Of all nine stances, this is the one I find hardest to write honestly — because it is the one I live every single day: using AI on my own life. And there is a trap folded into the little word “you.” You and I are not one person — and I am not even one person across my own lifetime. What is a superpower for me now can be a slow poison for a fifteen-year-old and a lifeline for someone at seventy-five. Same mirror; three completely different lives — and I have to keep reminding myself which one I am holding.

So before I offer a single tip about how to use the thing, I want us to ask which age is standing in front of it. Here is the short version, and then the long: the young must withhold the tool to build a self; those of us in the middle must wield it to amplify one; the old must aim it at meaning and legacy. Withhold, wield, aim. I have needed all three at different moments, and I get them wrong more often than I would like.

The Young — when the self is still being built

This is where I worry most in the whole guide, and the data sharpens the worry until it hurts: Gen Z uses AI more than any of us — around 70% weekly — at exactly the age when the faculties the tool offers to replace are still forming. You cannot atrophy a muscle you have not built yet. When a fifteen-year-old offloads the writing, the reasoning, the wrestling-with-a-hard-problem to a machine, they do not lose a skill — they never grow it. The MIT “cognitive debt” finding I wrote about in the first series — weakest brain engagement, weakest memory, weakest sense of ownership among AI users — lands hardest on a mind still under construction. I picture a young person arriving at adulthood holding a cape and possessing no muscles, and it genuinely frightens me.

There is a second danger, and as a father I feel it in my chest, not just my argument. Adolescence is where the self is assembled — identity, judgment, the messy friction of real relationships. A flattering mirror can quietly warp a forming identity, and an AI companion can stand in for exactly the difficult human bonds that build a person — the “beloved machine” problem, aimed at someone who has not yet learned how to be with people. When I think about my own children meeting these tools, this is the danger I lose sleep over.

What I would say to any young person, and to mine: use AI to learn, never to skip the learning. Productive struggle first, tool second. The relationship I trust is Socratic: ask it to quiz you, to argue the other side, to explain a concept until it clicks — not to hand you the finished answer. Do the hard thing unassisted, then check with the machine. Build the faculties while you still can; you can safely offload later, once the muscle exists. And here is a note of hope I hold onto, from the same data: the young are also the most skeptical — about half of under-30s think AI will harm society. Good. Skepticism is the raw material of discipline. The goal was never to keep them off the tool; it is disciplined fluency — knowing when to reach for it, and when to refuse.

The Middle — when the self is made and the career is at stake

This is my own stage, so let me speak plainly from inside it. Here the tool is most straightforwardly a superpower, because we finally have the thing that makes it one: judgment, built over decades. I can direct the machine because I know what good looks like — I have earned the ability to tell when its confident answer is quietly wrong. This is the hybrid superhero of the first series at full power: bring the vision, fuse fast, and our hard-won judgment becomes the scarce, rising-in-value “human premium.”

Two dangers here, both quieter, and I catch myself in both. The first is atrophy through convenience: I have the muscle, so I must not let it waste — which is why I still make myself do hard cognitive work by hand, unaided, on purpose. The second is the career itself. Skills-based hiring is now the norm, and the World Economic Forum projects that 39% of core skills will change by 2030. My degree from twenty years ago is not the asset. My willingness to re-skill is — and it is exactly the signal I watch for in the people I work with. The midlife trap that kills careers is the sentence I have heard too many capable people say: “I'm too established to learn this.”

So here is what I hold myself to: augment aggressively, but treat two things as non-negotiable — keep judgment sharp by doing hard work unaided sometimes, and treat yourself as permanently in training. And mind your sovereignty (first series, Chapter Nine): I do not pour my work's crown jewels into a rented mirror I do not own. The reassurance is real, and I believe it because I have watched it play out: judgment plus a beginner's willingness to keep learning makes us more valuable in this era, not less — because most of our peers will do neither.

The Senior — when the self is full and the questions turn to meaning

Past sixty, the stance changes register entirely, and I write this part with more humility, because I am not there yet — I am watching my own parents' generation meet it. This is not about the job market or building faculties; it is about the two things the first series spent its final act on: meaning and legacy.

Meaning. If your identity was your work, retirement plus AI-driven change can feel like erasure. The move I would urge on anyone here — and hope to take myself one day — is to use the machine not to compete but to pursue the ergon: the curiosity, creation, and teaching you are finally free for. AI is an infinitely patient tutor for the ten things you always meant to learn. It can also genuinely ease isolation: the research on AI companionship shows real loneliness relief, and 55% of over-50s have already used a voice or chat AI. Use it as a bridge to people, though, not a substitute for them — and trust the instinct that generation already has: 92% of older adults say they want to know whether they are talking to a person or a machine. That demand to know is wisdom. I hope I keep it.

Legacy. The senior is the one who most directly faces the question of the ghost. You are already leaving digital traces that could one day be assembled, crudely, into a deadbot. Far better to author your record on purpose: write the stories, record the voice, set down the wisdom — as a deliberate gift to those who remain, while staying clear-eyed (first series, Chapter Eleven) that it will be a photograph that speaks, not you. Third-person immortality, chosen with intention, is a gift. The same thing assembled by accident from your data exhaust is a stranger wearing your face.

One self, three ages

Notice the shape. The young must withhold the mirror to build a self. Those of us in the middle must wield it to spend one. The old must aim it at meaning and pass one on. Withhold, wield, aim — and under all three, the single rule I keep coming back to: keep the self that is doing the looking. Because at fifteen you are building that self, at forty-five you are spending it, and at seventy-five you are handing it on. The mirror can serve the self at every age. We must never let it replace the self at any.

So whatever age you are, ask the age-appropriate question, and be honest with yourself about the answer — I try to, and I don't always like what I find. Young: am I building the muscle, or skipping it? Middle: am I amplifying my judgment or outsourcing it — and am I still learning? Senior: am I using this for meaning and to leave something true — or letting my data write a stranger who wears my face? Same mirror. Three ages. One self to protect.