about the fulcrum. The fulcrum is human judgment. It is pattern recognition. It is the ability to frame a problem before solving it, to ask the right question before generating an answer, to know when the model is brilliant and when the model is lying to you eloquently. This fulcrum can only be built one way: through years of struggle, practice, error, and correction. There are no shortcuts. There never have been.

What AI Actually Disrupts

To understand why this matters, you need to understand what AI actually disrupts. It is not just automating tasks. It is disrupting the learning process that produces expertise.

Consider what happens when a student uses ChatGPT to write an essay. The obvious concern is cheating. But the deeper damage is invisible. When you write an essay yourself, you are not just producing text. You are clarifying a vague intuition. You are discovering what you actually believe through the discipline of articulating it. You are confronting the weakness in your own argument when you try to put it on paper, and it falls apart. The essay is the artifact. The thinking is the real learning. Outsource the thinking, and you have atrophied the learning.

This example applies to every knowledge task. When you ask AI to summarize a 30-page report, you skip the cognitive work of reading carefully and distinguishing what matters from what does not. That discrimination skill is the foundation of expertise. When you ask AI to write your code, you skip the debugging that builds your understanding of how systems actually work. When you ask AI to generate your strategy, you skip the messy human process of weighing tradeoffs under uncertainty that builds real judgment.

I call this problem premature abstraction, borrowing a concept from computer science. In programming, you are warned never to abstract too early, before you understand the underlying patterns. The same principle applies to learning. AI allows students and professionals to jump to high-level outputs before they have done the low-level work that makes those outputs meaningful. The product looks the same. The person behind it is fundamentally different.

Here is the sentence I want you to remember: you cannot supervise a process you have never done yourself. A senior partner at McKinsey can use junior analysts effectively because she has done the analysis herself thousands of times. She knows what good looks like. She can smell a flawed assumption in a spreadsheet. A student who has never built an argument from scratch cannot evaluate whether AI’s argument is sound. They lack internal calibration. They have no place to stand.

The Jobs Question, Honestly

Let me address the fear directly, because it hangs over this entire conversation. Will there be fewer jobs? Are our children walking into a world where human work is obsolete?

The leaders building these systems are not reassuring on this point. Sam Altman and Dario Amodei have both spoken publicly about the magnitude of the disruption ahead. Amodei’s vision of “radical abundance” through AI is optimistic in the long run but brutally honest about the transition. Altman has said that AI will eliminate many jobs and that society needs to prepare. These are not critics. These are the people building the most powerful AI systems on earth. When they express concern, it deserves serious weight.

Here is my take. Over a long enough horizon, say 20 to 30 years, I believe new forms of work will emerge that we cannot yet imagine, just as they have in every prior technological revolution. The internet did not produce net fewer jobs. It produced different jobs: jobs that no one in 1990 could have predicted. I expect the same pattern here.

But the transition will be savage in its distribution. And the speed is unprecedented. Previous technological revolutions played out over decades. AI is compressing that timeline to years. The middle tier of knowledge work, the credentialed-but-not-expert layer that processes information, synthesizes reports, and generates routine analysis, is being hollowed out right now. Not in five years. Now.. The people with genuine expertise can use AI as an extraordinary amplifier. The people who were coasting on credentials and process knowledge are discovering that AI can do what they do, faster and cheaper.

So the answer to “will there be fewer jobs?” is the wrong question. The right question is: fewer jobs for whom? And the answer is: for people who never built a place to stand. For people whose value was in execution, not judgment. For people who learned to produce outputs but never learned to think.

The new jobs, the ones that will be created, will go to people who can do what AI cannot: frame problems, exercise judgment under ambiguity, build trust with other humans, take ethical responsibility for outcomes, and orchestrate complex systems where AI is one component among many. These are the people with a fulcrum. AI is their lever. Everyone else is holding a stick.

Stories from Lived Experience

I want to make this concrete. Over the past few years, I have gone deeper into AI than any subject in my career. And I can tell you that the Archimedes principle is not a metaphor for me. It is my daily experience.

I built a custom AI system for writing business case studies. I encoded three decades of case-writing methodology into it: my structure, my style, my tone, my standards, my pedagogical logic. The foundation included insights from over 50 cases I have written over the years. The system now works the way I work. I can go from an idea to a solid first draft within a couple of hours. Astounding acceleration! But every creative decision, every structural choice, every judgment call is still mine. AI did not replace my expertise. It operationalized it. I had to build the expertise first before I could encode it.

I developed a framework called I-MOS, the Intelligent Marketing Operating System, for MIT Sloan Management Review. It maps seven core marketing workflows against an agentic AI operating stack. The architecture came from pattern recognition across hundreds of conversations with CMOs and decades of teaching and consulting. AI helped me iterate and refine at a pace that would have taken months of solo work. But the conceptual breakthrough, seeing the structure that organized what had been a fragmented landscape, was mine. AI did not see the pattern. I did. And then AI helped me articulate it with precision and speed.

Most recently, I have been building a comprehensive ontology for AI-driven marketing, a knowledge architecture that maps how concepts, workflows, technologies, and organizational capabilities connect. The depth and speed at which this work has progressed is, frankly, breathtaking. But the ontology reflects my mental model of how these domains relate. AI helped me externalize it, structure it, pressure-test it. The intellectual DNA is mine.

In each case, the pattern is identical. Human insight first. AI amplification second. Place to stand first. Lever second.

This Article as Proof of Concept

I want to tell you how this series came into being, because the story itself illustrates the thesis.

I was sitting at O’Hare, waiting to board a flight to Delhi. I had been carrying these ideas for months: conversations with anxious parents, observations from my own AI practice, a growing conviction that the education conversation was missing the point. At the gate, I opened my laptop, started a conversation with Claude, and began thinking out loud.

I did not ask AI to write an article. I brought the raw material: the instinct that skill security matters more than job security, the conviction that trades deserve more respect, the Archimedes metaphor that had been forming in my mind, the personal examples from my own work. I pushed. AI pushed back. I refined. AI helped me see structure in what had been a collection of instincts. I rejected some suggestions and sharpened others. The conversation surfaced the architecture of a three-part series.

By the time I boarded the flight, I had a solid first draft of three articles. Granted, the flight was delayed by 45 minutes. But still...

Now here is the question: was that AI slop? Was that a machine writing on my behalf? I would argue exactly the opposite. What happened at that gate was my accumulated expertise finding a lever powerful enough to match its ambition. The speed was not a shortcut. It was the result of the foundation. Every instinct I brought to that conversation was earned over decades of teaching, writing, consulting, and thinking. AI did not give me those instincts. It helped me articulate them faster than I could have alone.

A student with no experience in AI strategy, marketing transformation, or case writing could have sat at that same gate with the same tool and produced nothing of value. The lever was identical. The fulcrum was not. We all have word processors. But all of us are not Shakespeare!

What Comes Next

The most powerful cognitive lever in human history is upon us. It will reshape work, education, and expertise more profoundly than any technology since the printing press. The people who thrive will not be those who adopt AI fastest. They will be those who built something solid to stand on before they picked up the tool.

You have the lever. The question is whether you have a fulcrum.

In Part 2, I will get specific. If you are a student, or someone advising a student, what should you actually study? Which fields, skills, and habits build the kind of foundation that AI amplifies rather than replaces? The answer involves more philosophy and plumbing than you might expect, and less coding than the conventional wisdom suggests.

In Part 3, I will speak directly to parents. Because the hardest part of this story is not knowing what to build. It is having the courage to let your children struggle while they build it.

* * *

Most publications tell you what they’re about. I’d rather show you.

The Hidden Weave exists because the best ideas don’t stay in their lanes. Ancient wisdom shows up in boardrooms. Business frameworks illuminate personal decisions. History rhymes with headlines. My job is to find those threads and pull them together.

My loom produces three weaves (more to come!):

Strategic Weave is where I go beneath the surface and beyond the headlines to understand what's really going on and what to do about it. Frameworks, mental models, and deep dives into AI. Start with Two Moves, where I argue that AI will make you irrelevant if your value lives in a process rather than a perspective that only you can offer.

Inner Weave is about spirituality, mindfulness, and the inner work. I draw on ancient wisdom to illuminate how we must live and lead. Start with Don’t Offer Wilted Flowers, which reminds you that your best self is needed now, not saved for some perfect moment that may never come.

Personal Weave is about personal growth and reinvention. Here I will explore the habits, practices, and lived lessons that shape who we’re becoming. Start with The New Physics of Skills, because AI has made skills fluid, which means that you must become deeper in your anchor skills and broader in you radius of expertise.

You don’t have to read everything. Pick the stream that matches where you are right now.

I write at full resolution here, not compressed for a platform. I don’t publish on a rigid schedule. I publish when I have something worth saying. And in every post, I promise you will find at least one idea that will stay with you after you close the tab.

That’s the weave. Glad you’re here.

- Mohan

Subscribe now

Look at any product team today and you’ll notice something odd. The org chart hasn’t changed. The product manager still sits at the center. The designer still owns the interface. The engineer still writes code. The marketer still runs campaigns. Same titles. Same Slack channels. Same reporting lines.

But watch what actually happens in the meetings. The PM shows up with a wireframe she built in twenty minutes using AI. The marketer is running his own A/B tests without waiting for analytics. The designer is generating front-end code. The engineer is drafting product copy.

Nobody announced a reorganization. Nobody changed anyone’s title. But the boundaries between these roles have become porous in ways that would have been unthinkable a few years ago.

I have been thinking about this shift through a metaphor borrowed from physics. For decades, skills inside organizations behaved like solids. A skill belonged to a role. A role belonged to a function. A function belonged to a department. Movement across these boundaries was possible but slow, expensive, and rare. Crossing into another team’s domain was like walking through a wall. If a marketing manager wanted to run a complex data analysis, she needed months of training or dedicated help from an analyst. If an engineer wanted to do interface design, he needed to go to design school.

AI has changed the temperature. And when the temperature rises enough, solids become liquids.

The Phase Change

AI does not make everyone an expert at everything. That is simply not possible, and in fact it would be dangerous. What AI does is dramatically reduce the friction involved in acquiring and applying skills in adjacent domains. The cost and effort to stretch into a neighboring discipline has collapsed.

Research that once required a dedicated analyst can be synthesized in minutes. Designs can be mocked up from a text prompt. Code can be prototyped without a formal engineering background. Data analyses can be run by people with no SQL knowledge. The barrier hasn’t disappeared, but it has dropped from a wall to a curb.

This is why skills are becoming fluid. They no longer stay locked inside functional silos. They flow across roles, across teams, across the old boundaries that once kept everyone in their lane.

The implications are significant. LinkedIn’s workforce data suggests that by 2030, roughly 70% of the skills used in jobs will have changed, with AI as a primary catalyst. An Atlassian design leader, after the company ran an all-hands AI experimentation week, put it this way: what could have taken months of self-learning happened in days.

Anchor and Radius

When skills flow freely, we need a new way to think about professional capability. I propose a simple framework: anchor and radius.

Every professional has an anchor. This is their core domain of deep expertise and judgment, built through years of learning, practice, pattern recognition, and tacit knowledge. The anchor is what allows someone to make good decisions when information is incomplete, to exercise taste in their field. AI does not replace the anchor. If anything, anchors matter more in an AI world, because they guide the effective use of all the new tools. A weak anchor paired with AI leads to confidently wielding tools in foolish ways. As the saying goes, a fool with a tool is still a fool.

What AI does is extend the radius of competence around that anchor.

Think of each professional as having a circle of capability centered on their core expertise. AI increases the diameter of that circle. The product manager’s anchor might still be product strategy and orchestration, but her radius now extends further into design, analytics, even light engineering. The marketer’s anchor remains customer insight and creative storytelling, but his radius now reaches into data analysis, growth experiments, and user experience.

Without an anchor, a larger radius is dangerous. With a strong anchor, a larger radius is a force multiplier. Careers in the AI era will be defined not just by the depth of your anchor but by how far your circle extends around it.

Here’s the observation that changes how organizations work: as everyone’s circle expands, the circles inevitably overlap.

What Happens in the Overlap Zones

The PM’s circle now overlaps with design. The marketer’s overlaps with analytics. The engineer’s overlaps with UX. These adjacencies create very real tensions inside organizations.

Who owns the early user experience work when product managers can generate wireframes and user flows themselves? Who owns insight generation when a marketer can run experiments and queries without analyst support? Who defines technical feasibility when AI can generate a plausible code prototype for a non-engineer?

These aren’t turf battles or ego fights. They are structural consequences of reduced friction.

Consider what Airbnb did. When Brian Chesky announced in 2023 that they had merged product management and product marketing, it was widely misinterpreted as eliminating PMs. It wasn’t. It was a recognition that the overlap zone between building the product and marketing the product needed a single integrated owner. “You can’t develop products unless you know how to talk about the products,” Chesky explained. The skill sets had overlapped so much that maintaining separate roles was itself creating friction.

When designers at the Figma conference cheered Chesky’s announcement, they revealed the tension. Designers felt some PMs were encroaching on their terrain in unproductive ways, and the old role definitions weren’t resolving it.

The key managerial question is not whether overlaps exist. They do, and they’re growing. The question is: who prevails in the gray zones? I see three forces that determine influence in these overlap areas.

Anchor depth. When a decision truly hinges on deep craft or specialized mastery, the domain specialist retains authority. AI might help a PM create a draft design, but information architecture and interaction design nuances still require a seasoned designer’s judgment. Paradoxically, AI can increase respect for deep expertise by automating the easy parts and highlighting the hard parts.

End-to-end ownership. When decisions involve balancing trade-offs across the whole product or value stream, the person with holistic accountability prevails. Product managers typically own the end-to-end outcome, which naturally positions them to integrate across domains. AI-augmented breadth makes them even more effective here, because they can personally explore each side of a trade-off before deciding.

Decision frequency. Influence accrues to whoever makes the most frequent, iterative decisions. In a fast-paced, AI-powered workflow, there are dozens of micro-decisions every day: tweaking a prompt, choosing which experiment to run, adjusting copy based on feedback. The person consistently in that loop gains authority over time. High-frequency decision-making, compounded over dozens of iterations, beats a slow deliberative process that routes through hierarchy.

These forces suggest that overlapping skills won’t produce a free-for-all. We’ll see a new balance of power based on who brings deep craft when it matters, who integrates across functions, and who drives fast iterative cycles.

The Rise of the Full-Stack Product Builder

No role illustrates this shift more clearly than product management. Even before AI, great PMs were defined by their breadth. They sit at the intersection of customer needs, business strategy, design, and engineering. Their value has always come from synthesis rather than deep craft expertise in any single area. AI amplifies that inherent breadth dramatically.

LinkedIn saw this early and acted on it. In 2025, they sunsetted their Associate Product Manager program and replaced it with the Associate Product Builder program. The name change is telling. Participants learn coding, design, and product skills together, leveraging AI tools throughout. LinkedIn even created a formal “Full-Stack Builder” job title with its own career track, enabling anyone from any function to take a product from idea to launch.

The rationale, as LinkedIn’s Chief Product Officer Tomer Cohen explained it, was that the traditional model of highly specialized teams was too slow. Organizational bloat meant even small features took six months, whereas AI tools allow leaner teams to deliver faster. The application process reflects this philosophy: no resume required. Candidates submit a 60-second demo of a product they built and answer questions about how they used AI in the process.

Shopify has moved in a similar direction. Their PMs don’t just write specs. They use AI to generate prototypes and analyze data independently. Atlassian ran an internal “AI Product Builders Week” where over a thousand designers, PMs, and engineers set aside their normal tasks to experiment with AI together, producing dozens of production-ready prototypes in days.

The full-stack archetype integrates three capabilities: strategist (market understanding, customer insight, product vision), builder (hands-on creation, prototyping, technical execution), and growth driver (adoption, experimentation, metrics, iteration). The anchor remains judgment. The expanded radius shows up in the building and growth dimensions.

This isn’t just a PM story. Marketing is converging from the other direction. Today’s marketers segment customers with AI-driven clustering algorithms. They generate and test content variations at scale. They build personalized user journeys that start to look like product UX flows. They run A/B tests and attribution models that were once the domain of data scientists. At HubSpot, 74% of marketers were using at least one AI tool at work in 2024, up from just 35% the year before.

The pattern is the same across functions: T-shaped professionals, deep in one area with broad ability in others, are increasingly valuable. The horizontal bar of that T can now stretch much further than before.

The Risks of Expanding Your Radius

I want to be direct about the dangers, because the full-stack archetype has failure modes.

The first is the jack-of-all-trades trap. Expanding your radius without maintaining your anchor depth makes you broadly mediocre rather than distinctively valuable. Organizations still need deep specialists. Not every product manager, marketer, or engineer will become full-stack, and that is perfectly fine.

The second risk is burnout. Because full-stack people can do so much, organizations tend to pile responsibilities onto them. Just because your star PM can also do great design work and run analyses doesn’t mean she should do all of it all the time. The new bottleneck is cognitive bandwidth. No matter how talented an individual is, there are limits to how many domains they can integrate in a single day. Full-stack humans need protection from their own capability.

The third risk is starting-point bias. In complex work, whoever produces the credible first draft of an artifact wields outsized influence on the outcome. If AI allows PMs to generate the first draft in domains that used to be out of their reach, they increasingly set the direction before the specialist formally gets involved. A PM who comes to a meeting with a plausible wireframe is effectively steering the design conversation from the outset. This can accelerate progress. It can also marginalize expertise if not managed carefully.

What Leaders Must Do

The changing physics of skills places new demands on the C-suite.

Redefine roles around outcomes, not tasks. Rigid job descriptions that specify “you do X, then hand off to Y” are becoming obsolete. Define roles by the outcomes they own and let the tasks be fluid. Some companies are already dropping traditional titles in favor of “product builder” or “growth owner” to signal this shift.

Protect your integrators. Your best full-stack people are force multipliers. They are also the first to burn out if you load everything onto them. Ensure they have support. Monitor their workload. In performance reviews, check in on how they are managing breadth, not just depth.

Clarify decision rights in the overlap zones. As overlaps proliferate, ambiguity in decision-making creeps in. Define who has the call on what and when, even as collaboration increases. The PM might have final call on priority and scope, the designer on UX quality, the marketer on brand voice. Make it explicit. Overlaps without governance become stalemates.

Reward cross-functional value creation. If your compensation and promotion criteria only celebrate individual functional excellence, you are sending the signal that staying in your lane is safe. Highlight team wins that were only possible because someone ventured beyond their role. Create career paths that allow growth in breadth, not just depth.

Treat AI fluency as foundational literacy. Don’t leave this to chance. LinkedIn created an entire culture program around being AI-native in product development. Atlassian’s company-wide experimentation week gave every team member hands-on education. If you institutionalize it, you get a company-wide leap. If you leave it to individual initiative, you get pockets of innovation surrounded by inertia.

The Management Crossroads

We stand at a choice point. One path is to continue managing skills with the old assumptions, treating overlaps as anomalies or problems to be squashed. The other is to redesign organizations for skill liquidity, encouraging learning, experimentation, and collaboration across boundaries.

The companies already taking the second path are reporting not just faster output but higher employee satisfaction. Top talent wants to grow and contribute in multiple ways. If you provide the platform for it, they flourish. If you don’t, they leave for organizations that will.

Your organizational chart is a lagging indicator of how work actually happens. It describes the past, not the present. When skills become fluid, leaders who will win will let the skills flow and wisely guide the current.

Your company has an HR department. It recruits talent, onboards new hires, manages performance, handles compensation, ensures compliance, and plans for the future of the workforce. No serious company would operate without it.

Now ask yourself: who is doing any of this for your AI agents?

Large enterprises are deploying agentic AI across every function, from marketing to finance to customer service. These agents read documents, draft responses, make recommendations, flag risks, and interact with customers. Some operate around the clock. Some make decisions that affect revenue. Some touch sensitive data.

But these digital workers have no onboarding process, no performance reviews, no governance structure, no clear ownership, and no one tracking whether they’re actually delivering value. They are, in effect, feral employees: hired enthusiastically, deployed tactically, and managed by no one.

Your human workforce has an entire management discipline behind it. Your digital workforce needs one too. I call it Digital Labor Orchestration (DLO), and this article is its blueprint.

The Parallel That Changes Everything

The simplest way to understand DLO is through a mirror. Hold up your existing HR function and ask: what would this look like for AI agents?

Hiring employees becomes sourcing and selecting digital agents. Do you build your own? Buy from Salesforce or Microsoft? Outsource to a service provider? The decision depends on the same factors as human talent acquisition: strategic importance, IP sensitivity, and cost.

Onboarding and training becomes configuring and fine-tuning agents. Just as a new hire needs context, tools, and expectations, an AI agent needs prompt engineering, access permissions, guardrails, and escalation rules. Skip this step and you get the AI equivalent of an employee who was never told what their job actually is.

Performance management becomes monitoring agent effectiveness. What are the accuracy rates? Completion rates? How often does the agent escalate versus resolve? Are its outputs drifting over time? You wouldn’t tolerate a human employee who was making expensive mistakes and messing up their jobs. Why tolerate agents that hallucinate or go off the reservation?

Compensation and costing becomes understanding the true economics of digital labor. The sticker price of an API call is not the cost of digital labor, just as a salary is not the total cost of employing a human. You need to account for compute, licensing, monitoring, retraining, exception handling, and human oversight. I call this the Total Cost of Labor Ownership (TCLO): the sum of human labor cost, digital labor cost, and orchestration cost.

Compliance and ethics becomes guardrails, auditability, and explainability. When an AI agent makes a loan decision, who’s liable? When it generates customer-facing content, who reviews it? When it fails silently, who notices?

This isn’t a metaphor. It’s an operating model.

From Jobs to Flows

Here’s where DLO parts company with traditional workforce thinking. In the human world, work is organized in terms of roles and titles: “marketing manager,” “loan officer,” “customer service representative.” In the DLO world, the organizing principle is workflows, broken down into tasks and micro-tasks.

The first step in any DLO initiative is what I call a Workforce X-Ray: a deep diagnostic that decomposes roles into the actual tasks people perform, then evaluates each task for its potential to be handled by a digital agent. A European bank that performed this exercise on its loan origination function discovered over 120 micro-tasks, from intake and document review to fraud checks and credit decisioning. Some tasks were obvious candidates for automation. Others required human judgment. Most fell somewhere in between.

This decomposition is where the real insight lives. A “loan officer” is not one job. It’s dozens of micro-tasks stitched together by habit and job description. Some of those tasks are ripe for AI. Others are deeply human. The art is in the balancing.

The Four A’s: A Volume Dial for Autonomy

Once you’ve identified which tasks can be handled by digital agents, the next question is: how much autonomy should an agent have? This is not a binary choice between “automated” and “not automated.” It’s a dial with four settings.

• Assist. The agent enhances human productivity through insights, summarization, or suggestions. A marketing copilot recommends headlines. A research agent summarizes competitive intelligence. The human decides. The agent informs.

• Approve. The agent performs the work, but a human must approve before anything executes. A contract analysis tool identifies risk clauses, but legal counsel signs off. The agent proposes. The human ratifies.

• Audit. The agent operates autonomously, but its decisions are reviewed after the fact. A dynamic pricing agent adjusts rates in real time. A human audits a sample weekly. The agent acts. The human verifies.

• Autopilot. The agent owns the workflow end to end without human intervention. An inventory replenishment system orders stock based on real-time demand signals. The agent decides. The human has moved on to higher-order work.

The instinct of most executives is to jump straight to Autopilot. Resist it. The right autonomy level depends on the risk profile of the workflow, the maturity of the agent, and the trust your organization has built through experience. A billing dispute agent at a logistics company might start at Audit (handling tier-1 cases independently, with 10% weekly human review) and graduate to Autopilot in low-risk regions only after three months of demonstrated performance. High-risk markets might stay at Audit permanently.

The Four A’s give you a common language for a conversation that every enterprise is having in fragmented, inconsistent ways.

The Agentization Map: Where to Start

Not every workflow deserves digital labor, and not every promising workflow deserves the same level of investment. To prioritize, plot your workflows on a 2×2 matrix (See the diagram below):

• Y-axis: Agentization Potential (how well-suited is this workflow for AI agents?)

• X-axis: Strategic Importance or Volume

This produces four zones:

• Accelerate (high value, high potential): Fast-track these for digital labor redesign. This is where your biggest ROI lives.

• Activate (low value, high potential): Perfect test beds. Low risk, high learning. Use these to build organizational capability before tackling the high-stakes workflows.

• Augment (high value, low potential): Deploy copilots and human-in-the-loop systems. The work is too complex or too risky for full automation, but AI can make humans significantly more effective.

• Avoid (low value, low potential): Don’t waste resources here. Focus elsewhere.

This map turns a sprawling, overwhelming question (“where do we start with AI agents?”) into a portfolio decision with clear priorities

Building the Digital Labor Office

If DLO is the discipline, the Digital Labor Office is the institutional home. Think of it as the organizational equivalent of HR, but for your AI workforce. It’s headed by a senior executive (call it Head of Digital Labor) and staffed with:

• Workflow Architects who map processes and redesign them for human-AI collaboration

• Agent Designers who configure, prompt, and embed guardrails into digital workers

• Governance and Risk Officers who ensure compliance, auditability, and appropriate autonomy levels

• Human-AI Partnership Managers who handle change management and ensure human workers embrace (rather than fear) their digital colleagues

• Agent Product Managers who treat each digital agent as a product with a backlog, performance metrics, and lifecycle

This team works in deep partnership with HR. Together, they co-own workforce planning (how many humans, how many agents?), role redefinition (what does a “loan officer” do when AI handles 60% of the micro-tasks?), and cultural transformation (how do you move from “AI is taking my job” to “AI is making my job more interesting”?).

I recommend a hub-and-spoke operating model: a central DLO team sets strategy, tools, and policies, while embedded DLO Champions in each business unit adapt and execute locally. This mirrors the HR Business Partner model that is common in human workforce management.

Measuring What Matters

A common mistake in digital labor is measuring only cost savings. Yes, a bank that deploys AI agents in loan origination can cut processing costs by 60% and improve speed-to-decision by 40%. Those numbers matter. But they’re the floor, not the ceiling.

DLO measurement should span three dimensions:

• Productivity. Not just human productivity, but blended productivity. I propose a metric called Blended Workforce Productivity (BWP): total output divided by the sum of human labor input (in FTEs) plus digital labor input (in ATEs, or Agent-Time Equivalents). This puts humans and agents on the same scorecard.

• Cost efficiency. The TCLO model described earlier, tracking not just compute and licensing but orchestration costs: governance, exception handling, supervision, and change management.

• Strategic impact. Speed to market, customer experience uplift, innovation enablement, operational resilience. If a bank can cut loan approval time from 4 days to 40 minutes, it won’t just save money. It will increase application volumes because of the improved customer experience.

The Journey Ahead

Most enterprises today deploy AI agents sporadically, with no unified strategy for sourcing, governing, or measuring their digital workforce. Digital Labor Orchestration is a journey. To map this journey, I have created a DLO Capability Maturity Model with five stages of maturity, from Ad Hoc deployments with no formal structure to Institutionalized operations where digital labor is embedded in the company's operating model, talent strategy, and performance management systems. See the Table below for the full maturity model with dimensions, capabilities, and diagnostic indicators at each stage.

Digital Labor Orchestration is not a technology initiative. It is a management discipline. The companies that built great HR functions gained a durable advantage in the human capital era. The companies that build great DLO capabilities will gain the equivalent advantage in the age of AI.

Your AI agents are already working. Now you need to figure out how to manage them as employees.

The inspiration for this framework was a conversation with my friend Parimal Deshpande, who works at Adobe. He told me that ever since he started a new role, he had been able to tap into his creative side. He was energized in a way he hadn’t been before, because he has been a creator all his life. Suddenly his job felt meaningful. That conversation got me thinking about what actually drives professional fulfillment. This framework is the result.

There are three pillars. They are simple to name but difficult to align.

Performance

The foundation is performance. This is your competence and skill in the role you occupy. You must be good at your job, and that competence comes from experience, capability, and the relevance of your skills to the context you are in. That last part matters more than people realize. You might be a phenomenal guitar player. But if you’ve been asked to play the piano, your talent is real but misplaced. Skill without fit is potential without traction.

If your job were a human body, performance would be the muscle and the mind. It is the starting point. But functional competence must be maintained and matched to context for it to translate into results.

Passion

The second pillar is passion. If performance is the muscle, passion is the heart. This is what inspires you, motivates you, and makes you excited about what you do. Think about when you wake up in the morning. Are you looking forward to the work ahead?

For a salesperson, passion might be the thrill of closing a deal. For me, it is helping students learn and watching them grow their careers. Passion may sound lofty, but it can be about small things. You might be a barista who takes pride in creating the perfect cappuccino. You might be a hotel front desk clerk who delights in making guests feel welcome. Passion is your in-the-moment satisfaction. It is the rocket fuel that converts a competent employee into an exceptional one.

Purpose

The third pillar is purpose. This is your mission. Do you derive meaning from your job? Are you making a difference?

Some professions are naturally purpose-driven: nursing, teaching, firefighting, public service, the military. Some organizations also index higher on purpose within their industry. I served for ten years on the board of Reliance Jio, a company whose purpose was crystal clear - bring affordable data connectivity to a billion Indians. Microsoft, Nike, The Body Shop, Patagonia: these are organizations where purpose is baked into the culture. A sense of purpose is what makes people go for the moonshots.

The Trifecta

Performance is the body. Passion is the heart. Purpose is the soul. When all three are present, you have professional fulfillment.

Think of your career as a ship sailing the seas of professional opportunity. Performance is the sails: strong, broad, and able to push the ship forward. Passion is the wind that fills those sails. Purpose is the North Star: the destination you are sailing toward. If that destination is meaningful, and the sails and wind are working together, you have a ship that is not only seaworthy but headed somewhere worth going.

Here is the critical insight. These three elements do not add together. They multiply.

Fulfillment = Performance × Passion × Purpose

If any one of the three is zero, the product is zero. You can be highly skilled and deeply passionate, but if the work has no meaning, fulfillment collapses. You can be mission-driven and competent, but without passion, you are grinding. You can be passionate and purposeful, but without competence, you are ineffective.

The Japanese call this alignment Ikigai: the discovery of your reason for being. The three P’s are my way of making that idea actionable. Find the role where your skill meets your energy meets your meaning. That is where fulfillment lives.

Here’s a formula I can use to improve my life. I imagine that tomorrow is my last day. Knowing this, I ask if yesterday was a day well lived. If I could have done better, I try to make the most of today. Repeat daily.

If you keep your head in the clouds and your feet on the ground, you will become a very tall person.

The force of habit is very strong. If I could just make a few habits good ones, the Force would be with me.

Do not see yourself in the uneven mirror of other people’s perceptions.

If you want to see Heaven, you will have to die yourself.

Look for solitude among the crowd. Look for silence among the screams. Look for serenity among the chaos. Be like the lotus flower, that lives in the swamp yet rises above it.

The Universe only sends weak signals of opportunity. Listen carefully and act early.

The best conversations I can have are with myself. The best work I can do is inner work. But I am too busy talking with everyone around me. And I am too busy working on what matters to others.

If you feel you are fully in control of your life, you aren’t driving fast enough. If you aren’t failing, you aren’t aiming high enough.

The sweeter you want your tea to be, the more sugar you have to put into it.

A student called me for advice on choosing among three excellent job offers. I told her to choose the job that is closest to her passion and furthest from her comfort zone.

We weave the fabric of life with the warp of actions and the weft of relationships.

The only way to walk with you is to walk by your side. If I try to lead or to follow, I walk alone.

Social media has expanded my peripheral circle of acquaintances, but it has shrunk my core circle of friends.

I should click less on “Like” and connect more with those I like.

When you are angry with me, you offer me the gift of negativity. When I react, I accept your gift, and I am obligated to return it. Can we stop this tradition of exchanging gifts?

You said that it was good to see me again. But I am not who I was yesterday. I have laughed. I have cried. I have loved. I have learned. I have grown. Tomorrow will be a new day. And I will be a new me.

Write mindfully. Your words are like ripples on a pond. They will travel farther and touch more people than you know.

Even though the present is but a moment, treat it as eternity. This moment is all there is.

Imagine for a moment what you could lose in an instant. Your health. Your wealth. Your home. Your family. Your job. Your love. Only when we think about what can be taken away from us do we realize how much we have been given.

This morning, I woke up to see snow falling gently in my garden. The decaying leaves and naked trees were covered with a fresh white blanket. Nature was showing me the dance of death and renewal.

I hide behind so many masks. If it is such an effort to be myself, who is the real me?

A camera cannot photograph itself.

Crutches are not legs, lenses are not eyes, assumptions are not facts.

Give without expectations. The Universe will pay you back. Just not how you expect. Or when you expect.

This is the full article behind the “method” discussion in my launch post.

Professional value in the age of AI is under pressure from two directions. The unit in which knowledge work is priced. And whether its outcomes are inseparable from the person who produces them. Most knowledge workers are only paying attention to one.

AI has not merely automated professional tasks. It has forced two questions that knowledge workers prefer not to ask themselves. The first is economic: when AI can do what you charge for, charging for doing it stops making sense. You can still bill by the hour. But AI has stripped away the opacity that made input pricing viable. Hours are an input metric. Outcomes are an output metric. When a task that took a junior associate three hundred hours takes AI thirty minutes, the gap between what clients were charged and what the work costs is no longer hidden.

The second question is existential: does the value you create exist because of you, or merely through you? If a client, a patient, or an organization could attribute what they received to a process, a platform, or a credential rather than to a specific individual, you are replaceable. If the outcome and the person can be separated, the outcome can be sourced elsewhere.

The professionals who will thrive in the age of AI will have made two moves. The first: decoupling output from input, so that intellectual capital scales beyond the hours in a day. The second: coupling outcome tightly with identity, so that what they produce cannot be attributed to anyone or anything else. The first move is about production economics. The second is about personal irreplaceability. Most of the literature on AI and professional work has described the first. Almost no one has described the second. Both are now mandatory.

My father understood the first move before AI existed. He flew bombers for the Indian Air Force before he became an entrepreneur. He had a precise way of cutting through complexity. He gave me a piece of advice that stayed with me: don’t measure your costs and revenues in the same units. He was speaking as a small business owner trying to protect his margins in the fabrication business, where he bought steel in kilos and sold window frames by the piece. After forty years, I appreciate how much insight his advice carried.

What AI Threatens and What It Does Not

The vulnerability of a professional to AI is determined by three structural characteristics of how their value is created and attributed.

The first is input transparency. When your process is observable and reproducible, AI can substitute for it directly. Document review, financial modeling, diagnostic imaging, market research, first-draft content production: these are processes that can be specified, observed, and replicated.

The second is output separability. When the artifact you produce can be detached from the person who produced it without loss of value, you have a separability problem. The contract, the report, the diagnosis, the slide deck: if the client values the deliverable and not the person who made it, they do not need you specifically. They need a capable producer. Increasingly, AI is that producer.

The third is outcome attribution. When the result the client cares about is attributed to a process, a firm, a platform, or a credential rather than to a specific individual, the individual has no moat. The client was never buying you. They were buying the result.

The most exposed professionals score high on all three. The most protected score low on all three: opaque inputs built through experience that cannot be observed or reproduced by instruction, inseparable outputs whose value is irreducibly bundled with the person who produced them, and outcomes attributed personally and completely to a specific individual.

This is what the pricing literature missed, because it was concerned with fee structures, not with the more fundamental question of whether the value survives the removal of the person who created it. Outcome-based pricing changes how you bill. It does not change whether you are replaceable.

The Lawyer Who Gets You Four Billion Dollars

David Boies bills by the hour. But the hourly rate is a billing convention. What his clients are paying for is something else entirely. When he represented the government in its antitrust case against Microsoft or advised investors in the aftermath of the Theranos collapse, the value being exchanged operated on an entirely different logic than billable hours. He is one of a handful of lawyers who can change the outcome of a case worth billions of dollars. His preparation, his deposition strategy, his cross-examination instinct are inputs, but they are not what the client is buying. The client is buying an outcome that would not exist without Boies in the courtroom. The input is opaque. The outcome is personally his. Both protective moves made.

The threat to lawyers is not at that level. It is at every level below it.

The Am Law 100 business model runs on leverage. Senior partners develop client relationships and originate business. Junior associates execute the work: document review, due diligence, contract drafting, research memos, discovery production. The spread between what associates bill and what they cost is the profit engine. That spread is collapsing, because the work is pattern recognition across large document sets, which is precisely what large language models do better than humans at a fraction of the cost.

When that justification disappears, so does the leverage model. Partners who own client relationships, whose outcomes are inseparable from their personal judgment, will capture more value than ever. Associates who expected to climb a partnership ladder by accumulating billable hours will find the ladder shortened from below.

The lawyers moving toward outcome-based pricing are on the right track. But they are only halfway there. Fixed fees, retainers, and contingency structures are the first move. The second move is making the attribution of the outcome irreducibly personal. Not the firm’s judgment. Not the practice group’s methodology. This person, this mind, this accumulation of experience that exists nowhere else.

The Surgeon of Last Resort

There is a small cohort of surgeons the ultra-wealthy call when a diagnosis has been delivered and they are not ready to accept it. They are the physicians who see the case three specialists missed, who will perform the procedure no one else at their institution will attempt. No patient asks how long the consultation took. No family asks about the hourly rate. The value is irreducible and the attribution is personal.

The disruption of medicine by AI is hurting the top of the cognitive hierarchy rather than the bottom, which surprises people until you understand the mechanism. Radiology. Pathology. Diagnostic dermatology. These specialties require exceptional training and years of pattern exposure. They are also, at their core, visual pattern recognition tasks, and AI has reached or exceeded human performance in several of them for specific conditions. The radiologist reading two hundred chest scans a day is performing high-skill cognitive labor whose process is entirely observable and whose output is completely separable from the individual who produced it. The report says “radiology department.” The patient rarely knows the reader’s name.

All three vulnerability conditions are present. AI steps into the gap, not because the radiologist lacks skill, but because the structure of the work makes both protective moves unavailable. The individual and the value were already disaggregated before AI arrived. AI simply made the disaggregation economically decisive.

The physicians who are protected are those whose value is irreducibly bundled with a specific person. The concierge physician who has known a family across three generations. The surgeon whose judgment and hands have performed this procedure ten thousand times and whose name is on the outcome in a way no departmental attribution can replicate. Their configuration is the mirror image of the radiologist’s, and so is their exposure.

The McKinsey Senior Partner

One of my former students is a senior partner at McKinsey. He does not write slides. In fact, he has not written slides in two decades. What he does is appear in the right boardroom at the right moment, when a CEO faces the kind of decision that ends careers or transforms companies. He draws on thirty years of wisdom to offer an invaluable and uniquely personal perspective. Then he leaves. The slides come from someone else.

His input is almost entirely invisible. His output is inseparable from him. Nobody in the room is thinking about the process. They want to know what this brilliant mind sees that they cannot.

The consulting model under pressure is not his. It is the pyramid surrounding him: the analysts and associates who ingest data, build financial models, populate templates, and produce deliverables that justify the engagement fee. AI does that work faster and more consistently at a cost approaching zero.

There is a structural vulnerability in consulting that does not exist in the same way in law. The consulting outcome is frequently attributed to the firm rather than to the individual. “McKinsey recommended this” is weaker personal attribution than “Boies argued this.” The brand intermediates between person and outcome, which dilutes the individual moat. The consultants building durable practices are those who have made their personal attribution explicit, who are known for a specific way of seeing problems that cannot be found elsewhere.

The industry is barbelling. At one end, firms with powerful institutional attribution will retain their position because the brand intermediates effectively between the work and the buyer. At the other end, boutiques built around a named individual with a distinctive point of view will also thrive. The vast middle is most exposed: firms and practitioners trading on neither strong institutional brand nor strong personal attribution, whose value proposition rests on competent execution of processes AI can now replicate.

For individuals inside strong institutions, the barbell creates a false sense of security. The institutional brand may protect the firm. It will not protect you if your outcomes cannot be attributed to you personally when the institution is removed from the equation.

The Business Guru and the Forty-Year Draft

I want to be transparent about something directly relevant to this argument.

This article went from first conversation to complete draft faster than I could have dictated it longhand. An AI model helped me structure the argument, pressure-test the logic, refine the exposition, and craft a logical narrative arc at a speed that was unthinkable three years ago.

And yet the framework at the center of this article, the three vulnerability conditions, the two protective moves, the distinction between the first move the pricing literature describes and the second move it ignores, are uniquely mine. Rooted in thirty-five years of watching how knowledge markets work, teaching strategy to executives across six continents, building programs, making mistakes, and developing a distinctive point of view. AI accelerated the time from idea to execution. It did not generate the insight. Archimedes promised to move the world if given a lever and a place to stand. My decades of experience are the place to stand. AI is the lever that extends the reach.

When Picasso was challenged on the price he charged for a portrait completed in minutes, he replied: “It took me all my life.” That is the precise situation of any professional whose value lives in accumulated judgment rather than current effort. The output looks effortless. The input is a career. As I like to say: a ladder has rungs for a reason.

I have reached more than forty thousand executive education participants in six years through online programs, generating over a hundred million dollars in gross revenue. In the classroom, I teach at most sixty students per class, the same number I taught thirty years ago. The decoupling between my input and my output is complete. But the scaling is only protected because the outcome remains attributed to something that cannot be replicated by removing me from the equation. Executives seek this specific point of view, these specific frameworks, this particular voice. The content scales without limit. The attribution does not transfer to the platform, the course, or the institution. It stays with me. Both moves made. Neither alone is sufficient.

What the Pricing Literature Missed

The case studies illustrate the logic. Here are the two moves stated precisely.

The first move is decoupling. Break the link between your effort and your output so that your intellectual capital scales beyond the hours in your day. This is productization. Stop selling time. Build something that delivers value while you sleep. Don’t be lulled into a false sense of security by inertia. Clients do not defect the moment AI enters the picture. But inertia is a reprieve, not a defense.

The second move is coupling. This is less intuitive and more important. Make the outcome of your work inseparable from your individual identity, so that a client, patient, or organization attributes the result to you specifically and not to a process, a firm, or a credential. Coupling is not about how you produce value. It is about whether the value is recognized as yours. It is the only durable answer to the question AI is now forcing every knowledge worker to confront: does the value exist because of you, or merely through you?

The two moves work in tandem. Decouple the production. Concentrate the attribution. Scale where you are going. Own the credit for arriving.

My father’s aphorism reaches further than he intended. Don’t measure your costs and revenues in the same units. It is not advice about invoicing. It is a statement about the nature of value. What professional work costs and what it is worth have never been the same thing. The gap between them was sustainable because it was opaque. AI has made the gap transparent. And a visible gap cannot be sustained.

Where This Leads

The professionals who will define knowledge work in the next decade share a configuration. Their inputs are opaque, built through experience that cannot be observed or reproduced by instruction. Their outputs are inseparable from their individual identity. Their outcomes are personally owned.

The senior litigator whose cross-examination instinct was built across a thousand depositions. The surgeon whose judgment has been formed by ten thousand procedures. The strategy partner whose mastery over organizational dynamics came from sitting in a hundred boardrooms at the highest stakes. The educator whose frameworks were forged across forty years of research, teaching, and synthesis that no one else has done in quite this way.

What is happening to professional work is not replacement. It is disaggregation: the separation of expertise from the process that used to house it. The hours were never the value. The credential was never the moat. The process was never what the client was paying for. AI has not changed what professional value is. It has made it impossible to pretend otherwise.

The honest question every knowledge worker needs to answer is not whether AI can do their job. It is two questions, more uncomfortable and more precise. In what units is my value denominated? And is the outcome of my work inseparably mine?

If the value is in the hours, the repricing has already begun. If the outcome could be attributed to a process, a platform, a firm, or a credential rather than to a specific irreplaceable individual, the substitution is closer than it appears.

If your value lives in outcomes that only your specific accumulated judgment can deliver, and if those outcomes are irreversibly coupled to your identity, you are not competing with AI. You are using it.

My father gave me a pricing principle for his small business in India. As he smiles down at me from heaven, I salute him for his timeless insight.