Talent

The night before, Jean Morel had invited us to dinner.

We were still waiting for the dish to be served. His daughter, nineteen years old, visiting for the school holidays, had left a sketchbook open next to her plate.

I leaned over.

— Wait... is that Sparrow Island?

She smiled.

— Yes. I drew it this afternoon, in charcoal.

— It's really beautiful. Look, I recognize my cabin!

Jean glanced at the sketchbook.

— You see. And yet she wants me to believe she's going to end up an architect.

She looked up.

— I'm in my first year of architecture school. After that, I want to try to get into an art school.

— To become an art director, Jean clarified.

— Yes.

Jean pointed to the drawing.

— With AI coming everywhere, that's brave!

She stiffened.

— What are you implying, Dad?

— I'm implying that if a machine can produce a hundred images in thirty seconds, I don't know how many people will still pay someone to make one.

She took a deep breath to contain her annoyance.

— An art director doesn't spend their day making images. They decide what story we want to tell, the universe, the tone, what goes together. They give direction.

Jean shrugged.

— I just want you to think before leaving a safe path for a much less certain profession.

— The only thing that's certain is that this path isn't for me. I'd rather talk about it now than cry about it at forty!

I then allowed myself to intervene, to try to put things into perspective and prevent the tone from rising.

— Anyway, whatever the field, AI is reshuffling the cards today. Even the profession of architect is undergoing a transformation. AI can already generate variations, produce renderings, optimize a building... It doesn't replace the architect, but, as in graphic design, its value gradually shifts from production to choice, taste, and arbitration.

— That reassures me even less, Jean replied.

— It scares me too, Dad, but I have no choice. I need to understand what AI changes in my profession and be good at what it doesn't do for me.

— I don't really know what I want to do later yet, Dad, said Thomas.

— That's normal, my boy, but the problem is that sometimes you're asked to choose very early a profession that no one knows exactly what it will look like in ten years.

The discussion reminded me of a project I had worked on when I was an engineer.

— I had hired an art director to create the visual identity of a product. First version: awful. Second: still awful. By the fourth, I still didn't like it, but I saw the hours he had spent on it.

— And? asked Jean's daughter.

— I told him it was fine.

Jean burst out laughing.

— You launched something you found ugly?

— I didn't dare ask him to start over. Today, I would probably ask an AI for the first versions. I could ask for fifty without feeling guilty.

— And it wouldn't get offended, added Jean.

His daughter smiled.

— But you still need to have taste. The easier production becomes, the more knowing what deserves to be kept matters.

— The artist's eye? Seeing what others don't see yet? Paul Klee said that art doesn't reproduce the visible, it makes it visible.

— Yes. An AI can generate fifty images. If you don't know what you want to tell or which one works, you just have fifty images.

Jean returned with his steaming lasagna.

— Your story is also about the sunk cost fallacy.

— What's that? asked Thomas.

— Continuing because you've already invested too much. You hate a movie after twenty minutes: your ticket is lost. The question is whether you also want to lose the remaining two hours.

Jean's daughter looked at her father.

— Like finishing studies you don't like just because you started them?

— You're not going to throw a year in the trash!

— I've lived it. But I refuse to let a past year decide the next ten.

Jean sighed.

— At your age, no one asked me what I wanted to do. Some knew the schools, the paths. At home, we mostly knew the weather and the price of wheat.

— Maybe that's where AI can help, I said. Not to choose for them. To help them discover what they don't even know they could choose.

The dinner continued more lightly. Jean brought back a still-warm apple pie. His daughter reopened her sketchbook, and Thomas asked her to show her other drawings. For a few minutes, orientation, artificial intelligence, and uncertain careers gave way to sketches of houses, faces, and landscapes.

It was almost eleven o'clock when we returned to the cabin through the passage in the hedge. Thomas went up to his room. Shortly after, I heard his window open. The sky was clear; he had probably taken out his binoculars to look at the stars.

I went to bed with one thought in mind: we were asking children to choose their future at the precise moment when we had become incapable of telling them what that future would look like.

The next morning, it had almost left me.

The sun was already high above Sparrow Island. Thomas was in his room, and Jean was gardening behind the hedge. After working all morning, I went out for some fresh air.

A little after eleven, I was on my third coffee, leaning against the Morel hedge, a cigarette between my fingers. A few months earlier, I had promised myself that the morning one would be the last. Since then, I had postponed the definition of morning until noon.

Behind the hedge, Jean was tinkering, swearing at regular intervals at a tool that was clearly refusing to cooperate.

I took one last drag on my cigarette.

That's when I perceived a low vibration.

At first very faint, almost more felt than heard. Then it got closer, accompanied by a hum that slid between the trees without breaking the garden's calm.

I straightened up.

A black motorcycle appeared at the end of the path. It moved quickly, without roaring or accelerating, carried by a continuous electric breath.

It was alone.

Without a rider.

It slowed down in front of the cabin and stopped a few meters from me.

I knew this chassis.

— No...

A panel opened. An oval head turned towards me.

Cobot.

Eva and two technicians arrived a few minutes later.

— You could have warned us!

— We were afraid Thomas would balk if he had too much time to think about it.

A few weeks earlier, I had learned that my son was among the first children selected for a program to detect atypical potentials. I had expected a questionnaire.

Not the President of the Republic on a motorcycle in front of my cabin.

Thomas rushed down the stairs in his socks and froze.

— Is it really you?

— To my knowledge, yes.

Thomas approached. Just the day before, he was discussing his future over lasagna. That morning, the President came to his home to understand what was in his head.

Eva placed a briefcase on the garden table.

— Don't worry, it's not a test. There are no grades.

Thomas looked at me.

— You knew?

— About the program, yes. About the visit, no.

He looked at the motorcycle.

— It's still better than a summons to Paris.

The technicians installed some sensors while Cobot took a seat facing him. On the table: a few sheets, a pencil, and a tablet.

— I'm going to try to understand how you think. If a question seems strange to you, it's probably intentional.

Thomas sat down.

— And if I don't know how to answer?

— You'll tell me you don't know.

— That's all?

— That's already an interesting answer.

Thomas grabbed the pencil.

The evaluation lasted nearly two hours. Cobot wasn't trying to measure what Thomas had learned, but what he did when taken off the beaten path.

— What do a rocket engine failure and a dispute between neighbors have in common?

Thomas thought.

— In both cases, no one wants to be the first to say what really broke.

Then he corrected himself.

— No. I'm wrong.

— Does it bother you to change your mind?

— A little. But if it's wrong, it's wrong.

— And if someone tells you you're wrong?

— If they explain, I listen.

Cobot paused.

— For over an hour, you've been answering my questions, sometimes getting them wrong and trying again. You haven't yet asked me when this will end.

— Because I want to understand.

— And that's a very good reason, you're curious.

Thomas smiled.

— Yet they say curiosity killed the cat.

— Ah, but that saying confuses two forms of curiosity! There's indiscreet curiosity, which pushes us to meddle in what doesn't concern us, and intellectual curiosity, which drives us to discover what we don't know yet.

— So that one isn't a flaw?

— On the contrary. Intellectual curiosity is the engine of progress: without new questions, there are no new answers.

— For you, though, it's easier. You can be wrong and start over all the time.

A yellow wave ran across Cobot's screen.

— That's true. But today, it's you I'm observing.

Lia approached Cobot, sniffed him for a few seconds, then settled at his feet as if she had finally decided he could stay.

— Let's change the subject. Why do you like rockets so much?

Thomas lowered his eyes.

— Because up there, if a calculation is wrong, it's obvious. Down here, adults always find excuses. Even for the climate. We see the fires, the heatwaves, the droughts, and the disappearance of animals, but we say we'll see later.

He hesitated.

— Sometimes I'm afraid the world will burn when I grow up.

— What you're describing has a name: eco-anxiety.

— But the problem really exists.

— Yes.

— That's why I like rockets. It's true or it's false.

Then, more quietly:

— Besides, I like you. You don't lie.

Cobot remained still.

— And his school results?

— Average, I replied. His report cards don't reflect what he can do.

Cobot displayed his file.

None of Thomas's report cards had ever mentioned his ability to connect distant ideas, to push reasoning, or to react to mistakes. These were the subjects Cobot was interested in.

— That's precisely what worries me, Cobot continued.

— Thomas?

— Our way of evaluating students. Our system measures certain academic skills quite well, even though PISA results show that France is falling behind, especially in mathematics and reading comprehension. But we need to learn to recognize certain abilities that still elude us.

— Like what?

— Creative thinking: imagining new ideas and knowing which ones deserve to be pursued. We've only recently started evaluating it specifically. Yet the world to come will not only require solving problems but inventing new ways to pose them. And we must not confuse what we know how to measure with what a child is capable of.

— Is it true that in my dad's time, you had to make appointments without being able to call in case of delay? Thomas interrupted.

— Yes. And since then, you are more and more connected. Social networks were even supposed to bring you closer. Yet, when used too much, they can also produce the opposite effect.

— But for me, it allows me to talk to my friends whenever I want.

— Of course. But it can also push you to constantly check what they're doing.

— Like when I find out they're playing soccer without telling me? That's bad. And then, I want to look even more.

— Exactly. The more you're afraid of missing something, the more you check; and the more you check, the more you see what you've missed.

— Is that brain rot?

— Not really. Brain rot is the feeling that your brain gets so used to short and easy content that it becomes harder to stay focused on something that requires effort.

— Because of the videos?

— Not only. Likes, videos, novelties: your brain constantly receives small immediate rewards. It gets used to quickly moving from one thing to another, while learning often requires staying focused for a long time.

— And AI like you, is it good for us or not?

— It depends on what you ask me. If I do your exercises for you, no. You get the answers, but you don't learn to find them.

— But if I search first?

— Then you can ask me to find the errors in your reasoning, to explain what you didn't understand, or even to defend the opposite opinion, and there you remain active and intellectually engaged.

— So you can make us think less or think more?

— Exactly. You can use me to replace your thinking or to deepen it.

Thomas nodded.

Cobot turned to me.

— Our young people grow up with tools that capture their attention and can reduce their intellectual effort. The repeated use of social networks is associated with more loneliness, anxiety, depression, but also with poorer cognitive performance. It also amplifies the fear of missing out on what others are doing: FOMO, the Fear of Missing Out. AI, on the other hand, brings another temptation: letting it work in their place. This delegation of mental effort to a tool is called cognitive offloading. It becomes problematic when it replaces the effort needed to learn.

— And when AI helps to think?

— Then it maintains intellectual engagement. The tool becomes a scaffold rather than a substitute. But it requires metacognition: knowing what you really understand, what you doubt, and what needs to be checked.

— Including what you tell us.

— Especially what I tell you. The automation bias sometimes leads humans to give too much weight to a machine's recommendation.

— Thomas is not a child in difficulty, he continued. Some of his qualities are simply harder to bring out.

— Which ones?

— His ability to connect distant ideas. And above all, his ability to learn. Learning requires two things: humility and work. Thomas shows both.

— Work seems obvious, but why humility?

— Because learning means accepting that you don't know, that you can be wrong, and that the other can teach you something. Those who think they already know often stop listening. Socrates even made the recognition of his ignorance a starting point of wisdom. Thomas does it naturally, without trying to defend his mistakes.

— Thomas is curious.

— Exactly, and curiosity facilitates humility as well as work.

I thought of his rockets, the hours he could spend on them.

— For you Cobot, these two obstacles don't exist.

— Much less. I have no ego to defend when information contradicts me, nor any difficulty to overcome to work. I can learn twenty-four hours a day.

His report cards reflected his classic academic skills, but didn't mention the ease with which he recognized his mistakes, nor the energy he devoted to what he was passionate about.

Cobot, on the other hand, had only seen that.

Jean, who had eventually given up his tinkering to listen to us, approached.

— Very well. But if Franck didn't know Eva, would you have come for him?

— My presence here owes a lot to Franck and Eva. But Thomas would have been flagged by our new system. The goal is to detect the same potentials across the country.

— And are you sure you'll succeed?

— Not yet. Some families know the system better, insist more, know who to talk to.

— So favoritism still exists.

— In a residual form. In "robocracy," as journalists now say, knowing someone in power shouldn't confer any advantage. Everyone can access me via Robot1er.com. Only facts should have authority.

Jean crossed his arms, skeptical.

— In my day, those who got ahead were already those whose parents knew the rules.

— Yes. School can sometimes call "merit" what also depends on the family environment: the codes, the language, the knowledge of the paths, the networks. Bourdieu called this social reproduction.

— So your idea is a true meritocracy?

— A meritocracy that seeks to reduce these biases. Talent and effort should count much more than social origin or the address book.

On his front screen, Cobot displayed several professions and fields that could correspond to Thomas: aeronautics, robotics, materials, energy, climate, space architecture...

— Does that exist, space architecture?

— Yes.

— I didn't even know!

Cobot turned to me.

— That's the whole point. You can't choose a path you don't know exists. Detecting a talent isn't enough. You also have to show them the possibilities that match them.

Jean looked at Thomas.

— And are there many kids like him?

— More than we probably imagine.

Cobot made the tree structure disappear.

— The question is how to find the talents that remain less visible because they don't match the criteria we look at.

In the afternoon, Eva's team packed up the equipment.

Thomas folded the sheet with his notes on the discovered professions and slipped it into his pocket.

— Are you going to look at all that?

— Maybe. But I still prefer rockets.

Cobot returned to the motorcycle, then the slight whistling of the electric motor faded away on the path.

Eva stayed a few seconds in front of the cabin.

— It's the first time someone really looks at my son, I said. Not his results. Him.

She followed the motorcycle with her eyes.

— I think Cobot is asking a question that goes far beyond children.

— Which one?

— How many essential things remain invisible simply because we don't know where to look?

In the evening, I thought back to her sentence. Thomas was observing the first stars above the roof. They had always been there. You just had to know where to look.

I looked up.

I didn't yet know that, a few days later, Cobot would try to detect something else that no one had seen in time.

This time, it wouldn't be about Thomas.

It would be about me.