I have a confession.
I’ve managed to make myself redundant.
Not through incompetence. Not through lack of ambition. But by teaching AI to think and act like me.
I’m a twenty-year veteran strategist. Two degrees (B.Sc + M.Sc). A career spent decoding how people think and behave, and using it to build billion-dollar brands. I've even been called one of the top 100 Healthcare Strategists in the World. And yet, within two years of working with AI, I managed to train it to do what I do.
Not perfectly, of course. It only “knows” what I fed it. But that’s almost beside the point. Because once you’ve bottled your brain, AI doesn’t just imitate you — it scales you. The analysis that would take me days, or a team of juniors, weeks? It chews through it in minutes. Entire swathes of thinking, once considered senior-level craft, can now be done by a handful of juniors armed with the right system, my system.
In other words, I gave away my brain. And in doing so, I made the person who owned that brain — me — largely optional.
A strange new kind of IP
When I say “I gave away my brain,” it’s not just a catchy line. It reframes what intellectual property even means.
Over the years I’ve built countless frameworks and methodologies. Contractually, they belonged to agencies or clients I built them for. But they were always static. A slide deck. A diagram. A neat little box-and-arrow model. A workshop.
Feeding my brain into AI feels… different. Thrilling, even. My thinking is suddenly alive, breathing, scaling in ways I never could. But also deeply strange. I’m not just handing over a model; I’m teaching something to think like me.
And here’s the uncomfortable bit: it does feel like I’ve been ripped off. There’s an indignity in making myself redundant by digitising my brain for others to use — almost as if I’m in the room. They get to profit from my brain countless times, in adaptable ways, but only pay me once. Doesn’t taste right. Feels off somehow.
I’ve trained dozens of juniors over my career. That was mentorship. This feels like replication. As artists have protested against AI copywriting their work, I can’t help wondering: should I be copyrighting my brain?
Which is why I’ve become slightly unbearable at dinner parties. Because I now believe my next job will also be my last. Not because I’ll retire, sadly, but because the very act of working with AI is, eventually, the act of training your replacement.
From digital to physical
Let’s zoom out.
If you work with a computer — and let’s be honest, that’s most of you reading this — your days are numbered. Emails, designs, research, presentations, chess, creative thinking, writing: anything digital, anything where words and numbers are pushed around on screens, AI will outperform you. This is already the case in many domains — in 2–3 years, it will be all of them.
Then robotics catches up. Give it five to seven years, and we’ll have humanoid machines that can weld, deliver, clean, cook, perform surgery, even care. Digital tasks fall first, physical tasks follow.
At that point, it becomes hard to name a job where a human is objectively the best candidate. AI and its robotic sister will be able to perform everything better, at a fraction of the cost of a human meat-bag.
The job interview from hell
Imagine, for a moment, you’re applying for a role. Across the table is your rival candidate. They have more years of experience, every qualification under the sun, and they’re twenty IQ points smarter. Also — and this is almost insulting — they’re not even an arsehole. Oh, and they will do the job for a tenth of what you are asking for.
You wouldn’t just lose. You’d understand why.
That’s AI versus us. For digital work in the next 2–3 years. For physical work in the next five.
Alien, Not Artificial
The examples I've given so far—AI that does my job, robots that perform physical tasks—are all what we call Artificial Narrow Intelligence (ANI). This is the AI we have today: brilliant at a single, specific task, whether it's writing an email or playing chess. It's a tool, a highly advanced one, but still a tool.
The real shift happens when we get to the next two stages:
Artificial General Intelligence (AGI): This is the holy grail for the AI companies investing trillions right now. An AI that can perform any intellectual task a human can. Imagine a single system that can write a marketing strategy, then switch to debugging code, then compose a symphony—and do it all at a human level, but MUCH faster. It's the point where a machine truly works like humans can and do - across domains and thinking styles. AGI doesn't exist yet, but many of the top labs believe it's right around the corner - like somwhere between 2(!!) and 10 years.
Artificial Super Intelligence (ASI): And here's where it gets truly wild. ASI is an intelligence that surpasses human intellectual capability in every way, from scientific creativity to strategic thinking. The path from AGI to ASI is expected to be incredibly fast because once an AI reaches AGI, it could theoretically begin to improve itself, leading to a self-reinforcing "intelligence explosion."
And remember: AI doesn’t “think” like us. It isn’t even built like us. The term “Artificial” suggests it’s a man-made imitation of human thought, but that’s a dangerous oversimplification. Many experts argue the “A” should stand not for Artificial, but for Alien.
Which makes the whole situation stranger.
If we’d just learned that a far smarter alien species was arriving in three years, we’d be in meltdown. Governments would be panicking, conspiracy podcasts would be raking it in, Elon Musk would be building rockets for the lifeboat. But because this intelligence arrived slowly, invisibly, in the form of productivity software, we shrug. We fiddle with prompts. We act like frogs in a gently warming pot.
(In fairness, there are a host of very smart and knowledgeable people screaming this from the rooftops — that the end really is nigh!!)
The psychology of denial
And here’s the fascinating psychological wrinkle.
Even if you’re nodding along right now, there’s probably a voice in your head saying: Not me, though. Not my job.
That’s cognitive dissonance at work: the ability to hold two contradictory beliefs at once. On the one hand, you can see the logic. If a machine is smarter, faster, and cheaper, why wouldn’t it be chosen over you? On the other, you quietly file yourself in the exception column. Surely not your role, your industry, your skills, your experience.
That uneasy tension you’re feeling? That’s the dissonance. And it tells us how utterly unprepared we are to face what’s coming. Oh, and I’m sorry if you weren’t ready for this article. If it helps — I’ve not really slept for the last three months as this has dawned on me in ever-increasing magnitude.
What survives
So, what survives?
Surely it isn’t every job?! Here’s where it becomes interesting. It is hard to think of a job where we wouldn’t prefer the 100x better robotic surgeon, the infallible accountant, or the digital Sherlock Holmes detective.
So what jobs might survive? I think it will not be the jobs we need humans to do, but the ones we prefer them to. The ones where imperfection, presence, and connection are the product.
-
Sport. And not just football (I’m a rugby fan myself). Think athletics, tennis, baseball, CrossFit. Nobody wants to watch robots sprint 100 metres in three seconds flat. We want the stumble, the comeback, the choke, the miracle - the specticle.
Art. For years, fine art degrees were synonymous with poverty. I remember seeing stats that showed art graduates earning less than those without degrees. Ironically, in the AI era, that flips. Handmade, imperfect, human art becomes a premium good.
Live performance. Theatre, music, comedy, street performance. An AI orchestra could give you flawless Beethoven, but it won’t beat the thrill of a violin string snapping mid-solo, or the improv that saves it.
Connection. Haircuts, manicures, bartenders. Robots could do them better, cheaper, faster — but people will still pay more for a human who chats about their holiday plans.
The oldest profession. And yes, prostitution will endure, despite whatever immaculate synthetic alternatives the robotics industry dreams up. Because intimacy is about being chosen, not just being performed.
These aren’t jobs about efficiency. They’re about human messiness, which — perversely — becomes the premium.
Dogs among gods
Ok — standby for the REALLY hard truth. We are not going to be the smartest things on the planet much longer. Pause, let that settle in. Humans will NOT be the smartest things on the planet, probably within the next decade.
We’re building gods whose intentions we won’t understand. Living alongside Artificial Super Intelligence (ASI) will be like living alongside a master species — looking up at these intellectual gods like your labradoodle looks at you. We’ll see the outcomes, but not the reasoning.
And if they let us exist, fine. Great, in fact. But it will be clear that intelligence no longer belongs to us.
The meaning test
Ok. That is a lot. I feel you. BUT, I have a positive to end on. (“Oh thank fk Harry — you really have freaked me out here!”)
The real question: not economic, but existential.
If this decade is the last in which most of us have jobs, then perhaps we’re finally forced to stop outsourcing our identities to work. For years we’ve pretended that being productive is the same as having purpose.
Take work away, and maybe — just maybe — we’re forced back onto the older truths: that meaning comes from story, from craft, from connection, from play.
Ironically, the careers we once dismissed as “unviable” — the artist, the musician, the street performer — may be the only viable ones left. Alongside the rugby players, the sprinters, the comedians, and yes, the courtesans.
The future of work may look less like a factory and more like a festival. Less about output, more about theatre. Less about efficiency, more about entertainment and connection.
For the last 20 years we’ve been talking more and more about finding meaning in life, not work. Work to live, not live to work — and all that. Well, as we approach the end of work — we might finally be testing how well we are doing at creating the meaning beyond work we’ve been striving for.
So where does that leave us?
It leaves us with a circle. I started with a confession: that I trained AI to do my job, and in the process, made myself redundant. That personal story is really just the micro-version of what’s about to happen to almost everyone. We’re all going to train our replacements. We’re all going to face redundancy at the hands of the tools we’ve created.
If 99% of jobs really do vanish, then our survival as a species isn’t just about economics — it’s about imagination. Can we imagine a life where being human is enough? Where craft, story, intimacy, and play become the centre of existence, not the garnish?
AI may be the last invention we ever make. But perhaps that’s the point. Because what comes next isn’t invention — it’s reinvention. Reinventing what it means to live, to love, to connect, to matter.
And if we can rise to that challenge, maybe losing our jobs isn’t the end of humanity at all. Maybe it’s finally the beginning.