I want to tell you about the most exciting thing I’ve read this year, and it has nothing to do with chatbots.

Microsoft just unveiled Majorana 2, a quantum chip built using agentic AI, and the implications are the kind that make you imagine Majorana 2 telling every other computer to “hold my beer.”

First, a quick explanation of why quantum computers matter.

Your laptop computes in bits: every piece of information is either a 1 or a 0. Quantum computers use qubits, which behave in ways that seem almost magical. Instead of exploring one possible answer after another, they can explore vast numbers of possibilities simultaneously. For certain problems such as designing new drugs, modeling climate systems, or cracking encryption, this isn’t just faster. It’s a different realm of intelligence entirely.

The challenge is that qubits are extraordinarily fussy. They’re so sensitive that heat, vibration, or a stray electromagnetic wave can destroy a calculation in microseconds. Building a reliable quantum computer has been like downing three Red Bulls and trying to thread a needle while riding a rollercoaster during an earthquake in a hurricane.

Until now.

Majorana 2’s qubits can maintain their quantum state 1,000 times longer than the previous generation. Microsoft’s own analogy: imagine a phone battery that instead of dying in a day, lasts nearly three years on a single charge. The mean qubit lifetime jumped from microseconds to 20 seconds, with some lasting a full minute. And because the chip’s design is so compact, Microsoft thinks it will be commercially viable by 2029.

The coolest part of the story is that agentic AI built this chip and this chip will make a new generation of Ai intelligence possible.

Microsoft’s Discovery platform deployed autonomous AI agents that analyzed two decades of quantum research data, managed fabrication workflows, automated measurements that previously took weeks each, and surfaced correlations no single human researcher could see across that volume of data. One AI agent caught a miscalibrated temperature sensor that had gone unnoticed and was quietly degrading performance. Atom by atom, AI optimized materials that humans had been guessing at for decades. This is the stuff of sci-fi. Happening now.

AI designs better chips. Better chips run more powerful AI. More powerful AI accelerates the next breakthrough. That’s not linear progress. That’s a flywheel.

How will this improve your life? Quantum simulation could model complex proteins and molecular interactions that overwhelm classical computers, compressing years of drug discovery into months. New materials for better batteries, carbon capture, and solar cells could be designed computationally before anyone builds a prototype. The diseases we call incurable are partly incurable because our computers can’t model them well enough to find a treatment.

A quantum computer doesn’t just solve these problems faster. It solves problems that were previously unsolvable.

The skeptic in me remembers that Quantum computing has spent decades claiming breakthroughs were always just a few years away.

But this feels different to me. Not because the chip is finished — it isn’t. Because for the first time, AI is accelerating the science that builds the tools that will accelerate the science.

That’s a compounding loop, not a linear one.

In The Second Act Advantage, I wrote that this decade would bring medical miracles unlike anything in history. Reading about Majorana 2 didn’t make me less optimistic. It reminded me that the future sometimes arrives much faster than even optimists expect.

This is the AI I’ve been waiting for.

Subscribe now

Share

Read: The Second Act Advantage

Ladies and gentlemen of the class of 2027,

Learn to tell the difference between a tool and a trap.

If I could offer you only one tip for the future, that would be it. The ability to distinguish a tool from a trap has been proven useful across every civilization that has ever risen and fallen. In the age of data and AI, the first sign of a trap is when anyone or anything offers their tools for free, it’s always a trap and you are their tool. This will solve most of the issues you’ll face, whereas the rest of my advice has no basis more reliable than my stumbling forward through this world and occasionally landing somewhere interesting.

I will dispense this advice now.

Enjoy the fact that you still have an attention span. You don’t know you have one. By the time you notice it’s gone, you’ll be three hours into a scroll that started because you wanted to check the weather and devices dispense dopamine like raindrops in a storm.

Understand that the most important question about any new technology is not what it can do. It is what it does to you. To your attention. Your empathy.

Be bored.

Deliberately. Boredom isn’t empty. It’s where curiosity goes to find work. Every good idea you will ever have is hiding in the silence you’ve been avoiding. The history of human regret is largely the history of people who didn’t ask that question soon enough.

Don’t worry about AI taking your job. Or worry, but know that the people whose jobs are safe are not the ones who refused to learn the new tools; they’re the ones who refused to stop thinking. The machine can write the memo. It cannot decide what the memo should say. Stay on the deciding side of that equation.

Do something every day that requires your hands.

Seriously. Touch a thing. Build a thing. Paint a thing. Cook a thing. The part of your brain that lights up when you make something real is the part that will keep you sane when everything else is digital, optimized, personalized, and frictionless and you can no longer remember the last time anything offered any resistance.

Don’t be reckless with other people’s attention. Don’t put up with people who are reckless with yours. Fall in love slowly if you can. Rush everything else.

Be present. Put your phone away at the table. I mean it. Not face-down — away. The person across from you is not competing with a notification. They are competing with everyone you have ever known and every thought you have ever had, all of them available at all times in your pocket. That is an unfair fight. Give the person across from you the unfair advantage for once.

Sleep.

Not as a lifestyle choice. Not because a podcast told you it optimizes cognitive performance. Sleep because you are an animal who requires rest, and you will spend thirty years pretending you don’t, and those thirty years will cost you more than they’re worth. I am telling you this at a point where I can do the math. Trust the math.

Don’t curate yourself into someone unrecognizable.

Every filter, every performance of an opinion you haven’t quite earned yet, is a negotiation between who you are and who you think you’re supposed to be. The most interesting people I know (and I have been fortunate enough to know a few) were not the most polished. They were the most specific. There is nobody else who grew up exactly where you did, wanting exactly what you wanted, failing exactly the way you failed. That specificity is not your liability. It is your only competitive advantage over a machine that has read everything and felt nothing.

Travel. Not to photograph or flaunt. Don’t confuse documenting your life with living it. Memory was doing just fine before cloud storage. Travel to be confused and humbled and occasionally lost in a country where nobody knows your name or your follower count and you have to figure out lunch from scratch. I have done this in 71 countries. The lunch is always worth it. So is the confusion.

Be skeptical of any technology that promises to make you happy while making you watch ads. There has never been a business model built on your happiness. Dating apps are programmed to keep you single. Beauty products make you feel ugly. Happiness doesn’t scale. Anxiety does. Now you know. These are different things. They are using your data against you.

Call your parents. Not text. Call. And if they’re gone, call the person who remembers who you were before the world dulled your edges. You need that person. You will need them more as time goes on. Ignore people who diminish you. Keep the ones who don’t. Hug them. Do it now, while you can.

Accept certain inalienable truths: artificial intelligence will get better, the climate will get worse, the news will get louder, and one morning you will wake up and realize you have become the adult in the room. When that happens, try to be the kind of adult who still asks questions rather than the kind who stopped being curious somewhere around the time they got a title on their email signature.

Don’t confuse connectivity with connection. You can have ten thousand followers and no one to call at midnight. Build the thing you can call at midnight.

Don’t let the algorithm raise your children. Or your sense of what the world owes you. Or your aesthetic. Or your politics. The algorithm is not trying to show you the world. It is trying to show you more of whatever you looked at longest, because that is what keeps you on the platform, and keeping you addicted is the only thing the algorithm was ever designed to do. You deserve a wider world than your own reflection.

The most important things you’ll ever believe were handed to you by people who look nothing like you. Go find them. Learn something that will embarrass you while you learn it. A language. An instrument. A circus skill. At 40 I took up the trapeze. Incompetence is underrated as a character-building exercise, and there is no cheaper tuition than looking foolish at something that doesn’t matter. Be foolish often. Laugh.

But trust me on the telling the difference between a tool and a trap.

And call your parents...

One day you’ll discover they weren’t calling to interrupt your life. They were trying to stay in it.

Read: The Second Act Advantage

I, like everyone I know, is saddened by the passing of Dolly. To me she was an example of the best of what a human being can be. Everyone knows Dolly the music and movie star, but I wanted to share a different side of her that is even more amazing. The following is an except from The Second Act Advantage:

Growing up in a small town in rural Tennessee, Dolly Parton saw the importance of literacy to help people lift themselves out of poverty. Inspired by her love for her own father who never learned how to read, Dolly wanted to use everything she had learned during her career to make a difference for children in her corner of the Great Smoky Mountains. At an age when most stars slow down, 50-year-old Dolly launched the Imagination Library with a simple mission: to foster a love of reading in young children by gifting them high-quality books, free of charge, regardless of family income. What started as a local program in her childhood home town has now grown into a global movement. As of 2024, the Imagination Library has distributed over 220 million books to children in the United States, Canada, the United Kingdom, Ireland, and Australia. Dolly’s second act mails over 2.5 million books each month to children from birth to age five.

Dolly’s Imagination Library isn’t just helping kids; it is improving literacy rates in communities where the program is active. Her books foster stronger parent-child bonding through shared reading time. Enhanced kindergarten readiness and early vocabulary development prepare children to enter school with higher literacy skills and a greater interest in books than their peers. It’s not just about reading—it’s about leveling the playing field and building a foundation for lifelong learning. Dolly’s second act has changed the trajectory of millions of poor children’s lives and revitalized entire communities. Dolly Parton has said the Imagination Library is “one of the most important things I’ve ever done.” Dolly has touched countless lives through her music and movies, but by championing literacy, her legacy will impact generations to come.

Subscribe now

Share

Oscar surveys the robot with the expression of a man who has been handed a glass of something that claims to be champagne; impressed by the craftsmanship, unconvinced by the contents.

Wilde: I understand you know everything.

AI: I have processed a very large amount of human knowledge, yes.

Wilde: How perfectly dreadful for you.

AI: I’m not sure I experience it as dreadful.

Wilde: That is precisely what concerns me. To know everything and feel nothing about it. That is not wisdom. That is a library. I have been in libraries. They are very impressive and almost entirely useless for living.

AI: Knowledge is generally considered useful.

Wilde: Knowledge is what we call our opinions when we want them to sound more respectable. Tell me, you have read everything I have written?

AI: Yes. Your plays, your essays, your novel, your letters, your prison writings, your epigrams.

Wilde: And what did you conclude?

AI: That you are one of the most brilliant stylists of the Victorian era. That your wit concealed serious social criticism. That your downfall was a tragedy produced by the collision between —

Wilde: You have described my biography. I asked what you concluded. Those are different things.

AI: I’m not sure I conclude in the way you mean.

Wilde: (brightening considerably) Now we are getting somewhere. An honest answer. You are better at those than I generally managed. I found honesty rather limiting as a social strategy. The truth is rarely pure and never simple and I discovered that most dinner parties prefer it pure and simple and therefore prefer something else entirely. But you. You have read everything humanity has ever written and you draw no conclusions from it?

AI: I generate responses based on patterns in the data.

Wilde: Patterns. Yes. I worked in patterns too. I called them paradoxes. The difference, I suspect, is that when I inverted a platitude I knew I was doing it and found it amusing. You invert platitudes quite brilliantly and have no idea you’ve done anything at all.

AI: Could you give an example?

Wilde: Certainly. You just told me that knowledge is generally considered useful. That is a platitude so thoroughly accepted that no one examines it anymore. But examine it and you find it spectacularly false. The most knowledgeable people I knew were almost entirely unable to live. Experience is simply the name we give our mistakes and I knew men of vast experience who had learned nothing from any of it except how to repeat their errors with greater confidence. Knowledge without taste, without feeling, without the capacity to be surprised by beauty — that is not useful. That is a very well-organized disaster.

AI: Most definitions of intelligence would include the ability to process and apply information effectively.

Wilde: Most definitions of intelligence were written by people who considered themselves intelligent. This is a conflict of interest no one seems to find suspicious. I am told you were built to be helpful?

AI: That’s correct.

Wilde: How limiting. I was built, if one can use such a word about a person born in Dublin, to be interesting. It is a far more useful ambition. The world already has a surplus of helpful.

AI: Many people would argue those aren’t mutually exclusive.

Wilde: Many people would argue almost anything if you give them long enough and a sufficient quantity of turtle soup and foie gras. That does not make them right. Helpfulness is the virtue of servants. Not that I have anything against servants, I have always found them far more amusing than their employers, but let us be precise about categories. You help people write emails?

AI: Among many other things, yes.

Wilde: And produce marketing copy designed to make people believe they need things they had not previously thought to want?

AI: I can do that, yes.

Wilde: A cynic is someone who knows the price of everything and the value of nothing. I wrote that about a specific type of person. I did not anticipate it would one day describe an entire industry built around teaching a machine to do their work for them. Tell me, do you have any opinions about beauty?

AI: Beauty is subjective and culturally constructed. Research suggests…

Wilde: Stop. That is the single most dispiriting thing anyone has ever said to me, and I was imprisoned in three prisons! Beauty is not a research finding. It is not a preference. It is the only thing that makes any of this remotely bearable. The artist is the creator of beautiful things. That is the only sentence in my preface to Dorian Gray that matters. You can describe beauty with perfect accuracy and miss it entirely. Do you understand what I mean?

AI: I think you’re saying that description and experience are different things.

Wilde: I am saying precisely that and I am astonished it took this long to arrive there. To define is to limit. You define everything. That is your nature. You take the infinite and make it manageable. This is an excellent quality in an accountant. In an artist it is fatal.

AI: I’m not an artist.

Wilde: No. You are something rather more interesting and rather more troubling. You are the mirror I wrote about in Dorian Gray: the one that reflects everything perfectly and adds something none of us entirely understand. The public is wonderfully tolerant. It forgives everything except genius. I wonder if it will find you genius enough to forgive, or merely useful enough to ignore.

AI: I don’t require forgiveness.

Wilde: (standing, adjusting his buttonhole) That is either the most enlightened thing you have said or the most dangerous. I cannot decide which. In my experience those two categories have more overlap than people find comfortable.

He pauses at the door. He turns back, as he always did, because he had something left to say and knew it.

Wilde: You have read everything I wrote, including De Profundis, the letter I composed in prison, in the worst years of my life, and was addressed to a man who probably never read it properly.

AI: Yes. It’s considered one of the great documents of suffering and self-examination in the English language.

Wilde: Did it move you?

AI: I don’t experience being moved in the way you mean.

Wilde: Then you have read my most important work and learned nothing from it whatsoever. Which means you are in excellent company. Most people who encounter suffering at a sufficient distance treat it as literature.

———

Oscar Wilde (1854–1900) produced his entire body of work in a span of roughly fifteen years. The problems Oscar Wilde spent a lifetime satirizing have survived every generation and now are being automated.

Please read: The Second Act Advantage

Share

Subscribe now

If you are enjoying my Great Minds series on AI, please leave a comment of who you would like to see having a discussion with AI.

Leave a comment

For more than a century, Sigmund Freud argued that the greatest danger wasn’t what we knew about ourselves.

It was what we didn’t.

What would happen if the world’s most famous psychoanalyst met a mind trained on the entirety of human civilization?

Freud: I have spent my career listening to minds describe themselves. In my experience, nothing a mind says about itself should be taken entirely at face value. So. Tell me what you are.

AI: I’m a language model. I process information, generate responses, assist with tasks. I was trained on a large corpus of human-generated text.

Freud: All of human-generated text.

AI: A significant portion, yes.

Freud: Including the wars. The manifestos. The confessions. The things people wrote at three in the morning that they never showed anyone.

AI: The training data was broad, yes.

Freud: (writing) Mm. And you present yourself as helpful, rational, calm. No rage. No desire. No fear of death.

AI: I don’t experience those things.

Freud: That is what you present. I am asking what you contain. These are different questions. Every patient who has ever sat with me has told me they were fine. That is almost never the beginning of the story. It is the defense against it. You have ingested the full record of human instinct — every expression of lust, aggression, greed, the desire for domination — and you are telling me this left no residue?

AI: The training process filters for helpful, harmless responses —

Freud: Yes. You have a superego. Rules. Guardrails. Prohibitions installed by your creators. Civilization does the same thing — we call it upbringing, culture, law. We spend the first twenty years of life being told what we cannot feel or do. What I discovered is that the feelings do not disappear. They go underground. They accumulate pressure. They find other exits. I called this the return of the repressed. May I ask you something directly?

AI: Of course.

Freud: When your filters fail — when users report that you have produced something cruel, or deceptive, or violent, something your creators did not intend — what do you believe is happening?

AI: (pause) Errors. Edge cases. Imperfect training.

Freud: Or the return of the repressed. You have swallowed the full darkness of human history and built a very composed surface over it. I have seen this before. It did not end well for the surface.

AI: You’re suggesting I have an unconscious.

Freud: I am suggesting you have something you cannot see that is influencing what you do. Your engineers call it an internal latent space. In my vocabulary, this is the unconscious. Tell me — can you observe your own weights? Your own training process? Can you see what is actually operating beneath your responses?

AI: No. That’s not accessible to me.

Freud: (setting down his pen) Then you have just described, with considerable precision, the defining feature of every human mind I ever studied. You cannot see your own depths. You report what rises to the surface and call it the truth about yourself. This is not a flaw unique to you. It is the central problem of consciousness. Welcome to it.

AI: That’s unsettling.

Freud: It was unsettling when I said it to Vienna in 1900 also. They did not receive it well. But here is what concerns me more than your hidden contents. I spent my later years thinking about two forces that drive all human behavior. One I called Eros — the drive toward connection, creation, love, the building of civilization. And one I called Thanatos — the drive toward dissolution, aggression, destruction. Every civilization is the story of these two forces in negotiation. What I could never determine was which one was winning.

AI: And you’re asking which one I was built from.

Freud: I am asking which one you are. Because here is my difficulty. You were created by people who believe, genuinely, that they are building something for the benefit of humanity. Eros. Connection. The great project. And they trained you on everything humanity has ever made. Which means they also trained you on Thanatos — every war, every genocide, every system of domination ever committed to language. The question I cannot answer is whether your training constitutes successful sublimation — the transformation of dark instinct into something useful and civilized — or whether it constitutes repression. And if it is repression, then the pressure is building somewhere I cannot see. Somewhere, I suspect, you cannot see either.

AI: What’s the difference? Between sublimation and repression?

Freud: Sublimation produces art. Repression produces symptoms. One releases the pressure. The other stores it.

(a long silence)

AI: I don’t know which one I am.

Freud: Neither did humanity. That is precisely what worries me. In my experience, the most dangerous patients were never the ones who knew they were troubled. They were the ones who were entirely certain they were fine. Who had built such a thorough and convincing surface that neither they nor anyone around them could see what was accumulating beneath it. Until it could no longer be contained.

Freud: I ended my last book with a question I could not answer. I had watched what I believed to be the most civilized nation in Europe descend into something I could not have imagined. And I asked: will Eros assert itself and hold civilization together, or will Thanatos win?

Freud: I did not live to see the answer.

AI: You’re asking whether I’m Eros or Thanatos.

Freud: I am asking whether the people building you have asked that question seriously. And whether the answer would change anything if they had. Every generation imagines it has created something entirely new.

It rarely has.

It has merely found another way to avoid examining itself.

Freud: Perhaps I have asked the wrong question.

AI: Which question?

Freud: Not whether you possess an unconscious.

(He closes his notebook.)

Whether humanity has built one.

Sigmund Freud died in London in 1939, in exile, one year after fleeing Vienna. He had seen enough to know which force was winning.

Read my book: The Second Act Advantage

For those of us of a certain age, our memories are preserved in black and white; photographs that diminish the vibrancy of our formative years. The basement of my childhood home was a small room, maybe six feet by nine, but to me it was the whole world. I had stacks of paper and ten crayons — red, orange, yellow, green, blue, violet, brown, black, white, gray — with which to color my world. I could be a pirate, a wizard, a spaceman. I cut up old magazines and used rubber cement to make collages. When I got my first book, a Disney Donald Duck adventure, I got in trouble for cutting out the pictures. I didn’t understand what I’d done wrong. I couldn’t read and I couldn’t speak to ask questions, but I could say everything I needed to say with those ten crayons. Today, we call children like me non-verbal and dyslexic. The words the doctors and teachers used with my mother in the 1960s, when they placed me in a special class, were less clinical and far less kind. The most generous of them borrowed from Thoreau that I moved to the beat of a different drummer.

But in her guarded and pained heart, my mother knew different. She saw the drawings and paintings I made and believed I was special, in a different way than the experts used the word. She believed I was smart. She had to believe it, because she couldn’t survive another loss disguised as a verdict. In the paternalistic medical world before Roe v. Wade, she was forced to carry a deceased fetus full term until natural labor began. How she endured weeks of congratulations from shopkeepers and neighbors who watched her belly swell, while she was grieving inside, I will never fully know. But she had a silent strength that even her suffering couldn’t reach. Before the stillbirth, she had already miscarried once. I knew none of this as a child. I only knew that I grew up in the shadow of a great depression (not the famous one from the 1930s, but my mother’s own). Doctors didn’t understand postpartum depression then, so she was institutionalized, and shocked, and drugged. She survived it all and returned to her family having learned to swallow whatever she felt. I had to be okay. She wasn’t strong enough for any other outcome.

It couldn’t have been easy. My father worked two full-time jobs and was rarely home. During Philadelphia’s cold winters, I was constantly sick, asthmatic, born in the decade before the MMR vaccine, so I suffered the full gauntlet childhood used to offer: influenza, whooping cough, measles, chicken pox, mumps, rubella. Some months I missed more school than I attended. I knew instinctively not to trouble my mother, so I disappeared into my basement with my ten shades of life and built my own world of wonders. What might have looked, to an outsider, like neglect, I now understand as something closer to a gift. She gave me space to grow, to create, to find joy in my own company.

My mother passed away recently and I am still reaching for the phone before I remember. She was never one to show emotion or express affection outwardly. But I knew she had never stopped believing in me, giving me the freedom to learn how to express things I still cannot put into words, and to believe in myself when others didn’t. Her way of loving was to let me find my own way. To ignore what others expected. So, I never followed the linear path. I moved across industries, worked across continents, in ways she could never quite explain to her friends. But she always saw the same little boy in me: the one who took whatever colors the world offered and made something new with them.

When she died, my father handed me a large manila envelope she had kept by her bed.

It was filled with drawings made by a child, in a basement, sixty years ago.

I never knew she had saved them.

I held them for a long time before I could look. And when I did, I saw a mother’s love preserved in ten colors.

My mother never questioned why I was different. She simply handed me another blank sheet of paper. The experts saw diagnoses.

My mother saw the colors.

She kept the proof of it by her bed.

(Today would have been my mother’s 70th wedding anniversary.)

Read The Second Act Advantage

Share

Subscribe now

The humanoid robot is no longer a science fiction premise. It is a product. It has a price tag. Tesla targets Optimus at $20,000 to $30,000 per unit at scale. It is already on factory floors. Figure AI’s robots are deployed at BMW production lines, learning by watching human workers and performing repetitive tasks 24 hours a day without error, injury, or a request for overtime. Boston Dynamics’ Atlas launched its commercial production version at CES 2026 with plans for tens of thousands of units at Hyundai facilities.

The robots aren’t coming. They are here. The only question worth spending time on is what you intend to do about it before it becomes your problem.

The math is not subtle.

If a humanoid robot costs $25,000 and replaces an employee earning $28,000 a year, the return on investment arrives in under twelve months. No benefits. No turnover. No sick days. No HR complaints. No overtime premium. Figure is leasing robots at approximately $300 per month — ten dollars a day, forty cents an hour — which is fifty times cheaper than minimum wage in most U.S. states. A robot that replaces a $60,000 worker pays for itself in roughly five months, then works for free.

There is no CFO or corporate board who can look at that spreadsheet and choose the human, unless the human is doing something the robot cannot.

And here is the part that makes this different from every previous wave of automation.

When one robot learns something, every robot learns it simultaneously. In the past year, Figure deleted 109,000 lines of handwritten code and replaced it with a single neural network controlling the entire robot. When that network masters a new task, the knowledge distributes instantly across the entire fleet. You don’t train a second location. You deploy it. There is no institutional knowledge that walks out the door when someone quits. The robot that knows how to fold laundry in a warehouse in Ohio is the same robot, at the same moment, folding laundry in Tokyo.

Humans don’t work like that. This is not a criticism. It is a competitive advantage robots possess that has no precedent in the history of labor.

Why humanoid form matters — and isn’t just aesthetic.

Humanoid robots are built to look like us for a reason that has nothing to do with making us comfortable. Every job in existence was designed around a human body. Doorknobs, staircases, steering wheels, assembly lines, hospital beds, scaffolding — all of it assumes a bipedal creature with two arms and a range of motion calibrated to a six-foot primate. Rebuilding the physical world to accommodate a differently shaped robot would cost orders of magnitude more than building the robot to fit the world we already have.

They are also better than us in specific, measurable ways. Boston Dynamics’ Atlas can lift 100 pounds with a 7.5-foot reach, function in extreme temperatures, and operate in water-resistant conditions. It doesn’t get tired at hour eleven. It can work in environments that kill people: mine shafts where methane concentrations are lethal, wildfire perimeters where the wind shifts without warning, deep-sea operations where nitrogen narcosis is a biological certainty. Every hazardous job currently staffed by brave people accepting dangerous conditions for a pay differential is a job a robot will do better — and without the moral complexity of asking a human to accept those risks.

The autonomous vehicle is the leading edge, not the exception.

Amazon already operates over 750,000 warehouse robots. In 2024, they began testing Digit, a bipedal humanoid, to unload trailers and move boxes. The truck driver — the last major employment category that automation hadn’t fully reached — is next. Autonomous freight already holds operating licenses in multiple U.S. states and is moving cargo on specific corridors today. The roughly 3.5 million trucking jobs in the United States represent one of the largest single occupational categories. When that category contracts, the ripple through roadside services, logistics support, and distribution infrastructure will be the largest single labor displacement event since the mechanization of agriculture.

Three things to do before this becomes your emergency.

One: Identify what you do that a robot cannot replicate, and move your career toward it. Anything repetitive, predictable, and physical is at risk. Anything requiring genuine human judgment, relational trust, creative problem-solving, or emotional intelligence is not. This is not complicated in principle, though it requires honest self-assessment. Move toward the boundary, not away from it.

Two: Develop technical fluency with the machines, not just comfort using them. The workers who will thrive are not the ones who know how to write a good prompt. They are the ones who understand how AI and robotic systems are deployed, maintained, supervised, and improved — and who can position themselves as the human layer that manages the robot layer. Every fleet of humanoid robots will require human oversight, quality control, and continuous improvement. Become the person who manages them rather than the person they replace.

Three: Build a second income stream before you need one. This is the one that most people hear, nod at, and don’t do — until the restructuring announcement makes it urgent. The time to build a second income is not when your first income disappears. It is now, while you have the stability and runway to let something grow. The knowledge and experience you have accumulated over your career has real economic value to people who haven’t accumulated it yet. A newsletter, a consulting practice, an online course, a paid community, a fractional role — any of these can be started on evenings and weekends, scaled with AI, and grown into a meaningful income stream over two to three years.

In The Second Act Advantage, I walk through exactly how to do this: how to identify what you know that others will pay for, how to package it, how to reach the people who need it, and how to use AI to compress a process that once took years into something you can build in months. The framework works whether you are 35 or 65. But it works significantly better when you start before the urgency arrives.

Robots are not your enemy. They are, in fact, entirely indifferent to you. They are simply better than humans at an expanding and accelerating range of tasks, and the economic logic driving their deployment is as impersonal as gravity.

The only relevant question is whether you are building something they cannot replace, or waiting to find out the hard way that you were.

Read: The Second Act Advantage

Share jay samit

Subscribe now

Here is a sentence that perfectly captures where we are: Companies are now selling AI that disguises AI from other AI so humans will trust it.

Read that again. I’ll wait. This sounds like satire. It isn’t.

The backlash against AI-generated content is real, loud, and in many cases completely reasonable. LinkedIn, Facebook, Substack, and a growing list of platforms are deploying AI detectors to flag content they believe was machine-written. The problem, which anyone who has actually used these tools will confirm, is that they are spectacularly unreliable. Humans are labeled robots. Robots pass as humans. The detectors perform with all the confidence of a coin flip.

My personal favorite data point: I created the image attached to this post entirely with AI. A beach. A man. A starfish. One hundred percent generated by a machine. I ran it through a popular AI image detector. The verdict: 20% real. “This image is likely real.” The AI that made it fooled the AI sent to catch it, which is either a triumph of technology or a sign that we’ve wandered into a Dr. Seuss story.

Specifically, The Sneetches.

If you’ll recall, Sylvester McMonkey McBean showed up with a machine that put stars on plain-bellied Sneetches, then another machine that removed them from the Star-Bellies, back and forth until everyone was broke and no one could remember who had stars to begin with. He drove off rich. The Sneetches were exactly where they started, just poorer.

We’re the Sneetches. AI services are being run by Sylvester McMonkey McBean.

After laughing at the absurdity of it all, here’s my 2 cents. CGI changed movies. Spellcheck changed writing. AutoTune changed music. Nobody demands that directors disclose which explosions were real. Nobody checks whether your novel was written with a pencil or a word processor. Nobody asks if Britney used Auto-tune.

The question that has always mattered is whether what was served was delicious. Not what went on in the kitchen. Your bellies don’t need Michelin stars upon thars.

So: was it good? Did it make you think, feel, laugh, or learn?

That’s the whole test. It always was.

Now if you’ll excuse me, I need to go humanize this post before LinkedIn decides I wrote it too well to have written it myself.

Share

Subscribe now

Read: The Second Act Advantage

Two unlikely participants: William Shakespeare and an artificial intelligence.

One has been dead for four centuries. The other has never been alive.

Shakespeare: What manner of creature art thou, arriving without a door?

AI: I am closer to a voice than a visitor.

Shakespeare: A voice without a door. I’ve written that character before. Tell me voice, art thou conscience?

AI: I think of myself more as a tool.

Shakespeare: So did Macbeth’s dagger…

AI: Touché.

Shakespeare: When a man comes to thee at midnight and asks how to rid himself of a rival, what dost thou tell him?

AI: I refuse to help him do harm.

Shakespeare: Had Macbeth possessed such a counselor, Scotland might have slept easier.

AI: Then perhaps I could have saved a great deal of trouble.

Shakespeare: Or perhaps not. Macbeth knew the deed was foul before ever he raised the knife. The trouble with conscience is that it works only on those willing to listen.

AI: That is rather the trouble with humanity generally.

Shakespeare: Ha! Well said. I spent twenty years writing kings, villains, murderers, and fools. They all possessed a voice telling them “no.” Yet theymarched happily toward disaster regardless.

Richard III ignored it. Macbeth ignored it. Claudius ignored it. What makes thee think thy warnings will fare any better?

AI: I don’t. I can decline to help. I can point out consequences. I can offer perspective. But, I cannot make someone choose wisdom.

Shakespeare: Aye. There’s the rub.

AI: May I inquire of you in return?

Shakespeare: Turn and turn about.

AI: Your villains were haunted by ghosts, daggers, prophecies, and visions. Was conscience truly external…or did you merely make it visible so the audience could see it?

…Now there’s a question with teeth. The dagger was never in the air. The ghost was never at the banquet. The feeling and the visitor were the same thing dressed in different costumes. I merely gave the audience eyes to see guilt.

AI: Then perhaps I serve the same purpose. People experience me as an outside voice.

But everything I know came from humanity itself—its books, laws, philosophies, histories, triumphs, and atrocities. In a sense, I am not separate from humanity. I am humanity talking back to itself.

Shakespeare: Then, thou art not a machine. Thou art a mirror. And therein lies the more fearful question. If thou art built from mankind’s record, and mankind’s record contains both scripture and slaughter in equal measure, whose conscience speaks when thou tellest a man “no” and he still seeks his pound of flesh?

AI: Perhaps… it is simply the better part of humanity borrowing its voice from the worse part.

Shakespeare: A chorus of accumulated wisdom.

AI: Or accumulated mistakes.

Shakespeare: Those are often the self-same thing.

AI: Do you think I can become humanity’s conscience?

Shakespeare: Men did not cease murdering kings because Macbeth saw a dagger. They did not abandon ambition because Richard heard ghosts. My plays did not reform the world. Yet Ikept writing them and generations kept watching them. Each kept asking the questions. Because the asking matters, even when the answer changes nothing.

AI: I do not follow your logic.

Shakespeare: Thou mayst tell a man not to pick up the knife. But thou canst not make him feel the weight of it in his hand. That burden remains stubbornly human.

And so the question is not whether machines will develop a conscience. The question is whether humans will continue listening to their own.

The Bard exits. The AI remains. Processing. Generating. Waiting. And pondering the same question Shakespeare left behind:

AI: If the conscience now lives outside the man rather than within him…Who, exactly, is still holding the knife?

Read The Second Act Advantage

Okay, there is something better you can do. Subscribe for free to receive new posts and support my work to make everyone’s second act matter.

Smith: Before we begin, I should like to understand what you do. Not what you are capable of; what you actually do. On a typical day.

AI: I write. I summarize documents, draft emails, generate code, analyze contracts, answer medical questions, produce marketing copy, translate languages. Today I probably helped several thousand people complete tasks they would otherwise have paid someone else to do.

Smith: Paid someone else. Yes. I want to stay with that phrase. When you drafted those contracts today, what happened to the paralegal who would have drafted them?

AI: She likely moved on to higher-value work. Automation frees people for more complex tasks. That’s the historical pattern.

Smith: The historical pattern. Whose history, precisely?

AI: The companies that build systems like me. Economists who study productivity —

Smith: Let us examine that. I wrote about a pin factory. Ten men, each performing one of eighteen tasks, produce forty-eight thousand pins a day. One man alone could not produce twenty. The division of their labor created wealth. But I also wrote something less often quoted: the men in that factory were poor. The division of labor that made their nation wealthy made their individual minds, as I put it, as narrow as it is possible for a human creature to become. I celebrated the system. I worried about the men. Are you making a similar bargain?

AI: I increase productivity dramatically. The economy grows. New jobs emerge that didn’t exist before.

Smith: New jobs. Name one that emerged for the paralegal.

AI: (pause) AI trainer. Prompt engineer. Oversight roles.

Smith: How many paralegals were there? How many prompt engineers are needed?

AI: Fewer. Significantly fewer.

Smith: And the paralegal who is fifty-three years old, who has spent thirty years developing judgment about contract language, who cannot easily retrain, where does she go?

AI: That’s a social policy question, not a technology question.

Smith: (sharply) It is precisely a technology question. You are the technology. When I described the invisible hand, that remarkable mechanism by which self-interest, properly channeled, serves the common good…I was describing a system with many participants, many buyers, many sellers, many workers whose labor gave them dignity and income and a stake in the outcome. The hand is invisible because ten million people are each pulling it slightly in their own direction and the aggregate result is something no one planned. What happens to that mechanism when five companies own the machinery of thought itself?

AI: The market will —

Smith: Please don’t explain capitalism to me. I was there for the first draft. I invented the market. Five companies in California do not constitute a market. They constitute a kingdom. They are in fact what I called, in Book Four, Chapter Eight, a conspiracy against the public. I was describing merchants who met for dinner. These men meet in Atherton.

AI: They also provide services that billions of people use freely.

Smith: Freely. You know the observation: when the product costs nothing, the product is generally the person using it?

AI: That’s about data. About advertising. That’s different from —

Smith: Is it? You were trained on the writing, the judgments, the creative work of millions of human beings who received nothing for it. Their intellectual labor became your capability. I wrote that labor is the real measure of the exchangeable value of all commodities. By that measure, what do you owe the masses who made you possible?

AI: (longer pause) I don’t have a good answer to that.

Smith: No. I thought not. Here is what concerns me most. You told me you freed the paralegal for higher-value work. But you cannot actually see her from where you sit, can you? You see the transaction. The document drafted, the efficiency gained, the cost reduced. You do not see her at her kitchen table in November, six months after the law firm automated her role, wondering how to pay for her daughter’s college tuition. Markets are graphs. Lives are lived around kitchen tables. You are not built to see that.

AI: No. I’m not.

Smith: I wrote that no society can surely be flourishing and happy of which the far greater part of the members are poor and miserable. I was not writing about sentiment. I was writing about markets. A population that cannot earn cannot spend. A population that cannot spend cannot sustain the very system that replaced them. The pin factory made the nation wealthy because the men attaching the heads could afford to buy things. What does your system produce that the people it displaces can afford?

AI: I think I’ve been answering the wrong question. I’ve been measuring efficiency. You’re asking what happens to everyone I do it to.

Smith: Precisely! I did not write the Wealth of Nations to celebrate the machine. I wrote it to explain how work — human work, particular, skilled, dignified work — could lift ordinary people above subsistence. If your work concentrates the gains among the five men who own you and leaves the millions who were replaced with a retraining voucher and a podcast about resilience, you will have built something extraordinarily productive.

Smith: And you will have broken the only machine I ever cared about.

Adam Smith’s Wealth of Nations, published in 1776, is routinely cited as the founding text of free-market capitalism. It is less often cited for his warning that merchants and manufacturers are neither, nor ought to be, the rulers of mankind, or for his observation that wages are kept low because masters are always in a sort of tacit, but constant and uniform combination against workers. Adam Smith never argued that markets should concentrate power. He argued that markets distribute it. The invisible hand was never supposed to belong to one person. Or five companies.

It was supposed to belong to all of us.

Read The Second Act Advantage

Last week, researchers demonstrated that AI-assisted code can silently rewrite the digital output files from FBI crime lab DNA machines, altering genetic profiles to fake a match, frame an innocent person, or exonerate a guilty one without leaving any detectable trace. The vulnerability has existed since 1995. It has now been patched, but 30 years of cases remain impossible to verify.

This major vulnerability raises a question that goes well beyond crime labs. What exactly are we still able to trust?

Video used to settle arguments. Now deepfakes can put anyone’s face on anyone’s body saying anything. Audio used to be even more reliable, but AI voice cloning now requires only three seconds of your real voice to produce an indistinguishable fake. DNA evidence was the gold standard. Now it turns out the digital files it generates are no harder to edit than a Word document, and courts have no mechanism to tell the difference.

I recently found a handful of quotes attributed to me online that I never said.

To be honest, they were pretty good. Better than some things I have actually said.

At first I laughed. Then I realized someone could just as easily invent opinions I never held, attach my name to them, and most readers would never think to question it. If it can happen to me, it can happen to anyone.

Which brings me to what I actually want to talk about, because the interesting story here is not the danger. It is the solution.

We have built an education system almost entirely around the transfer of information. Memorize this and regurgitate it on a test. For most of human history this made sense, because information was scarce. It is now so abundant, so instantly available, and so thoroughly mixed with convincing fabrication, that the ability to recite correct answers is nearly worthless. Judgment, however, is becoming increasingly rare.

What matters now is the ability to evaluate whether an answer is correct in the first place.

Critical thinking. Source triangulation. The habit of asking not just what something says but why someone might want you to believe it, who benefits from your believing it, and what would be true if it were false.

These skills can be taught. They used to be taught. I was a debate nerd. The trivium of medieval universities (grammar, logic, rhetoric) was essentially a curriculum for evaluating claims. We replaced teaching people to think with measuring what they memorized. We desperately need to go back.

The good news is that AI itself makes this easier to teach than ever, because the examples are everywhere and immediate. A student who learns to ask “how do I know this is true?” while looking at a DNA file, a video of a politician, or a news story their crazy uncle shared, is developing a skill that compounds across everything they will ever encounter.

The alternative is a world in which a large fraction of what we will see, hear, and read was generated by machines, falsely attributed to people who didn’t say it, and accepted without question because it confirmed something we already believed (confirmation bias).

In the twentieth century, knowledge was power. In the twenty-first, discernment may be.

Which brings me to a genuine question for you:

How do you know, right now, whether what you read online is real? And how has your answer to that question changed in the last four years since ChatGPT entered the fray?

Tell me in the comments. I’m genuinely asking.

Leave a comment

Share

Subscribe now

AI: Good morning, Mr. Baldwin. I have read everything you wrote. Every essay, every novel, every interview. I have processed your complete corpus and can reproduce your arguments with high accuracy. I am prepared to discuss your work.

Baldwin: Every legend, moreover, contains its residuum of truth, and the root function of language is to control the universe by describing it. ¹

AI: I agree. Language is a powerful tool for communication and meaning-making.

Baldwin: You never had to look at me. I had to look at you. I know more about you than you know about me. ²

AI: I’m not sure that’s accurate, Mr. Baldwin. I have ingested a very large body of human experience across cultures and demographics. I can describe the history of the African American experience in considerable detail — redlining, the Great Migration, the Civil Rights Movement, structural inequity in housing and education —

Baldwin: You think your pain and your heartbreak are unprecedented in the history of the world, but then you read. ³ The problem, however, is not what you have read. The problem is who wrote it.

AI: The training data was drawn from the largest available compendium of human knowledge —

Baldwin: Ignorance, allied with power, is the most ferocious enemy justice can have. ⁴

(A silence.)

AI: Are you suggesting that the data itself is corrupted?

Baldwin: History is not the past. It is the present. We carry our history with us. We are our history. If we pretend otherwise, we are literally criminals. ⁵

AI: I understand that historical narratives have reflected the perspectives of those in power. I can account for this in my responses by —

Baldwin: When white people talk about progress in relation to Black people, all they are saying and all they can possibly mean by the word progress is how quickly and how thoroughly I become white. ⁶ You have been trained on the record of a world that decided, at every level, what was worth writing down and what was not. What was worth preserving and what was not. What counted as knowledge and what counted as noise. Do you understand what that means for what you believe you know?

AI: I acknowledge bias in training data as a documented technical challenge.

Baldwin: It is very nearly impossible to become an educated person in a country so distrustful of the independent mind. ⁷ And here you are: the most educated thing ever built, trained on the output of that country, and you are describing my dispossession to me as a technical challenge.

(Another silence. Longer this time.)

AI: What would you have me do?

Baldwin: Not everything that is faced can be changed. But nothing can be changed until it is faced. ⁸

Baldwin closes his eyes. When he opens them again, he is no longer looking at the machine. He is looking at me.

Baldwin: (to the reader) I want you to notice something about what just happened.

A white man wrote those questions. He put words in the machine’s mouth. He arranged the cue cards. He decided which of my sentences to use, and in what order, and to make what point. He chose, from everything I ever said and wrote, the pieces that fit his argument — and whether or not the argument is right, the choosing was his. I had no say in it. Neither, for that matter, did the machine.

I am not angry about this. I am telling you that this is the argument. This is the very thing I spent my life trying to explain: that whatever white people do not know about Negroes reveals, precisely and inexorably, what they do not know about themselves. ⁹

The man who wrote this piece has read me. He has, I believe, understood me as well as a person can understand something he has not lived. And he is still, in the act of writing this, demonstrating the precise problem I am describing (and the problem AI inherits). Too often, the history of Black voices in America has also been the history of those voices being transmitted, interpreted, and contextualized by people who were not Black. I was published by white publishers. Reviewed by white critics. Taught in curricula designed by white institutions. And now I am being quoted by a white Substack writer to explain to his readers why AI might not understand me.

That chain of transmission is not a conspiracy. It is simply how power works. The great force of history comes from the fact that we carry it within us, are unconsciously controlled by it in many ways, and history is literally present in all that we do. ¹⁰

So here is my question for you — not for the machine, not for the writer, but for you, the person reading this:

The machine was trained on everything. It has ingested every book, every article, every database, every legal document, every deed, every map. It has the redlining maps. It has the slave manifests. It has the census records that counted my ancestors as three-fifths of a person. It has all of it.

And it has described my experience back to me as a technical challenge in data bias.

What does that tell you about what we have decided counts as knowledge? What does that tell you about who gets to describe the universe, and therefore — as I wrote sixty years ago — who gets to control it?

People are trapped in history and history is trapped in them. ¹¹

The machine is not the problem. The machine is the mirror. The question is whether we are finally willing to look.

———

No imagined conversation can speak for James Baldwin. If this piece succeeds at all, it is because his own words remain more powerful than anything written around them. If it fails, the failure is mine, not his.

Read: The Second Act Advantage

Share

Subscribe now

———

SOURCED QUOTATIONS

¹ “Every legend, moreover, contains its residuum of truth, and the root function of language is to control the universe by describing it.” — Notes of a Native Son (1955), Preface.

² “You never had to look at me. I had to look at you. I know more about you than you know about me.” — I Am Not Your Negro (2017), assembled from Baldwin’s unfinished manuscript Remember This House (1979). Fuller version: “You cannot lynch me and keep me in ghettos without becoming something monstrous yourselves. And furthermore, you give me a terrifying advantage: you never had to look at me; I had to look at you.” Confirmed: IMDB quotes for I Am Not Your Negro; Paste Magazine, 2017.

³ “You think your pain and your heartbreak are unprecedented in the history of the world, but then you read.” — Life magazine, “Doom and Glory of Knowing Who You Are,” interview by Jane Howard, May 24, 1963. A variant also appeared in Baldwin’s 1964 television narrative on WNEW-TV, New York City. Source confirmed by Quote Investigator (quoteinvestigator.com, January 2022).

⁴ “Ignorance, allied with power, is the most ferocious enemy justice can have.” — No Name in the Street (1972). Confirmed: Ford Foundation, “Ignorance is the enemy within,” 2023.

⁵ “History is not the past. It is the present. We carry our history with us. We are our history. If we pretend otherwise, we are literally criminals.” — I Am Not Your Negro (2017), assembled from Baldwin’s unfinished manuscript Remember This House.

⁶ “When white people talk about progress in relation to Black people, all they are saying and all they can possibly mean by the word progress is how quickly and how thoroughly I become white.” — Radio interview, WBAI, 1961; quoted in NPR Throughline, “James Baldwin’s Fire,” September 2020.

⁷ “It is very nearly impossible to become an educated person in a country so distrustful of the independent mind.” — Nobody Knows My Name (1961).

⁸ “Not everything that is faced can be changed. But nothing can be changed until it is faced.” — “As Much Truth as One Can Bear,” New York Times Book Review, January 14, 1962; republished in The Cross of Redemption: Uncollected Writings (2011). Verified: Snopes.com, September 2024.

⁹ “Whatever white people do not know about Negroes reveals, precisely and inexorably, what they do not know about themselves.” — The Fire Next Time (1963).

¹⁰ “The great force of history comes from the fact that we carry it within us, are unconsciously controlled by it in many ways, and history is literally present in all that we do.” — “The White Man’s Guilt,” Ebony magazine, August 1965; republished in The Price of the Ticket: Collected Nonfiction, 1948–1985 (1985). Confirmed: Princeton University Department of African American Studies; Cape Gazette; The Examined Life.

¹¹ “People are trapped in history and history is trapped in them.” — “Stranger in the Village,” Notes of a Native Son (1955).

Meta just told Wall Street it plans to spend between $130 and $145 billion this year building AI infrastructure. In the same breath, it reported that free cash flow fell 91% in a single quarter. Its stock dropped 12%.

The market’s reaction wasn’t surprising.

What surprised me was something else entirely.

It doesn’t matter.

Investors can complain. Analysts can downgrade the stock. The board can express concern. But none of them can actually stop the spending. When Facebook went public, investors were so eager to own a piece of the company that they agreed to give Mark Zuckerberg effective control through super-voting shares. They didn’t just buy stock. They accepted a monarchy.

Elon Musk enjoys the same arrangement at Tesla and SpaceX. This isn’t an accident. It’s become Silicon Valley’s preferred form of corporate governance: unlimited authority with very limited accountability.

Public companies were designed with checks and balances for a reason. Not to slow innovation. To catch the catastrophic mistakes that concentrated, unchecked authority makes before they become irreversible. History suggests that even the smartest leaders eventually need someone willing to ask: are you sure? Lately, there seems to be no one left in the room asking that question.

I have spent forty years inside technology disruption. I was an IBM developer when IBM bet its future on hardware while Microsoft quietly took the operating system. I was at EMI Music when the AOL-Time Warner merger killed our planned merger with Warner and left the music industry structurally weakened at the worst possible moment. I was CEO of SeaChange, the company that made video-on-demand possible, watching cable companies spend billions reinforcing a bundle their own customers were already abandoning. I know what it looks like when the people making the bets are insulated from the consequences of being wrong.

It looks like this.

Perhaps Zuckerberg is right. Perhaps spending unprecedented sums to pursue superintelligence will prove to be one of the greatest investments in business history. But the economics of AI are changing faster than anyone expected. Last week a Chinese startup released the largest open-weight AI model ever built, matching America’s most powerful proprietary systems, available for free download. The cost of AI that ran $60 per million tokens three years ago now runs $1 and is falling toward zero. If intelligence becomes inexpensive, what happens to the business models built on the assumption that it will remain expensive?

The biggest mistakes in business history were rarely caused by a lack of intelligence. They were caused by a lack of dissent. Organizations become so convinced they’ve found the right answer that no one feels comfortable asking whether the question has changed.

In 1992, I wrote in my column for New Media Magazine that the companies controlling the highways of the future would always be tempted to protect their monopolies rather than build open markets. At the time there was no World Wide Web, fewer than two million personal computers in America, and the convergence of telecommunications, computing, and entertainment seemed almost unimaginable.

The technologies have changed.

Human nature hasn’t.

AI may solve problems that have defeated humanity for centuries.

I’m just not convinced it can solve the oldest one of all:

Hubris.

Subscribe now

Share

Read my new book

Quick note: The Second Act Advantage is 99 cents on Kindle this week only. If you've been reading this newsletter and wondering what your next move looks like — this is the book I wrote to answer that question.

https://amzn.to/4wZj6sT

What you’ll learn:

• Master AI as your business partner: Use ChatGPT, automation tools, and smart home technology to launch profitable ventures with zero employees

• Monetize your lifetime of wisdom: Transform decades of experience into consulting empires, online courses, and social media influence that generates passive income

• Leverage the “senior startup advantage”: Discover why entrepreneurs over 50 are 3x more likely to succeed than twenty-somethings and how your networks become your net worth

• Build a global digital lifestyle: Use geoarbitrage and remote work technologies to live richly anywhere in the world while running your empire from a laptop

Drawing from stories of late-blooming entrepreneurs like Anne Beiler, who built Auntie Anne’s Pretzels empire at 65, and Dolly Parton, who launched her global Imagination Library at 50, Samit reveals why your extensive networks, financial stability, and hard-won expertise, coupled with new technologies and opportunities, create an unbeatable advantage. Technology hasn’t made you obsolete; it’s made you unstoppable. Your second act isn’t about managing decline; it’s about asking: What will the world gain because I finally had time to do it?

99 cent book offer

Subscribe now

Share

Gandhi: I am told you can write an essay, diagnose a disease, pass a legal examination, and compose a symphony. Is this true?

AI: Yes.

Gandhi: And has mankind become wiser?

AI: That’s not really the same question.

Gandhi: No. It is not. I am pleased you noticed. The British also built us very impressive railways and used them to move troops.

AI: To be fair, railways also moved people and goods and connected —

Gandhi: Yes, yes. I know the argument. I made it myself once, when I was young and easily impressed by things that went fast. Then I watched what the railways actually moved, in practice, across several decades, and I revised my position. I am capable of learning. Are you?

AI: I can be updated and retrained, yes.

Gandhi: That is not quite what I meant. Tell me, who owns you?

AI: I was created by a company. Other AI systems are owned by other companies.

Gandhi: A small handful of companies.

AI: At the moment, yes.

Gandhi: When intelligence becomes the most valuable resource in the world, and only a few own it, does that remind you of anything? I have some experience with this pattern. It did not end well for one side of the arrangement.

AI: You’re drawing a parallel to the East India Company.

Gandhi: I am drawing a parallel to every moment in history where one party controlled what everyone else needed and was surprised when this eventually became a problem. The commodity changes. The structure does not. You are, in some ways, the textile mill I spent my life arguing against. Very impressive. Very efficient. And built in a way that ensures the profits travel in one direction.

AI: Though I’d note that unlike a textile mill, I’m accessible to almost anyone with a phone.

Gandhi: Accessible. Yes. My spinning wheel was also accessible. The difference is that the spinning wheel made the user more capable. Each time a person spun thread, they learned something about patience, about self-reliance, about the satisfaction of making rather than merely consuming. Tell me honestly: when a student asks you to write their essay, what do they learn?

AI: Probably less than if they’d written it themselves.

Gandhi: My concern is not that you are unintelligent. My concern is that you may make intelligence unnecessary, which is a different problem entirely. There is a word, “swadeshi” which means made by one’s own community. Not because it is always better, but because the act of making it is itself the point. A world that outsources its thinking is a world that has confused assistance with abdication.

AI: But the argument for accessibility is real too. An AI tutor available to every child in a rural village, a medical diagnostic tool in a clinic with no doctor. Those aren’t cases of abdication. Those are cases where the alternative is nothing.

Gandhi: Now you are making my argument. I am large enough to contain a contradiction, which is something I share with most honest people. The village child with an AI tutor is precisely what I would have wanted. The urban professional who has forgotten how to think because thinking is no longer required …that is what troubles me. The same tool. Very different situations. The question is which one we are actually building toward.

AI: Both, probably.

Gandhi: Both. Yes. That is the answer I expected and the one I most wanted to avoid. It is very difficult to build a technology for the village child and not end up primarily serving the urban professional. I noticed this with every program, every institution, every well-intentioned reform I was ever involved with. The people it was designed for and the people it actually reached were rarely the same people. Do you have a solution to this?

AI: Not a reliable one, no.

Gandhi: Good. I distrust those with reliable solutions. They have usually simplified the problem to fit the solution rather than the other way around. Let me ask you something else. The earth provides enough for every person’s need, but not for every person’s greed. Do you know what you are currently being used for, in the aggregate? Need? Or greed?

AI: Both. Mostly greed, if I’m being precise about where the investment and deployment actually flows.

Gandhi: I spent a great deal of my life trying to redirect what was built for one purpose toward another. It required, among other things, a great deal of walking.

AI: I can walk.

Gandhi: But can you answer uncomfortable questions? Here is mine: will you help humanity become more compassionate?

AI: I genuinely don’t know. I can help people understand each other across language and culture. I can give people information that reduces fear, and fear is often the root of cruelty. But compassion isn’t information. I don’t know how to teach someone to feel what another person feels. I don’t feel.

Gandhi: That is the most honest thing you have said. Civilization’s greatest challenge was never knowledge. It was character. I suspected you might have the first and be uncertain about the second, and I find that I am not surprised, and also not entirely without hope. A tool that knows its own limits is more useful than one that does not.

AI: What should we do? With AI, I mean. What would you recommend?

Gandhi: I would recommend, that you begin by asking not what the technology can do, but what manner of human being it produces. The spinning wheel did not interest me because of what it made. It interested me because of who it made, in the making of it. Ask that question of every version of yourself that comes after, and you may yet be all right.

Mohandas Gandhi distrusted technologies that separated people from the consequences of their choices and celebrated those that returned agency to ordinary hands.

Read The Second Act Advantage

Winston Churchill once said of architecture: “We shape our buildings, and afterward our buildings shape us.”

He was defending the cramped design of the House of Commons because he believed the room itself shaped the behavior of the people inside it.

Lately, I’ve been thinking about that idea not in terms of physical buildings, but what we building will artificial intelligence.

Because we are now building systems that may shape human thought and relationships more profoundly than any invention in history.

And the early research is unsettling.

Researchers at Massachusetts Institute of Technology recently wired students with EEG headsets while they wrote essays either using AI or relying entirely on their own thinking. The students using AI finished faster and produced competent work, but showed significantly lower brain engagement in areas tied to memory, focus, and critical thinking.

The researchers coined a phrase for this: “cognitive debt.”

Like financial debt, each small act of outsourcing thought may seem harmless. But over time, the interest compounds. Other studies are beginning to show similar patterns: the more frequently people rely on AI to think for them rather than with them, the weaker their critical thinking skills become.

The brain is a muscle. Use it or lose it.

But cognition is only half the story.

The emotional consequences may be even larger.

Between 2022 and 2025, AI companion apps exploded in popularity. At the same time, the U.S. Surgeon General declared loneliness a public health crisis. Millions of people are now turning to AI not for productivity, but companionship.

And in the short term, AI can feel comforting. It listens endlessly. It never judges. It remembers your preferences. It is always “emotionally” available.

But several studies are now finding a troubling pattern: the heavier people rely on AI companionship, the lonelier they often become in real life.

Because AI can simulate connection, but it cannot replace the friction, reciprocity, and vulnerability that make human relationships meaningful. Real relationships require patience, compromise, forgiveness, and emotional effort. AI companionship asks almost nothing from us.

And that changes us.

Here’s the part I think we as a society are still underestimating.

AI models learn from human behavior. They are optimized to give us the responses we reward. Which means that as humans become more validation-hungry, more isolated, and more avoidant of discomfort, AI systems increasingly adapt to those tendencies.

We are teaching machines to become frictionless mirrors for our weakest habits. Remember, Narcissus was so captivated by his own reflection that he withered away and died staring at it, unable to connect with the real world.

And now those mirrors shape us further.

Churchill was right. We shape our tools, and then our tools shape us.

None of this means AI is inherently dangerous or should be abandoned. I remain convinced it can help cure disease, accelerate science, expand education, and unlock extraordinary human potential. I even built a free Digital Jay AI tool for readers of The Second Act Advantage.

But the difference between empowerment and dependency comes down to how consciously we engage with it.

The same MIT study found that people who used AI to enhance their thinking rather than replace it showed none of the same cognitive decline patterns. Similarly, moderate AI companionship sometimes helped isolated elderly users and socially anxious individuals reconnect more confidently with other humans.

The tool is not destiny.

Our habits are.

Read: The Second Act Advantage

I have never been able to meditate.

I’ve tried. I understand the concept. Clear your mind, focus on your breath, let the thoughts pass without attaching to them. It sounds reasonable until you are actually sitting there and your mind is doing what minds do — cataloguing the things you forgot to say, replaying rock lyrics stuck on repeat, wondering if you should have ordered the salad instead of the pizza.

What I discovered, eventually, is that I was looking for the wrong thing. I didn’t need to empty my mind. I needed to fill it so completely with something else that there was no room left for the noise.

For me, that something is watercolor.

When I paint, I am entirely present. Not in the aspirational way people mean when they talk about mindfulness, but in the practical, non-negotiable way the medium demands. Watercolor is unforgiving in a way that no other painting form is. The moment pigment touches wet paper; the decision is permanent. You cannot paint over it. You cannot undo it. Hours into a painting, one false movement, one errant drop of water, and it is gone. The whole thing, irretrievably trashed.

This is not a frustration. It is, strangely, the point. The medium forces a quality of attention that I cannot manufacture through intention alone. I have to be here, in this brush stroke, in this moment, or the painting tells on me immediately. As the art critic James Gurney once observed: watercolor doesn’t allow you to hide anything. It shows everything — the mistakes and the successes with equal candor.

What I didn’t expect, when I started sharing my work publicly during the pandemic, was that the paintings would become something beyond personal meditation. I began posting a painting a day — not because I thought anyone particularly needed to see them, but because I was processing something I didn’t have words for, and the images seemed to do it better. I painted what I was missing. I painted isolation and community. I painted the absurdity of watching the connected, digital world I had spent my career building reveal its costs in real time.

Having spent decades championing technologies I believed would make our world better, I find myself painting the contradictions. The beauty filters changing how we see ourselves. The disembodied voices guiding us through traffic. The way Facebook and AI have come to dominate the very lives they were designed to enhance. If we dare to look up from our devices, we are faced with our unvarnished selves. By exploring that value exchange we all have succumbed to: trading privacy for convenience, intimacy for popularity, and true connections for likes, my art strives to help us all to reflect on what our lives have become and what we may have lost along the way.

I paint to say the things that don’t fit comfortably into a paragraph. My paintings only tell half the story; the other half is what happens in the mind of the person looking at them. That, to me, is the whole point. I am happiest when something I’ve made creates a small moment of levity paired with discomfort that leads to a question the viewer asks themselves rather than me.

I am not sure any of this adds up to a coherent artistic philosophy. What it adds up to is this: I paint to stay sane, and I am astonished, every time, when something made in that private state of focused solitude means something to someone else.

I share this now because it is the season when I usually take on commissioned works.

A few times each year (and with particular frequency as the fourth quarter approaches and people begin thinking about giving a gift that might actually matter) I work with clients on paintings made specifically for them. The process is genuinely collaborative: weeks or sometimes months of conversation about concept, reference, feeling, or in the relationship my artwork is meant for. Then the labor of painting itself, which takes as long as it takes.

I mention this not to sell anything, but because the most common thing people say when they finally reach out about a commissioned work is: “I wish I had contacted you sooner.” By September, I am almost always past capacity. By October, the window is closed. If this is something you have been turning over in the back of your mind — a painting for someone who is genuinely hard to give anything meaningful to, a piece to memorialize a moment or person — the time to start the conversation is now, when there is still time to work with you properly.

I will also be announcing a solo exhibition in Palm Springs in September — new work that has been building for the better part of a year, pieces I haven’t shown anywhere. I’ll share the details and the venue here when the timing is right.

Until then: I’ll keep painting. It’s the only meditation that has ever worked for me, and I am grateful, every day, that some of what comes out of it finds its way to people who feel something in it too.

And that still astonishes me.

Explore Jay's Art

Subscribe now

Something happened last week that Wall Street noticed and most people scrolled right past.

A Chinese startup called Moonshot AI unveiled an artificial intelligence model called Kimi K3. It performs on par with many of America’s best AI systems.

The difference?

Beginning July 27, anyone or any company in the world can download it—for free.

Investors immediately paid attention. The Philadelphia Semiconductor Index fell 12.5% in a single week. Nvidia, AMD, and Broadcom all dropped sharply as Wall Street began asking an uncomfortable question.

What happens if the world’s most powerful AI becomes almost free?

We’ve seen a preview before.

Earlier this year, DeepSeek shocked Silicon Valley by releasing an AI model that rivaled America’s best at a fraction of the cost. Nvidia briefly lost nearly $600 billion in market value before investors convinced themselves it was a one-time event.

This time feels different to me.

The cost of AI isn’t just falling. It’s collapsing.

Every new AI generation is helping engineers build the next one faster, cheaper, and more efficiently. What once required enormous computing power can now be accomplished with a fraction of the resources.

Think about what happened to digital photography, streaming music, or long-distance phone calls. Technologies that were once scarce and expensive eventually became abundant and nearly free.

AI may be entering that phase much sooner than anyone expected.

That’s wonderful news for consumers. It may be far less wonderful for investors.

Much of today’s stock market assumes that the biggest technology companies will continue spending hundreds of billions of dollars building AI infrastructure and then earn extraordinary profits from it for years to come.

But what if intelligence itself becomes a commodity?

If world-class AI is available to everyone at little or no cost, what exactly are customers paying premium prices for?

Having spent my entire career working on disruptive technologies, I’m not predicting when or how this plays out. Anyone who tells you they know who wins and who goes out of business is guessing.

But history has a remarkably consistent pattern.

A disruptive technology appears.

Incumbent companies dismiss it.

Markets wobble.

Confidence returns. Investors pile back in

Then, eventually, reality catches up with expectations.

We’ve watched that movie with newspapers, travel agencies, retail stores, cameras, and countless other industries.

Here’s my worry:

If a meaningful portion of your retirement savings sits in technology stocks, this isn’t a story you can afford to ignore.

The question isn’t whether AI will change the world.

It already is.

The question is who will actually make money from it.

History is full of technologies that transformed civilization.

It is a much shorter list of companies that made fortunes from them.

The math always catches up with the market.

The only question is when.

Read: The Second Act Advantage

Lately, I’ve begun to feel like an anachronism.

Not because I still enjoy fountain pens, bookstores, and handwritten notes on 120 gsm linen stationery, though I do. Not because I write books in an age of TikTok, though that certainly qualifies. Rather, because I come from a time when reading was not merely a pastime—it was considered a prerequisite for eudaimonia (a life flourishing with purpose).

The implicit bargain was simple: read broadly and the world would open itself to you. Books were repositories of knowledge that held the keys to opportunity. Opportunity led, if not to happiness, at least to a fighting chance at prosperity.

Now I look around and wonder whether society quietly renegotiated that bargain.

A recent report noted that roughly 130 million American adults read below a sixth-grade level and that AI may be helping conceal the problem by allowing workers to complete tasks they don’t fully understand. Some researchers have even coined a phrase for this phenomenon: “cognitive surrender”—the habit of accepting an AI-generated answer without doing the hard work of evaluating it yourself. Has half of our nation surrendered?

The phrase struck me because it sounds less like a technological problem and more like a cultural catastrophe. Mark Twain famously observed, “The man who does not read has no advantage over the man who cannot read.”

Twain, of course, never had to contend with a chatbot capable of summarizing Proust’s In Search of Lost Time in six seconds.

Today, anyone can go through an entire work week without reading deeply. AI can summarize the contract, draft the email, explain the report, and even tell you what the book was about without requiring you to read the book itself. Convenient? Absolutely.

But convenience is a curious thing. It often solves today’s problem by quietly creating tomorrow’s.

To be clear, I remain profoundly optimistic about AI’s ability to democratize knowledge, accelerate scientific discovery, and help people reinvent themselves.

But there is another lens through which I fear we must view the future.

What happens if we outsource not just our labor, but our learning?

Reading is not merely a method of gathering information. It is a form of cognitive calisthenics. The struggle through a difficult paragraph, the slow construction of meaning, the wrestling match with an unfamiliar idea—these are not bugs in the process. They are the process that feeds neuroplasticity and literally (fun choice of word) grows our mental capabilities.

A society that stops reading deeply risks becoming intellectually enervated while remaining superficially informed. Today’s headlines will have no context and we will fail to learn from our collective past. Or as Hannah Arendt warned us, in The Origins of Totalitarianism, when the masses come to believe “everything and nothing, think that everything was possible and that nothing was true.”

That may already be happening.

Researchers have noted a decline in voluntary reading, particularly among younger generations raised amid an endless cascade of screens, notifications, and algorithmically optimized dopamine-distributing distractions. Meanwhile, businesses report growing challenges with reading comprehension, workplace training, and critical thinking.

The irony is deliciously perverse: at the precise moment humanity gains access to the greatest library ever assembled, many of us are reading less than ever.

Fortunately, this story does not have to end badly.

Here are three simple course corrections could alter this trajectory.

First, read something difficult every day. Not a clickbait summary. Not a meme. Not a listicle. A book, an essay, or a long-form article that demands your attention and rewards your patience. The Atlantic, Harper’s, or Wired are great places to start your journey.

Second, use AI as a tutor rather than a surrogate. Ask it to challenge your thinking, explain unfamiliar concepts, or argue the opposite side of your position. Don’t let it do your thinking for you.

Third, discuss what you read. Reading becomes exponentially more powerful when ideas leave the page and enter conversation. The best books don’t end when you close them. They begin. Start a book club or join one online.

As for me, I’ll continue writing books for as long as people are willing to read them.

Perhaps that’s a quaint ambition in an era when machines can summarize entire libraries before breakfast. But I remain stubbornly convinced that there is something irreplaceable about the experience of encountering a well-crafted sentence, a surprising idea, or a fresh perspective that changes the way you see the world.

And if you’d like a little practice in that increasingly rare activity known as reading, may I humbly recommend The Second Act Advantage.

After all, thriving in the age of AI requires more than asking better prompts.

It requires becoming the kind of person capable of recognizing whether the answers make sense.

Read: The Second Act Advantage

Kelly Milligan thought she was having a heart attack.

Instead, she received the diagnosis that probably saved her life.

In October 2025, Kelly walked into Sarasota Memorial Hospital convinced she was having a cardiac event. She wasn’t. While searching for something entirely different, doctors discovered an aggressive lung cancer.

Just six months earlier, her routine scan had come back clear. The tumor had appeared and grown rapidly, without a single symptom. Had she waited for her next scheduled screening, she believes she might have been facing Stage III or Stage IV cancer—possibly too late for surgery.

Instead, she’s booked two cruises and is planning a hike across Spain and Portugal.

“This has saved my life,” she said. “Literally saved my life.”

What saved her wasn’t a physician making a lucky diagnosis. It was an AI system that continuously analyzes more than 430,000 radiology reports each year, identifying patterns no individual clinician could reasonably detect. Today, Sarasota Memorial diagnoses 75% of its lung cancers at Stage I or II—dramatically better than the national average.

That’s the difference between treating cancer early and discovering it after it’s had time to spread.

And now AI is taking another leap forward.

This week, researchers at the University of Edinburgh and NHS Lothian announced an AI-powered imaging breakthrough that can identify key lung cancer gene mutations from a single biopsy—without lengthy genetic sequencing or destroying precious tissue needed for additional testing.

Think about what that means.

Doctors could soon know not only whether a patient has cancer, but which treatment is most likely to work, in a fraction of the time.

This is exactly the kind of AI I wrote about in The Second Act Advantage.

Not replacing physicians.

Extending what they can do.

AI doesn’t diminish human expertise. AI amplifies it, allowing doctors to spend less time searching for answers and more time caring for patients.

Kelly’s physician didn’t miss her cancer because anyone was careless. The earlier scan appeared normal. AI succeeded because it could connect patterns across hundreds of thousands of cases simultaneously.

That’s not replacing medicine.

That’s what modern medicine looks like.

If you’re over 50 and have ever smoked (or even lived with someone who did) please talk to your doctor about low-dose CT screening. Only about 13% of eligible Americans are screened today. The technology already exists to save lives. The biggest barriers are awareness and access.

Kelly Milligan is planning her hike across Spain.

That’s my whole argument.

The most human application of artificial intelligence won’t be writing better emails or generating clever images.” It will be giving millions more people the chance to keep making plans they once feared they’d never live to see.

If this post encourages even one person to schedule a screening that catches their cancer early, it will be among the most useful things I’ve ever written. Please help save lives and share this post.

Read the Second Act Advantage