What’s striking is that the question itself seems slightly at odds with the essence of the piece.
Framing a “mental equivalent of the gym” assumes attention is something to maintain or optimize. But that framing already reflects the same cognitive loop many of us are caught in. You can feel it even while reading: most of the time, we’re not fully with the text, we’re scanning, anticipating, selecting what fits, sometimes as if something were at stake—a pattern that likely once served immediate physical action, but now mostly runs without discharge.
In that sense, both the gym and the “mental gym” can look like compensations for a system that rarely settles. And the moments you describe—shared presence, resonance, that quiet “mmh”—seem to arise precisely when this loop softens and attention drops back into the body
You are right but the loop persists because it's rewarded. Scanning, optimizing, producing faster aren't cognitive failures. They're rational responses to an incentive structure that prices execution and ignores orientation.
The gym metaphor appears because we're treating a structural problem as a personal discipline problem.
The moments described (shared presence, resonance, attention dropping back into the body) arise when the loop softens. But they're not compensated. They don't appear on any report. The system sees the scan, but not the presence.
When the scarce input changes, that is when human orientation, judgment, and genuine connection become what the economy prices rather than what it ignores, the cognitive loops change with it. Not through discipline. Through incentive.
The infrastructure that makes the new scarce input economically legible isn't just an economic argument. It's what creates the conditions where the "mmh" stops being a stolen moment and starts being the point.
Incentives shape behavior. But I’m not sure a nervous system conditioned for years by low-friction reward and constant scanning suddenly regains presence because the incentives changed.
At some point discipline matters too—not as optimization, but as the ability to reclaim yourself instead of constantly reacting.
And not the kind of discipline most people practice at the gym, where attention is either absorbed in some future self-image to endure the present difficulty, or just totally somewhere else
AI " intelligence" is imitation, and imitation is cheap. I was surprised how miserably it fails approaching stuff that was not in training data. Maybe true human intelligence will be the only valuable thing that let us survive crazy world filled with (left)half-brained ai installations...
Super intéressant et inspirant ! Merci pour tes articles :)
Tu as raison, je pense que l’optimisme et l’enthousiasme sont presque les seuls vrais remparts face à cette peur qu’on peut tous avoir parfois : celle d’être remplacé par l’IA du jour au lendemain, et de finir par se demander à quoi on sert encore.
Je te rejoins vraiment sur le fait que ce qui fera toujours la différence, ce sera l’humain. La confiance, le feeling, l’intuition, la manière de transmettre une idée, de rassurer, de fédérer… tout ça, c’est impossible à cloner par une IA.
On pourra peut-être imiter des formes, des mots, des raisonnements, mais pas cette présence, pas cette "energie" humaine. Et je suis convaincu que ce sera justement l’une des plus grandes forces des leaders de demain.
De mon côté, il y a des choses que je ne voudrais jamais voir remplacées par l’IA.
Tout ce qui touche à la formation, à l’éducation, à la transmission du savoir. Oui, aujourd’hui on a accès à l’info en deux secondes. Mais l’info ne remplacera jamais quelqu’un qui t’explique, qui te transmet, qui partage son expérience, ses erreurs, son vécu. C’est là qu’on apprend vraiment, selon moi.
Et pareil pour le médical. Que l’IA aide à mieux comprendre un symptôme pourquoi pas. Mais elle ne remplacera jamais un médecin. Et surtout, il ne faudrait pas qu’elle pousse à l’automédication.
Au plaisir d'échanger sur ces sujets de vive voix !
- the craving we have for shared experience and it’s growing value
- the singularity of artistry - as you say I don’t think people will fall for AI dua lipa
- the way that Ai will increasingly feel magical - echoing the Arthur Clarke point on “any sufficiently advanced technology is indistinguishable from magic” - my early post about Ai were all featuring a magician
- And that "stop hiring humans" ad, we had it in Paris around VivaTech, it stopped me cold too. Wrong on every level.
For sure the destination you describe is bright and worth working for …. But one thing I feel is unresolved though is the road to get there which may be creating more asymmetries … between those with capex and those without, those with robot hunger and those with fear, those who can afford agents and those who can’t … How to navigate this messy road is another question that can keep me awake …
When execution becomes free, the source of execution becomes the scarce signal. Who produced it — and whether you trust them — replaces what was produced as the differentiating question. That's the foundation everything else in this piece builds on correctly.
The question it opens for me: does the infrastructure exist to make that trust economically legible at scale? If human orientation, taste, and judgment become the scarce currency — but the systems distributing economic reward are still calibrated to measure execution — the value is real and the compensation doesn't follow.
The carpenter whose physical craft becomes more valuable in an AI-abundant world still needs the market to actually see that scarcity and price it.
I've been attempting to build the infrastructure argument — attribution systems, productive entity registration, the architecture that makes human creative direction structurally owned rather than just culturally valued. Your piece identifies the destination precisely. That work is trying to build the on-ramp.
Your closing question is exactly right. I'd add one: what infrastructure needs to exist to make that refusal economically viable rather than just principled?
Imagination indeed. This question is posed with the notion that humans will not evolve. Imagine if you will, how humans could evolve with AI ? The species as we understand as human will no longer be the same...
As a rising senior in college, I have this written above my desk.
"Strong / free personality and actively unselfish."
The prior - Expressing myself more artfully has been a ceiling-less pursuit. For me, there is no destination because it's a continual unfolding. This is where I feel a childlike sense of play / flow state.
The latter - Seeing that spark in someone else's eyes when I have a positive impact on them is the most rewarding external event.
(This is almost directly "As a Man Thinketh" -James Allen)
To put time into something, I need it to check both of these boxes.
Answering your question concretely, I find this changes daily and is unpredictable. To show this, I'll take my answer down the ladder of abstraction to events in my life the past couple of days : building an AI chrome extension aiming to make a dent in American illiteracy, partying, finding new clothes, making a meal for my roommates, a friend's album release concert at an apartment, a potluck, planning summer trips(half the fun it spitballing ideas), cards and coffee in the sun.
What would you refuse to let AI do for you, even if it could do it better?
I refuse to let AI own my work completely. Even as we move away from the "Copilot" narrative into the "Agentic" one, I still find that using it as a copilot enhances my thought process, ownership of work, and intelligence (or at least my own perception of my intelligence). This means never relying on it to output something I do not understand, and still owning the problem framing and most of the narrative behind decisions (although AI does bake that in to responses sometimes now).
Now as the ability of it increases, being stuck to using it this way may make me slower at output than my peers, so it is something I am going to have to navigate especially since the world rewards output more than anything.
I'm probably becoming a little too reliant and could benefit from a detox where I cut them out completely.
I agree with Houda: when intelligence gets cheap, the valuable stuff gets weirdly human.
But I’d add one unromantic footnote: taste, trust, judgment, presence, legitimacy, none of these scale by vibes alone. They have to be organized. A brilliant person with ten AI agents is not a company. It is a small weather system with Slack.
The scarce thing will not be intelligence. It will be agreement about what intelligence is for.
Bad teams will use AI to produce more plausible noise, faster. Good teams will use it to make the hidden work visible: the tradeoffs, the exceptions, the judgment calls, the apprenticeship nobody wanted to budget for.
Cheap intelligence does not make humans ornamental. It makes poorly designed humans-in-groups impossible to hide.
First of all, banger fucking essay. (I said it that way so you’ll know, or at least have some indication, that I am a real person and not an AI agent.)
I watch a lot of American football. A favorite player of mine (from some time back; maybe 10 years ago); he was raised by two parents who picked tomatoes for a living for minimum wage. This guy was a true “baller”; a person whose effort level was off the chart; every game and every play. At the time I loved to watch him play. I loved the drive.
But over time I came to realize: his motivation was born of his parents’ struggle. He was working for them; balling out to earn a contract, to buy his parents a decent house and get them out of the fields. That deflated the joy for me as I understood it more. I was watching his intensity, as if it was born of joy and love of the game. But in reality it came from a darker place.
You made me think of this guy and his parents, in this article. I wonder how many farm workers will participate in the coming human-effort-revaluation? I hope everyone one of them. But, let’s admit we have a long way to go. I’m still buying my tomatoes at the supermarket. This dudes parents, or the faceless grunt workers who replaced them in the fields, are still invisible to me as I shop. I’m not yet selecting for human connection. I want a good tomato at a low price. Which is both natural, and perverse.
And I wonder if we can still excel without struggle to motivate us? If life is easy (and I think it’ll be much easier over time), what will we struggle against? My baller would have been lost without his struggle. Many of us are the same: we need conflict to bring out our fight. For in our fight resides our genius, if we have any genius to give.
Perhaps the future belongs to those who are still driven, even in a world of abundance.
This essay arrived for me at a very specific moment. I've spent the last few days writing in my notebook about the same question from the other side — not "what will AI replace," but "what did I already give up without noticing."
I kept two lists. What I refused to outsource, and what I outsourced without realizing.
I never asked AI to write a blog post from scratch. I never asked it to outline my tangible, handwritten notebook — because the pen is where I feel most present. I never asked it to summarize my day for me, because the act of remembering yesterday to connect with today is the reflection itself. I never asked it to sit with me in my grief, or to write my prayers, or to translate scripture into my own life. I never missed it the way I miss people. I never felt jealous of it. I never asked it for love.
But I did let it quietly take over the steering wheel of my own learning for almost two years. I didn't notice until I noticed. I had outsourced the part where I sit with not-knowing — and that turned out to be the most important part.
So your question lands with weight. What I would never hand over is the act of noticing. Choosing, in real time, what is worth looking at. AI can summarize anything after the fact. But the "this, right now, matters" — that has to stay mine, or I lose the thread of my own life.
Thank you for writing this. I'm reorganizing how I live because of a few sentences here.
What’s striking is that the question itself seems slightly at odds with the essence of the piece.
Framing a “mental equivalent of the gym” assumes attention is something to maintain or optimize. But that framing already reflects the same cognitive loop many of us are caught in. You can feel it even while reading: most of the time, we’re not fully with the text, we’re scanning, anticipating, selecting what fits, sometimes as if something were at stake—a pattern that likely once served immediate physical action, but now mostly runs without discharge.
In that sense, both the gym and the “mental gym” can look like compensations for a system that rarely settles. And the moments you describe—shared presence, resonance, that quiet “mmh”—seem to arise precisely when this loop softens and attention drops back into the body
You are right but the loop persists because it's rewarded. Scanning, optimizing, producing faster aren't cognitive failures. They're rational responses to an incentive structure that prices execution and ignores orientation.
The gym metaphor appears because we're treating a structural problem as a personal discipline problem.
The moments described (shared presence, resonance, attention dropping back into the body) arise when the loop softens. But they're not compensated. They don't appear on any report. The system sees the scan, but not the presence.
When the scarce input changes, that is when human orientation, judgment, and genuine connection become what the economy prices rather than what it ignores, the cognitive loops change with it. Not through discipline. Through incentive.
The infrastructure that makes the new scarce input economically legible isn't just an economic argument. It's what creates the conditions where the "mmh" stops being a stolen moment and starts being the point.
Incentives shape behavior. But I’m not sure a nervous system conditioned for years by low-friction reward and constant scanning suddenly regains presence because the incentives changed.
At some point discipline matters too—not as optimization, but as the ability to reclaim yourself instead of constantly reacting.
And not the kind of discipline most people practice at the gym, where attention is either absorbed in some future self-image to endure the present difficulty, or just totally somewhere else
AI " intelligence" is imitation, and imitation is cheap. I was surprised how miserably it fails approaching stuff that was not in training data. Maybe true human intelligence will be the only valuable thing that let us survive crazy world filled with (left)half-brained ai installations...
A substantial piece. At present we’re at the emergent epi-phenomena.
Imagine when we enter the phenomena. Yale professor Santos sets aside complexity and defaults to simple acts of loving others - not an obsession on the plight of SELF: https://www.thedartmouth.com/article/2026/05/yale-professor-laurie-santos-says-happiness-is-about-doing-stuff-for-others?shem=rimspwouoe,
Super intéressant et inspirant ! Merci pour tes articles :)
Tu as raison, je pense que l’optimisme et l’enthousiasme sont presque les seuls vrais remparts face à cette peur qu’on peut tous avoir parfois : celle d’être remplacé par l’IA du jour au lendemain, et de finir par se demander à quoi on sert encore.
Je te rejoins vraiment sur le fait que ce qui fera toujours la différence, ce sera l’humain. La confiance, le feeling, l’intuition, la manière de transmettre une idée, de rassurer, de fédérer… tout ça, c’est impossible à cloner par une IA.
On pourra peut-être imiter des formes, des mots, des raisonnements, mais pas cette présence, pas cette "energie" humaine. Et je suis convaincu que ce sera justement l’une des plus grandes forces des leaders de demain.
De mon côté, il y a des choses que je ne voudrais jamais voir remplacées par l’IA.
Tout ce qui touche à la formation, à l’éducation, à la transmission du savoir. Oui, aujourd’hui on a accès à l’info en deux secondes. Mais l’info ne remplacera jamais quelqu’un qui t’explique, qui te transmet, qui partage son expérience, ses erreurs, son vécu. C’est là qu’on apprend vraiment, selon moi.
Et pareil pour le médical. Que l’IA aide à mieux comprendre un symptôme pourquoi pas. Mais elle ne remplacera jamais un médecin. Et surtout, il ne faudrait pas qu’elle pousse à l’automédication.
Au plaisir d'échanger sur ces sujets de vive voix !
When intelligence becomes cheap, the scarce layer moves upward.
Not just to “humanity” in the abstract, but to judgment.
What becomes valuable is:
- taste
- trust
- courage
- coordination
- responsibility
- original ontology
- the ability to choose what matters
AI can make answers abundant.
It cannot automatically make meaning abundant.
That may be the real shift:
from producing more output
to deciding what is worth producing at all.
So much here I agree with
- the craving we have for shared experience and it’s growing value
- the singularity of artistry - as you say I don’t think people will fall for AI dua lipa
- the way that Ai will increasingly feel magical - echoing the Arthur Clarke point on “any sufficiently advanced technology is indistinguishable from magic” - my early post about Ai were all featuring a magician
- And that "stop hiring humans" ad, we had it in Paris around VivaTech, it stopped me cold too. Wrong on every level.
For sure the destination you describe is bright and worth working for …. But one thing I feel is unresolved though is the road to get there which may be creating more asymmetries … between those with capex and those without, those with robot hunger and those with fear, those who can afford agents and those who can’t … How to navigate this messy road is another question that can keep me awake …
When execution becomes free, the source of execution becomes the scarce signal. Who produced it — and whether you trust them — replaces what was produced as the differentiating question. That's the foundation everything else in this piece builds on correctly.
The question it opens for me: does the infrastructure exist to make that trust economically legible at scale? If human orientation, taste, and judgment become the scarce currency — but the systems distributing economic reward are still calibrated to measure execution — the value is real and the compensation doesn't follow.
The carpenter whose physical craft becomes more valuable in an AI-abundant world still needs the market to actually see that scarcity and price it.
I've been attempting to build the infrastructure argument — attribution systems, productive entity registration, the architecture that makes human creative direction structurally owned rather than just culturally valued. Your piece identifies the destination precisely. That work is trying to build the on-ramp.
Your closing question is exactly right. I'd add one: what infrastructure needs to exist to make that refusal economically viable rather than just principled?
“As we delegate,*trust* becomes the binding currency of the new economy.” What a beautiful and truthful perspective.
Imagination indeed. This question is posed with the notion that humans will not evolve. Imagine if you will, how humans could evolve with AI ? The species as we understand as human will no longer be the same...
As a rising senior in college, I have this written above my desk.
"Strong / free personality and actively unselfish."
The prior - Expressing myself more artfully has been a ceiling-less pursuit. For me, there is no destination because it's a continual unfolding. This is where I feel a childlike sense of play / flow state.
The latter - Seeing that spark in someone else's eyes when I have a positive impact on them is the most rewarding external event.
(This is almost directly "As a Man Thinketh" -James Allen)
To put time into something, I need it to check both of these boxes.
Answering your question concretely, I find this changes daily and is unpredictable. To show this, I'll take my answer down the ladder of abstraction to events in my life the past couple of days : building an AI chrome extension aiming to make a dent in American illiteracy, partying, finding new clothes, making a meal for my roommates, a friend's album release concert at an apartment, a potluck, planning summer trips(half the fun it spitballing ideas), cards and coffee in the sun.
What would you refuse to let AI do for you, even if it could do it better?
I refuse to let AI own my work completely. Even as we move away from the "Copilot" narrative into the "Agentic" one, I still find that using it as a copilot enhances my thought process, ownership of work, and intelligence (or at least my own perception of my intelligence). This means never relying on it to output something I do not understand, and still owning the problem framing and most of the narrative behind decisions (although AI does bake that in to responses sometimes now).
Now as the ability of it increases, being stuck to using it this way may make me slower at output than my peers, so it is something I am going to have to navigate especially since the world rewards output more than anything.
I'm probably becoming a little too reliant and could benefit from a detox where I cut them out completely.
I agree with Houda: when intelligence gets cheap, the valuable stuff gets weirdly human.
But I’d add one unromantic footnote: taste, trust, judgment, presence, legitimacy, none of these scale by vibes alone. They have to be organized. A brilliant person with ten AI agents is not a company. It is a small weather system with Slack.
The scarce thing will not be intelligence. It will be agreement about what intelligence is for.
Bad teams will use AI to produce more plausible noise, faster. Good teams will use it to make the hidden work visible: the tradeoffs, the exceptions, the judgment calls, the apprenticeship nobody wanted to budget for.
Cheap intelligence does not make humans ornamental. It makes poorly designed humans-in-groups impossible to hide.
First of all, banger fucking essay. (I said it that way so you’ll know, or at least have some indication, that I am a real person and not an AI agent.)
I watch a lot of American football. A favorite player of mine (from some time back; maybe 10 years ago); he was raised by two parents who picked tomatoes for a living for minimum wage. This guy was a true “baller”; a person whose effort level was off the chart; every game and every play. At the time I loved to watch him play. I loved the drive.
But over time I came to realize: his motivation was born of his parents’ struggle. He was working for them; balling out to earn a contract, to buy his parents a decent house and get them out of the fields. That deflated the joy for me as I understood it more. I was watching his intensity, as if it was born of joy and love of the game. But in reality it came from a darker place.
You made me think of this guy and his parents, in this article. I wonder how many farm workers will participate in the coming human-effort-revaluation? I hope everyone one of them. But, let’s admit we have a long way to go. I’m still buying my tomatoes at the supermarket. This dudes parents, or the faceless grunt workers who replaced them in the fields, are still invisible to me as I shop. I’m not yet selecting for human connection. I want a good tomato at a low price. Which is both natural, and perverse.
And I wonder if we can still excel without struggle to motivate us? If life is easy (and I think it’ll be much easier over time), what will we struggle against? My baller would have been lost without his struggle. Many of us are the same: we need conflict to bring out our fight. For in our fight resides our genius, if we have any genius to give.
Perhaps the future belongs to those who are still driven, even in a world of abundance.
Intelligence isn’t arbitrary.
This essay arrived for me at a very specific moment. I've spent the last few days writing in my notebook about the same question from the other side — not "what will AI replace," but "what did I already give up without noticing."
I kept two lists. What I refused to outsource, and what I outsourced without realizing.
I never asked AI to write a blog post from scratch. I never asked it to outline my tangible, handwritten notebook — because the pen is where I feel most present. I never asked it to summarize my day for me, because the act of remembering yesterday to connect with today is the reflection itself. I never asked it to sit with me in my grief, or to write my prayers, or to translate scripture into my own life. I never missed it the way I miss people. I never felt jealous of it. I never asked it for love.
But I did let it quietly take over the steering wheel of my own learning for almost two years. I didn't notice until I noticed. I had outsourced the part where I sit with not-knowing — and that turned out to be the most important part.
So your question lands with weight. What I would never hand over is the act of noticing. Choosing, in real time, what is worth looking at. AI can summarize anything after the fact. But the "this, right now, matters" — that has to stay mine, or I lose the thread of my own life.
Thank you for writing this. I'm reorganizing how I live because of a few sentences here.