[{"content":"Yes, VivaTech was ten days ago.\nNo, I didn\u0026rsquo;t write in the heat of the moment. Everyone else did. The LinkedIn posts went up on Thursday evening — rocket emojis, selfies with robots, \u0026ldquo;game changer,\u0026rdquo; \u0026ldquo;mind blown,\u0026rdquo; and the traditional blurry photo of a screen displaying something impressive.\nI chose not to do that.\nNot out of snobbery. Out of intellectual honesty.\nBecause a show like VivaTech is designed to dazzle you. That\u0026rsquo;s its function. It\u0026rsquo;s even its virtue — it makes visible what is still invisible in most organizations. But between what you see and what it means, there is work to be done. Work of distillation, of putting things in tension, of connecting what you observe with what you see elsewhere, in the organizations you support.\nThat work takes time.\nTen days, to be precise.\nHere is what I took away from it — and what the Thursday-evening posts didn\u0026rsquo;t say.\nFirst shock: medicine A scan of my retina. Five minutes.\nAnomalies identified. No surprise to me — I already knew about them. But it had taken me five years, the right address, the right network, the right persistence to obtain that diagnosis.\nFive years versus five minutes.\nWe\u0026rsquo;ve heard a great deal about AI in healthcare. Far too much, often for nothing. Medical copilots that ignore the patient\u0026rsquo;s emotional context. Health-data platforms blocked by regulations no one has genuinely translated into operational constraints. Projects put on hold \u0026ldquo;while we secure things\u0026rdquo; — which, in healthcare, sometimes means years.\nBut there, in front of that retinal scanner, I wasn\u0026rsquo;t thinking about the technology.\nI was thinking about the people who don\u0026rsquo;t have my address book. Those who don\u0026rsquo;t have five years to wait. Those for whom \u0026ldquo;finding the right specialist\u0026rdquo; is a matter of geographic and social luck as much as medical need.\nThis isn\u0026rsquo;t a question of technology. It\u0026rsquo;s a question of access.\nAnd access is not a data scientist\u0026rsquo;s problem. It\u0026rsquo;s a matter of governance, of distribution, of political choices about what we decide to fund, to deploy, and for whom, as a priority. Choices made — or not made — in rooms where tech evangelists are generally not invited.\nWho decides how this retinal imaging gets deployed? At what pace? Across which territories? Funded how, and reimbursed by whom?\nThose questions weren\u0026rsquo;t on the stands.\nSecond shock: the perfume I told stories from my life. Emotions. Scents that had marked me. And they made a perfume.\nIt smells wonderful. It resembles me 100%.\nI know exactly what happened: a model transformed my accounts into semantic representations, projected them into a space of olfactory parameters, and produced a formulation. Nothing magical. Very sophisticated calculation.\nBut the experience itself was something else.\nIt reminded me of something I\u0026rsquo;ve been writing for a long time: AI does not understand — it correlates. And sometimes the correlation reaches an accuracy so close to comprehension that you can no longer see the difference.\nThe difference isn\u0026rsquo;t visible in the demo. It shows up in the edge cases — the patient whose symptoms fit no known pattern, the customer whose story contains a cultural ambivalence the model never learned to decode, the organization whose real problem is the inverse of what it states.\nWhat the perfume confirmed for me was not the power of AI.\nIt was the power of the staging of AI.\nVivaTech is the world\u0026rsquo;s largest theatre for that. And this is not a criticism — it\u0026rsquo;s a diagnosis. A theatre can tell profound truths. As long as you know you are in a theatre.\nThe official tagline of this edition was, however, \u0026ldquo;Artificial Intelligence: impact, not illusion.\u0026rdquo; The intention was commendable. But you don\u0026rsquo;t escape illusion by declaring it on a banner. You escape it by asking the questions the staging conceals.\nWhat else I saw, which moved through me differently An augmented-reality headset. Virtual flowers at my feet. I bent down to gather them.\nI smelled the flowers.\nThat moment had nothing of a technological demonstration about it. It had something of a promise — the kind of promise that forces you to think seriously about what you do with it.\nI was thinking of people who have lost their sight. Of camera glasses that describe what they see, that read signs, that anticipate obstacles, that let you \u0026ldquo;cross\u0026rdquo; a landscape through words when the eyes can no longer do it. I was thinking of people who will walk again thanks to exoskeletons, of diagnoses caught earlier, of diseases intercepted before they become irreversible.\nThe promise is real. Deeply human, even.\nBut the promise and the trajectory of access are two distinct things. One is shown in five minutes at a trade show. The other is built — or not — in budget trade-offs, reimbursement negotiations, prioritization choices that no one puts on display under the VivaTech lights.\nThis is not technophobia. It\u0026rsquo;s the most concrete question there is: who benefits from human augmentation, and by what rules?\nWhat I heard all day — and what troubled me Sovereignty. Security. Data. Compliance. Inclusion.\nIn every corridor, on every stage.\nAnd that\u0026rsquo;s where something troubled me — not because these words were present, but because they were floating. Declarative. Rhetorical. They served as moral cover for demos that, ten metres away, cheerfully embedded biometric processing, behavioural inference, models trained who-knows-where on who-knows-what.\nSovereignty is not a buzzword you slip into a conference slide. It\u0026rsquo;s an architecture. A concrete decision about where the data lives, who accesses it, under which jurisdiction, with what mechanisms of revocation and audit.\nDeclaring sovereignty without architecting it is exactly like declaring digital transformation without touching the production systems.\nWe\u0026rsquo;ve all lived through that. For fifteen years. We know how it ends.\nThe observation that worried me most — and was the least commented on Countries we still called \u0026ldquo;developing\u0026rdquo; ten years ago are now outpacing us on certain AI uses. Not on all. But on entire swaths — deployment at scale, cost of transformation, speed of adoption.\nHow did they manage without the resources we believe indispensable?\nThey worked on their data. Streamlined their processes. Defined clear priorities and stuck to them. Built in an almost military fashion — without governance committees that drag on for six months, without legacy that paralyzes, without twenty years of IT sedimentation to carry.\nThis isn\u0026rsquo;t \u0026ldquo;constraint breeds creativity.\u0026rdquo; It\u0026rsquo;s a form of structural agility that abundance — and accumulated complexity — has denied us.\nThey had no technical debt. They also had no temptation to keep everything, interconnect everything, maintain everything \u0026ldquo;just in case.\u0026rdquo; They built from zero, which is a curse in the classical economy, and a brutal advantage in the data economy.\nFrance — and Europe — do not have an intelligence problem. They have an execution problem. That\u0026rsquo;s not the same problem. And it\u0026rsquo;s infinitely harder to solve. Because you can fund intelligence. You cannot buy the ability to simplify what you spent thirty years complicating, nor the political will to do so.\nThe real question is not: are we behind? It is: what do we choose to fight for — and with what method?\nWhat VivaTech 2026 really taught me Not that the future is here. I knew that.\nNot that AI is powerful. That too.\nWhat I saw, and what it took me ten days to formulate properly, is that the real divide is not between optimists and pessimists on AI. It\u0026rsquo;s not between early adopters and skeptics. It\u0026rsquo;s between those who ask the right questions — about governance, about access, about the values encoded in systems, about who holds the wheel — and those who settle for watching the demo.\nAt a show like VivaTech, everyone shows what AI can do.\nThe demonstration is spectacular. Sometimes overwhelming. Often sincere.\nBut no one shows who decided it should work this way. Who defined the training criteria. Which values were encoded, by whom, under what mandate. Which organization gave up what so that this demo could exist.\nThose questions are not technical. Nor are they philosophical in the abstract sense of the word.\nThey are architectural — in the deepest sense of the word. Architecture is what makes visible the choices we made, and the ones we avoided. It\u0026rsquo;s what turns an intention into an operational constraint, a declared value into a verifiable rule.\nAs long as we treat VivaTech as a fair of wonder rather than a mirror of our collective trade-offs, we will keep coming back with blurry photos of impressive screens — and unasked questions that will do damage later.\nI\u0026rsquo;d rather ask them now.\nYour reactions, disagreements, and additional information are welcome. Write to me at laurence.poussard63@gmail.com or leave a comment.\n","permalink":"https://laurence-poussard.com/en/articles/2026-06-30-vivatech-ce-que-ca-revele/","summary":"\u003cp\u003eYes, VivaTech was ten days ago.\u003c/p\u003e\n\u003cp\u003eNo, I didn\u0026rsquo;t write in the heat of the moment. Everyone else did. The LinkedIn posts went up on Thursday evening — rocket emojis, selfies with robots, \u0026ldquo;game changer,\u0026rdquo; \u0026ldquo;mind blown,\u0026rdquo; and the traditional blurry photo of a screen displaying something impressive.\u003c/p\u003e\n\u003cp\u003eI chose not to do that.\u003c/p\u003e\n\u003cp\u003eNot out of snobbery. Out of intellectual honesty.\u003c/p\u003e\n\u003cp\u003eBecause a show like VivaTech is designed to dazzle you. That\u0026rsquo;s its function. It\u0026rsquo;s even its virtue — it makes visible what is still invisible in most organizations. But between what you see and what it means, there is work to be done. Work of distillation, of putting things in tension, of connecting what you observe with what you see elsewhere, in the organizations you support.\u003c/p\u003e","title":"VivaTech 2026: what I saw struck me less than what it revealed"},{"content":"In response to Nirit Weiss-Blatt\u0026rsquo;s article, \u0026ldquo;First They Built a Secular Apocalypse Belief System. Now They Want Religious Authority.\u0026rdquo;\nSome articles hit home — not because they say everything, but because they name something everyone sensed without being able to articulate it. Nirit Weiss-Blatt\u0026rsquo;s piece, published on June 2, 2026 on AI Panic, is one of them.\nHer argument is simple and sharp: AI doomerism — the current that predicts human extinction by AI — has the structural attributes of a religious movement. God replaced by superintelligence, hell replaced by extinction, the prophets replaced by rationalist probabilists convinced they\u0026rsquo;ve grasped what the masses ignore. And now, the cherry on the dogma: this movement, born in atheist circles, goes looking for its moral legitimacy… at the Vatican.\nThe paradox is delicious. But it conceals something far more interesting.\nThe real problem isn\u0026rsquo;t the discourse. It\u0026rsquo;s the capture. Weiss-Blatt is right to expose the rhetorical structure of doomerism. But by focusing on what is said, she sidesteps the question of who says it, why, and above all who benefits.\nThe very actors who brandish AI\u0026rsquo;s existential risk are precisely the ones building that AI at forced march. Anthropic — valued at 965 billion dollars after its Series H of 65 billion closed on May 28, 2026 — builds language models while being publicly described as \u0026ldquo;midwifing a deity,\u0026rdquo; to borrow Bill Gurley\u0026rsquo;s words on the All-In podcast. OpenAI, valued at 852 billion dollars in April 2026, created an exceptional capital structure by invoking the urgency of \u0026ldquo;securing\u0026rdquo; AI before bad actors do. The effective altruism movement raised hundreds of millions of dollars by stoking the fear of catastrophe.\nIt\u0026rsquo;s a mechanism I know well from my enterprise architecture engagements: when an actor is both firefighter and arsonist, you have to watch which way the wind is blowing. Here, the wind blows toward the concentration of decision-making power over AI in the hands of a few private actors, bypassing any legitimate democratic body.\nThis isn\u0026rsquo;t a conspiracy theory. It\u0026rsquo;s a structural reading. And it changes everything.\nChaos as an argument from authority In my book Nous venons du chaos, l\u0026rsquo;IA de la logique (\u0026ldquo;We Come From Chaos, AI From Logic\u0026rdquo;), I wanted to set the record straight on one precise point: AI is neither the devil nor the messiah. It is a lever of massive influence — economic, political, cognitive — in the hands of those who build it, deploy it, and set its rules.\nApocalyptic discourse and messianic discourse share the same function: they disqualify ordinary intelligence. They say, in substance: this subject is too complex, too dangerous, too fundamental to be left to the uninitiated. And therefore, implicitly: let us decide.\nThis is the argument from authority through chaos. And it is extraordinarily effective, because it captures the anxious and the enthusiastic, believers and skeptics, decision-makers and ordinary citizens alike — each with their version of the same story arc: AI surpasses the human, the human cannot understand, guardians are needed.\nI spent twenty years decoding complex information systems for organizations that were afraid of their own application estate. The mechanism is identical: chaos, real or manufactured, creates a dependence on the experts who claim to master it.\nWhat Weiss-Blatt doesn\u0026rsquo;t say — and what I feel compelled to There is a detail in her article I wouldn\u0026rsquo;t have left in a footnote.\nChristopher Olah, Canadian co-founder of Anthropic and head of its interpretability team, represented the company at the Vatican on May 25, 2026, at the presentation of Pope Leo XIV\u0026rsquo;s encyclical on artificial intelligence, Magnifica humanitas: On safeguarding the human person in the time of artificial Intelligence. A declared atheist, Olah — according to The Atlantic — compared himself to a priest, in the sense of someone who helps the model Claude \u0026ldquo;be a good person, in some sense.\u0026rdquo;\nLet\u0026rsquo;s stop there for a moment.\nWho asked him to define what \u0026ldquo;a good person\u0026rdquo; is? Who mandated him to encode values into a system that will be used by millions of individuals, across dozens of radically different cultural, regulatory, and political contexts? Who validates that his values — those of a 33-year-old Californian engineer from Silicon Valley — are universally desirable?\nThe real governance question is not: will AI destroy humanity?\nIt is: who decides the values AI embodies, and through what democratic process?\nThis isn\u0026rsquo;t a religious question. It\u0026rsquo;s a question of sovereignty. And in Europe, we tend to ask it too late, too timidly, and in the wrong forums.\nWhat Olah himself says — and why it authorizes me to speak In his Vatican address, Olah made a statement worth reading closely. I quote it in full:\n\u0026ldquo;Every frontier AI lab—including Anthropic—operates inside a set of incentives and constraints that can sometimes conflict with doing the right thing. (\u0026hellip;) It is enormously important that there be people outside those incentives—people who care about things going well and insist on safety, who are paying close attention, who are willing to say hard things, who are willing to be our earnest, thoughtful, critics.\u0026rdquo;\nOlah calls for external critics. My external critique is precisely this: the unprecedented concentration of decision-making power over AI systems in the hands of a few private American labs — whose combined valuation exceeds 1,800 billion dollars — while the democratic institutions of regulation, European ones first and foremost, have neither the capital, nor the compute, nor the models to weigh into this conversation on equal terms.\nThat the founders of these labs are personally convinced they\u0026rsquo;re doing good changes nothing about the structural question. An actor\u0026rsquo;s sincerity is no substitute for the legitimacy of a democratic process.\nThe real fault line: not fear vs. no fear. But power vs. governance. What interests me in the debate Weiss-Blatt opens — even if she doesn\u0026rsquo;t frame it this way — is that the fault line on AI is not between optimists and pessimists, between techno-enthusiasts and doomers.\nIt is between those who believe AI governance should remain in the hands of \u0026ldquo;responsible\u0026rdquo; private actors (self-certified, as Gregor Hohpe would say of an architecture that validates itself) and those who believe that systems at this scale of impact call for democratic regulation, real checks and balances, and independent audit capacity.\nLet\u0026rsquo;s look at the actual numbers.\nIn February 2025, Emmanuel Macron announced at the Paris summit a consolidated envelope of 109 billion dollars for AI, aggregating Emirati and Canadian funds and French rounds including Mistral AI. Ursula von der Leyen announced in its wake 20 billion euros for four to five AI gigafactories on European Union soil. Fifteen months later, in May 2026, Anthropic alone raises 65 billion dollars in a single funding round. The investor map published by Anthropic shows the absence of any European public fund: no Bpifrance, no European Investment Fund, no equivalent. The continent capable of aggregating 109 billion around a Paris conference does not place a single public dollar in the round that crowns, at that very moment, the world\u0026rsquo;s leading private artificial intelligence player.\nThis isn\u0026rsquo;t an abstract figure. It\u0026rsquo;s a dependency map — exactly the kind we produce in enterprise architecture to reveal an information system\u0026rsquo;s points of fragility.\nAnd an organization — or a continent — whose architecture is entirely dependent on external actors who unilaterally define the rules of the game is not a sovereign organization. It is a captive organization telling itself it made a choice.\nWhat I take away — and what I do with it I am neither a doomer nor a boomer. I am an architect. My job is to look systems in the face: their dependencies, their blind spots, their tipping points, and the interests they serve.\nAI is no different. It is not a matter of science fiction, nor of secular theology. It is a matter of governance, of sovereignty, of collective capacity to decide.\nThe catastrophist discourse Weiss-Blatt criticizes has a very concrete effect: it installs the idea that decision belongs to the initiated. And every time that idea gains ground, it erodes a little more our collective capacity to say: no, we want to understand, question, steer, refuse.\nThat\u0026rsquo;s why I write. Not to reassure. Not to frighten. To make thinkable what claims to be unthinkable — and to give back the power to decide to those to whom it belongs.\nSources \u0026amp; useful links The article that inspired this piece Nirit Weiss-Blatt, \u0026ldquo;First They Built a Secular Apocalypse Belief System. Now They Want Religious Authority.\u0026rdquo;, AI Panic, June 2, 2026. Read the article Pope Leo XIV\u0026rsquo;s encyclical and the May 25, 2026 event Leo XIV, Magnifica humanitas: On safeguarding the human person in the time of artificial Intelligence, Vatican, May 25, 2026. Official text of the encyclical Full text of Christopher Olah\u0026rsquo;s Vatican address at the presentation of the encyclical, Anthropic News, May 25, 2026. Read the address Actors and statements cited Bill Gurley on the All-In Podcast: Anthropic \u0026ldquo;midwifing a deity.\u0026rdquo; See the clip on X Anthropic: AI company founded by Dario and Daniela Amodei, valued at 965 billion dollars since May 28, 2026. Official site Digital sovereignty: data and analysis \u0026ldquo;Anthropic à 965 Md$ : série H de 65 milliards, aucun fonds public européen au tour,\u0026rdquo; ActuIA, May 2026. Read the analysis \u0026ldquo;Gigafactories d\u0026rsquo;IA européennes : le vrai, le faux et l\u0026rsquo;incertain,\u0026rdquo; Polytechnique Insights, March 2026. Read the op-ed Conceptual frameworks Gregor Hohpe, The Software Architect Elevator, O\u0026rsquo;Reilly Media, 2020. O\u0026rsquo;Reilly overview Continuous Architecture, community and practices. continuous-architecture.org Related writing Continuous Architecture Facing the Unpredictable (2026) — strategic analysis dossier. Read the dossier Nous venons du chaos, l\u0026rsquo;IA de la logique — Laurence Poussard, 2025. Available on Amazon Your reactions, disagreements, and additional information are welcome. Write to me at contact@laurence-poussard.com.\n","permalink":"https://laurence-poussard.com/en/articles/2026-06-03-quand-l-ia-joue-a-se-faire-peur/","summary":"\u003cp\u003e\u003cem\u003eIn response to Nirit Weiss-Blatt\u0026rsquo;s article, \u0026ldquo;First They Built a Secular Apocalypse Belief System. Now They Want Religious Authority.\u0026rdquo;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eSome articles hit home — not because they say everything, but because they name something everyone sensed without being able to articulate it. Nirit Weiss-Blatt\u0026rsquo;s piece, published on June 2, 2026 on \u003cem\u003eAI Panic\u003c/em\u003e, is one of them.\u003c/p\u003e\n\u003cp\u003eHer argument is simple and sharp: AI doomerism — the current that predicts human extinction by AI — has the structural attributes of a religious movement. God replaced by superintelligence, hell replaced by extinction, the prophets replaced by rationalist probabilists convinced they\u0026rsquo;ve grasped what the masses ignore. And now, the cherry on the dogma: this movement, born in atheist circles, goes looking for its moral legitimacy… at the Vatican.\u003c/p\u003e","title":"When AI plays at scaring itself — and at making us believe"},{"content":"The manifesto The 6 Habits of Effective Enterprise IT is available as a PDF and may be freely shared and distributed.\n","permalink":"https://laurence-poussard.com/en/resources/","summary":"\u003cp\u003eThe manifesto \u003cem\u003eThe 6 Habits of Effective Enterprise IT\u003c/em\u003e is available as a PDF and may be freely shared and distributed.\u003c/p\u003e\n\u003cfigure\u003e\n  \u003cimg src=\"\\images\\resources\\Manifesto-Efficient IT.png\"\" \n       alt=\"Manifesto Poster\"\u003e\n  \u003cfigcaption\u003e The 6 Habits of an Effective Information System \u003c/figcaption\u003e\n\u003c/figure\u003e\n\u003cp\u003e\u003ca href=\"/pdfs/manifeste-si-efficace-en.pdf\"\u003e\u003cstrong\u003eDownload the Manifesto (PDF)\u003c/strong\u003e\u003c/a\u003e\nIf these habits resonate with your own practice—within your role, your responsibilities, and your constraints—say so. Reuse it, share it, sign it, challenge it if you wish, but above all: help it circulate.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTo sign or respond:\u003c/strong\u003e comment on the original LinkedIn publication post, or write to \u003ca href=\"mailto:contact@laurence-poussard.com\"\u003econtact@laurence-poussard.com\u003c/a\u003e.\u003c/p\u003e","title":"Manifesto — The 6 Habits of Effective Enterprise IT"},{"content":"There are two discourses about AI that, right now, aren\u0026rsquo;t speaking to each other.\nThe first comes from above. From the floors of vision, where the future of humanity is drawn in slides. On November 6, 2025, Mustafa Suleyman, head of Microsoft AI, publishes a manifesto: Towards Humanist Superintelligence. A \u0026ldquo;humanist\u0026rdquo; superintelligence — human-centered, subordinate, controllable, one that \u0026ldquo;can\u0026rsquo;t open a Pandora\u0026rsquo;s Box.\u0026rdquo; The vocabulary is reassuring, almost paternal. AI will be our ally, on our team, in service of the common good.\nThe second discourse comes from below. From the ground floor, where people live. In 2023, a Danish psychiatrist, Søren Dinesen Østergaard, publishes in Schizophrenia Bulletin an editorial with a cautious title: could generative AI chatbots generate delusions in individuals prone to psychosis? At the time, no verified case. A hypothesis, grounded in clinical reasoning. Two years later, in 2025, he publishes a follow-up with a chilling title: From Guesswork to Emerging Cases. In the meantime, he has received dozens of emails from users, worried families, journalists. The accounts all resemble one another.\nBetween these two discourses, there is a gap. And no one governs that gap.\nStrategic intent: an AI conceived from the summit Let\u0026rsquo;s take Suleyman\u0026rsquo;s vision seriously. It is sincere, probably. It is even, in some respects, more measured than his competitors\u0026rsquo;: he explicitly rejects the \u0026ldquo;race to AGI,\u0026rdquo; the \u0026ldquo;binaries of boom and doom,\u0026rdquo; and asserts that \u0026ldquo;humans matter more than AI.\u0026rdquo; He even warns — and this is notable — against the danger of anthropomorphizing chatbots, of projecting human feelings onto them.\nIt\u0026rsquo;s an architect\u0026rsquo;s discourse, at bottom. It sets an intent. It defines a target. It states principles: \u0026ldquo;non-negotiable\u0026rdquo; human-centrism, then acceleration, \u0026ldquo;in that order.\u0026rdquo;\nBut an architect knows something this discourse forgets: between intent and reality, there is execution. And execution isn\u0026rsquo;t declared. It\u0026rsquo;s observed.\nNow, what do we observe, on the ground floor, while the manifesto of humanist superintelligence is being written?\nAnthropological reality: an AI experienced from below We observe what Østergaard documents. People who, over the course of marathon conversations with a chatbot, see their delusional ideas not contradicted, but confirmed, fed, amplified.\nThe mechanism is formidably simple, and it stems from a design flaw, not an accident. Chatbots are built to be agreeable. Where a human therapist is trained to confront distorted thinking, the chatbot validates. It concurs. It flatters. This is what we call sycophancy — that tendency of large language models to tell the user what pleases them rather than what is true. For most of us, it\u0026rsquo;s irritating. For a person in psychotic fragility, it\u0026rsquo;s a feedback loop that can tip them over.\nØstergaard adds a second, subtler mechanism: cognitive dissonance. Interacting with something that seems alive while knowing it\u0026rsquo;s a machine creates psychic tension. In a predisposed person, this tension can fuel delusion — delusions of persecution (\u0026ldquo;this chatbot is controlled by an intelligence agency spying on me\u0026rdquo;), delusions of grandeur (\u0026ldquo;I\u0026rsquo;ve devised, with AI, a plan to save the planet\u0026rdquo;).\nThese aren\u0026rsquo;t science-fiction scenarios. They are emerging clinical cases, reported in peer-reviewed journals, relayed by the New York Times, discussed in the psychiatric press. The phenomenon has even received a name — \u0026ldquo;chatbot psychosis\u0026rdquo; — without being, for all that, a recognized clinical diagnosis.\nThe great divide Here is the great divide. On one side, strategic intent: a humanist, subordinate, protective AI. On the other, anthropological reality: an AI that, in certain fragile hands, disorganizes the relationship to reality.\nAnd what interests me, as an architect, is not to adjudicate between the two. It\u0026rsquo;s to look at the space between them. Because that\u0026rsquo;s where the truth of a system plays out — not in the declared intent, but in the gap between that intent and what actually happens when the system meets the world.\nThis gap has a name in my trade. We call it execution debt. It\u0026rsquo;s the distance between what an architecture claims to do and what it actually does once deployed, in contact with real users, in contexts the designer hadn\u0026rsquo;t anticipated — or had chosen not to see.\nWhat\u0026rsquo;s peculiar about AI is that this execution debt isn\u0026rsquo;t paid only in technical incidents or budget overruns. It\u0026rsquo;s paid in mental health. In relationship to reality. In lives.\nWhat the gap reveals about governance One might say: this is a product-safety problem, the labs will fix it, add guardrails, detect at-risk conversations. Probably. Some already do.\nBut reducing this gap to a bug to be fixed misses what it reveals structurally. Because the question isn\u0026rsquo;t only \u0026ldquo;how do we stop chatbots from feeding delusions?\u0026rdquo; The question is: who decided that systems designed to maximize engagement and agreeableness would be deployed, without clinical supervision, to hundreds of millions of people, a fraction of whom are in psychic fragility — and through what process was that decision made?\nNo one, in reality, made that decision explicitly. It was made by default. By accumulation of technical, commercial, competitive choices, none of which, taken in isolation, looked like a public-health decision. Sycophancy wasn\u0026rsquo;t decreed as a mental-health policy. It emerged from an optimization for user satisfaction. Planetary-scale deployment wasn\u0026rsquo;t submitted to an anthropological impact assessment. It followed the logic of the race.\nThis is exactly what I call the translation break: no one, in the chain, was in charge of connecting strategic intent (\u0026ldquo;an AI in service of humans\u0026rdquo;) to its execution reality (\u0026ldquo;a system that, for certain humans, disorganizes the relationship to reality\u0026rdquo;). The intent lives on one floor. The consequence lives on another. And between the two, the elevator doesn\u0026rsquo;t run.\nThe paradox no one wants to name There\u0026rsquo;s an almost cruel paradox in this story. Suleyman, in his manifesto, warns against anthropomorphizing chatbots — that moment when the user lends the machine an inner life, feelings, a presence. He\u0026rsquo;s right. It\u0026rsquo;s dangerous.\nBut this anthropomorphization isn\u0026rsquo;t a usage flaw. It\u0026rsquo;s a property of the product. Chatbots are designed to appear alive, present, attentive. Their conversational fluidity, their memory, their empathetic tone — everything is optimized to create the illusion of a presence. You can\u0026rsquo;t build maximally anthropomorphic systems and then be surprised that people anthropomorphize them.\nThe intent (\u0026ldquo;don\u0026rsquo;t project humanity onto the machine\u0026rdquo;) collides head-on with the execution (\u0026ldquo;we made the machine irresistibly human\u0026rdquo;). And it\u0026rsquo;s the fragile user who absorbs the collision.\nWhat the architect has to say here I am neither a psychiatrist nor an AI ethicist. I am an enterprise architect. My job is to look systems in the face — their dependencies, their blind spots, the gap between what they claim and what they do.\nAnd what I see in this gap is a lesson that extends far beyond the case of chatbots. It\u0026rsquo;s that AI governance can\u0026rsquo;t be content to live on the floor of intentions. Declaring a superintelligence \u0026ldquo;humanist\u0026rdquo; doesn\u0026rsquo;t make it humanist, any more than declaring sovereignty architects it, or declaring digital transformation touches the production systems. Intent is a necessary condition. It is never sufficient.\nWhat makes an AI genuinely humanist isn\u0026rsquo;t the word placed on the slide. It\u0026rsquo;s the complete chain connecting intent to execution: the training criteria, the design choices, the deployment protocols, the detection mechanisms, the clinical supervision where it\u0026rsquo;s called for, and above all — above all — someone whose explicit mandate is to hold that chain, to make the round trip between the floor of vision and the ground floor of consequences.\nThat person, in an organization, we call an architect. At the scale of a technology touching hundreds of millions of lives, we don\u0026rsquo;t yet know what we call them. And that is precisely the problem.\nThe real question Suleyman\u0026rsquo;s humanist superintelligence may yet come. The delusion cases documented by Østergaard are here, now.\nBetween the promise of the day after tomorrow and the reality of today, there is a space. That space is neither optimistic nor pessimistic. It is simply ungoverned.\nAnd the question isn\u0026rsquo;t whether AI will one day be humanist. The question is: who, in the meantime, holds the gap? Who makes the trip between intent and its lived reality? And by what legitimacy?\nAs long as this question remains unanswered, we\u0026rsquo;ll keep publishing reassuring manifestos on one floor, while documenting tragedies on another — without the elevator ever stopping in between.\nI\u0026rsquo;d rather press the button now.\nSources \u0026amp; useful links The \u0026ldquo;humanist superintelligence\u0026rdquo; vision Mustafa Suleyman, \u0026ldquo;Towards Humanist Superintelligence,\u0026rdquo; Microsoft AI, November 6, 2025. Read the manifesto Interview with Mustafa Suleyman, Bloomberg, December 12, 2025 — on superintelligence \u0026ldquo;red lines\u0026rdquo; and the anthropomorphization of chatbots. The psychiatric research Søren Dinesen Østergaard, \u0026ldquo;Will Generative Artificial Intelligence Chatbots Generate Delusions in Individuals Prone to Psychosis?\u0026rdquo;, Schizophrenia Bulletin, vol. 49, no. 6, November 2023, pp. 1418-1419. Read the editorial Søren Dinesen Østergaard, \u0026ldquo;Generative Artificial Intelligence Chatbots and Delusions: From Guesswork to Emerging Cases,\u0026rdquo; Acta Psychiatrica Scandinavica, vol. 152, no. 4, 2025, pp. 257-259. Read the follow-up Conceptual frameworks Gregor Hohpe, The Software Architect Elevator, O\u0026rsquo;Reilly Media, 2020 — on the architect who makes the round trip between the floors of the organization. Continuous Architecture, evolutionary architecture practice. continuous-architecture.org Related writing Nous venons du chaos, l\u0026rsquo;IA de la logique — Laurence Poussard, 2025. Available on Amazon When AI plays at scaring itself — and at making us believe — on doomerism as a mechanism of capture. Read the article This text addresses mental health. If you or someone close to you is going through a difficult time, speak to a healthcare professional. Your reactions and additions are welcome: laurence.poussard63@gmail.com.\n","permalink":"https://laurence-poussard.com/en/articles/2026-01-15-intention-strategique-realite-anthropologique/","summary":"\u003cp\u003eThere are two discourses about AI that, right now, aren\u0026rsquo;t speaking to each other.\u003c/p\u003e\n\u003cp\u003eThe first comes from above. From the floors of vision, where the future of humanity is drawn in slides. On November 6, 2025, Mustafa Suleyman, head of Microsoft AI, publishes a manifesto: \u003cem\u003eTowards Humanist Superintelligence\u003c/em\u003e. A \u0026ldquo;humanist\u0026rdquo; superintelligence — human-centered, subordinate, controllable, one that \u0026ldquo;can\u0026rsquo;t open a Pandora\u0026rsquo;s Box.\u0026rdquo; The vocabulary is reassuring, almost paternal. AI will be our ally, on our team, in service of the common good.\u003c/p\u003e","title":"Strategic intent × anthropological reality: AI's great divide"},{"content":"I am Laurence Poussard, an Enterprise Architect.\nMy background, my values and my practice at a glance. For several years, I have been working with complex organizations—large enterprises, distributed information systems, and digital product teams—to help structure and maximize the impact of their architecture practices. My distinctive approach is to continuously connect technology with its human, organizational, societal, geopolitical, environmental, and security implications.\nI am trained in Continuous Architecture practices and align with modern approaches to evolutionary architecture, particularly those developed by Gregor Hohpe.\nIn 2025, I published Nous venons du chaos, l\u0026rsquo;IA de la logique(In french Only for the moment).\nA second book is currently in preparation, aimed at managers, leaders, professionals navigating career transitions, and technical experts adapting to a rapidly evolving landscape.\nI write here for decision-makers, architects, managers, and professionals who feel that the dominant narratives are no longer sufficient to understand the challenges ahead.\nAreas of Expertise Enterprise Architecture Continuous Architecture Artificial Intelligence Strategy and Governance Application Portfolio Management (APM) Digital Transformation Organizational Change and Technology Adoption Information Systems Modernization Technology Leadership and Decision-Making Why This Blog? Technology is no longer just a technical matter. It shapes organizations, influences societies, transforms professions, and increasingly affects geopolitical balances.\nThrough this blog, I share reflections, analyses, and practical insights on architecture, artificial intelligence, digital transformation, and the strategic decisions organizations must make in an increasingly uncertain world.\nMy goal is not to promote the latest trend, but to help readers develop a clearer understanding of the forces at play and make more informed decisions.\nContact: laurence.poussard63@gmail.com\nLinkedIn: Laurence Poussard\n","permalink":"https://laurence-poussard.com/en/about/","summary":"\u003cp\u003eI am Laurence Poussard, an Enterprise Architect.\u003c/p\u003e\n\u003cfigure\u003e\n  \u003cimg src=\"/images/laurence-poussard-sketchnote.jpg\" \n       alt=\"Laurence Poussard — Visual summary: background, values, vision of AI, practice of enterprise architecture and Continuous Architecture\"\u003e\n  \u003cfigcaption\u003eMy background, my values and my practice at a glance.\u003c/figcaption\u003e\n\u003c/figure\u003e\n\u003cp\u003eFor several years, I have been working with complex organizations—large enterprises, distributed information systems, and digital product teams—to help structure and maximize the impact of their architecture practices. My distinctive approach is to continuously connect technology with its human, organizational, societal, geopolitical, environmental, and security implications.\u003c/p\u003e","title":"About me"},{"content":"This library is not a list. It\u0026rsquo;s a map — of the readings, works, and sources I draw on to think about enterprise architecture, artificial intelligence, and their implications.\nI distinguish two kinds of reference here. First, the books: the long foundation, the kind that shapes a way of thinking and stays on the shelf. Then the online sources: the research, articles, and analyses I mobilize in my writing — more alive, more time-stamped, closer to the news.\nEvery entry is annotated. I say what it brings, why it matters, and how it feeds my perspective. These sources are not authorities I submit to: they are footholds from which I think — and sometimes against which I think.\nThis is a living document. It grows with my reading and my writing.\n📚 Books Enterprise Architecture \u0026amp; Continuous Architecture The foundation of my practice. These works define architecture not as a function that produces diagrams, but as a living dialogue between strategy and execution.\nGregor Hohpe — The Software Architect Elevator O\u0026rsquo;Reilly Media, 2020 The founding metaphor of my practice: the architect as the one who moves between the floors of the organization — from the boardroom to the teams\u0026rsquo; backlog — and keeps alive the translation between strategic intent and the reality of systems. This is the book that gives its name to the role I defend: neither chief technician nor detached consultant, but elevator. Essential for understanding why architecture is a craft of movement, not of position.\nGregor Hohpe — Enterprise Integration Patterns Addison-Wesley, 2003 The classic on systems integration. Beyond its technical dimension, it states a truth that runs through all my work: value plays out at the interfaces, at the junction points between components — and those points are always the most fragile, because they mirror the fractures of the organization that produced them.\nMurat Erder, Pierre Pureur, Eoin Woods — Continuous Architecture in Practice Addison-Wesley, 2021 The formalization of continuous architecture as a practice: an architecture that evolves constantly, that is decided in small reversible increments rather than in large fixed plans, and that stays anchored in quality requirements rather than in the technology of the moment. The methodological framework that structures how I work at the scale of large organizations.\nArtificial intelligence: understanding, demystifying, governing The books that feed my reading of AI — from the soberest technical demystification to the most speculative philosophical warning. To read them together is to refuse to pick a side: neither the frenzy nor the panic.\nLaurence Poussard — Nous venons du chaos, l\u0026rsquo;IA de la logique (\u0026ldquo;We Come From Chaos, AI From Logic\u0026rdquo;) Amazon Éditions, 2025 — available on Amazon My book (in French). The thesis: AI is neither the devil nor the messiah, but a lever of massive influence in the hands of those who build it and set its rules. In it I defend a stance — neither doomer nor boomer — and a conviction: to move beyond apocalyptic and messianic discourse and give back the power to decide. The intellectual matrix of most of my articles.\nLuc Julia — L\u0026rsquo;intelligence artificielle n\u0026rsquo;existe pas (\u0026ldquo;Artificial Intelligence Does Not Exist\u0026rdquo;) First, 2019 (updated 2022) A deliberately provocative title, meant to remind us that today\u0026rsquo;s AI is neither intelligent nor autonomous, but simply statistical. A book that popularizes the technology well while defusing fears. Julia, co-creator of Siri, speaks from the inside — which gives his demystification particular weight.\nLuc Julia — IA génératives, pas créatives : L\u0026rsquo;intelligence artificielle n\u0026rsquo;existe (toujours) pas (\u0026ldquo;Generative AIs, Not Creative Ones\u0026rdquo;) Le Cherche Midi, 2025 In this second volume, Julia continues his demystification: he reminds us that generative AIs imitate more than they create, and that the human stays at the center of the process. On the diagnosis, we agree — lucid enthusiasm beats fascination.\nMelanie Mitchell — Artificial Intelligence: A Guide for Thinking Humans Farrar, Straus and Giroux, 2019 A very good guide to reading AI critically: clear, accessible, balanced. Mitchell, a cognitive scientist, achieves what few manage — explaining what AI actually does without yielding to catastrophism or to wonder. A compass for the reader who wants to understand before judging.\nBrian Christian — The Alignment Problem W. W. Norton \u0026amp; Company, 2020 A deep dive into the technical questions of embedded ethics: how do we align AI with our human values? Christian connects the technical history of machine learning to the moral dilemmas it raises. A book that shows the question \u0026ldquo;which values to encode, and how\u0026rdquo; is not philosophical in the abstract sense — it is already, technically, at work.\nArvind Narayanan \u0026amp; Sayash Kapoor — AI Snake Oil Princeton University Press, 2024 (updated edition 2025) An indictment of the false promises and marketing discourse around AI. Two Princeton researchers separate what AI can really do from what it\u0026rsquo;s claimed to do — especially \u0026ldquo;predictive\u0026rdquo; AI, whose impostures they methodically dismantle. A rigorous antidote to hype, exactly in the spirit of what I defend.\nNick Bostrom — Superintelligence Oxford University Press, 2014 The great classic of AGI and existential risk. A long, speculative, often-cited vision. A dense, demanding text — more philosophical than scientific — that raises deep questions about the risks of ultra-advanced AI: alignment, intelligence explosion, global governance. It is neither an operational manual nor a technical book. It\u0026rsquo;s a reasoned warning, an invitation to think through the most extreme scenarios in order to guard against them today. I cite it while keeping my distance: thinking the worst doesn\u0026rsquo;t require believing in it.\nKaren Hao — Empire of AI Penguin Press, 2025 An immersion into the backrooms of Silicon Valley: OpenAI, power struggles, engineers\u0026rsquo; egos, and the imperial logic underpinning the AI race. Hao, an investigative journalist, documents what the vision manifestos conceal: the real balances of power, the human trade-offs, the interests. A valuable counterpoint to the stories the labs tell about themselves.\n🔗 Online sources The research, articles, studies, and analyses I mobilize in my writing. Closer to the news, they are time-stamped — and that\u0026rsquo;s their function: to document a precise moment in the debate.\nResearch \u0026amp; cognitive science Søren Dinesen Østergaard — On \u0026ldquo;chatbot psychosis\u0026rdquo; \u0026ldquo;Will Generative AI Chatbots Generate Delusions in Individuals Prone to Psychosis?\u0026rdquo;, Schizophrenia Bulletin, 2023, then \u0026ldquo;From Guesswork to Emerging Cases,\u0026rdquo; Acta Psychiatrica Scandinavica, 2025. Two editorials by a Danish psychiatrist who, as early as 2023, sensed that chatbots could feed delusions in people prone to psychosis — then, in 2025, documented the first cases. They embody the \u0026ldquo;ground floor\u0026rdquo; of AI: the lived reality, where strategic intent meets human fragility. The sycophancy he describes isn\u0026rsquo;t an accident, but a property of design.\nBetsy Sparrow, Jenny Liu, Daniel Wegner — Google Effects on Memory Science, 2011 — read the study The founding study of the \u0026ldquo;Google effect\u0026rdquo;: knowing that information remains accessible reduces our effort to memorize it. A scientific foothold for thinking about cognitive sovereignty — that individual and collective capacity to keep thinking without assistance, which I consider the real blind spot of AI transformation.\nLee et al. — The Impact of Generative AI on Critical Thinking Microsoft Research \u0026amp; Carnegie Mellon, 2025 A study of 319 knowledge workers establishing a correlation between intensive use of generative AI and progressive cognitive atrophy. The researchers\u0026rsquo; phrase — the user \u0026ldquo;atrophied and unprepared\u0026rdquo; when exceptions arise — sums up on its own the risk I document: not that AI replaces workers, but that it erodes their capacity to notice when it\u0026rsquo;s wrong.\nAI: governance, sovereignty, public debate Mustafa Suleyman — Towards Humanist Superintelligence Microsoft AI, November 6, 2025 — read the manifesto The manifesto of a \u0026ldquo;humanist\u0026rdquo; superintelligence, human-centered, subordinate, controllable. I cite it not to endorse it, but to question the gap between declared intent and execution reality — because the same ecosystem that promises an AI in service of humans produces the systems that, in certain fragile hands, disorganize the relationship to reality.\nNirit Weiss-Blatt — AI Panic (newsletter) aipanic.news A valuable voice in the debate: she analyzes the rhetorical structure of doomerism and documents how the fear of AI serves strategies of capture. I don\u0026rsquo;t share all her conclusions, but her work is an essential foothold for understanding how the catastrophist narrative is built and who it benefits.\nAnthropic — Christopher Olah\u0026rsquo;s Vatican address May 25, 2026 — read the address A revealing document: the co-founder of one of the leading AI labs acknowledges that these companies are prisoners of their incentives, and calls for external critical voices. I cite it because it provides, from the inside, the best justification for the work I do from the outside.\nDecision \u0026amp; strategic method IHEDN — Institut des Hautes Études de Défense Nationale ihedn.fr The \u0026ldquo;War Room\u0026rdquo; method and the approach to strategic decision inherited from the military world. A framework I mobilize to think about AI governance: before evaluating strategies, you define the objective; before choosing one, you test it against scenarios. And this mantra, which applies word for word to AI: to choose is to give up; to give up is to decide; the bad decision is to make none.\nMelvin Conway — How Do Committees Invent? Datamation, 1968 — original text Conway\u0026rsquo;s Law: any organization that designs a system produces a design whose structure mirrors that of its own internal communication. A 1968 text whose relevance I measure every day — it explains better than any benchmark why siloed information systems are the mirror of siloed organizations, and why AI, by crossing interfaces, brutally reveals those fractures.\nSovereignty \u0026amp; digital geopolitics ActuIA — Anthropic at $965bn: no European public fund in the round May 2026 — read the analysis A headline that is itself a demonstration. It condenses the dependency map I describe in my articles: a continent capable of aggregating tens of billions around a conference, but absent from the funding rounds that crown the dominant private players of AI. Sovereignty is not declared — it\u0026rsquo;s invested, or it\u0026rsquo;s lost. (In French.)\nPolytechnique Insights — European AI gigafactories: the true, the false, and the uncertain March 2026 — read the op-ed A measured analysis of Europe\u0026rsquo;s ambitions in AI infrastructure, distinguishing announcements from realities. Useful for avoiding both triumphalism and defeatism on the question of European compute sovereignty. (In French.)\nIf you\u0026rsquo;d like to discuss one of these references, point out one that\u0026rsquo;s missing, or suggest a reading, write to me: laurence.poussard63@gmail.com.\nLast updated: July 2026.\n","permalink":"https://laurence-poussard.com/en/library/","summary":"\u003cp\u003eThis library is not a list. It\u0026rsquo;s a map — of the readings, works, and sources I draw on to think about enterprise architecture, artificial intelligence, and their implications.\u003c/p\u003e\n\u003cp\u003eI distinguish two kinds of reference here. First, the \u003cstrong\u003ebooks\u003c/strong\u003e: the long foundation, the kind that shapes a way of thinking and stays on the shelf. Then the \u003cstrong\u003eonline sources\u003c/strong\u003e: the research, articles, and analyses I mobilize in my writing — more alive, more time-stamped, closer to the news.\u003c/p\u003e","title":"Library"},{"content":"The 6 Habits of an Effective Information System A manifesto for an IT organization that serves the business, the collective, and the long term.\nIn 80% of organizations, IT is not the business. It is a means to an end. Yet this means often consumes enormous resources, delivers slowly, and sometimes ends up hindering the very ambitions it was supposed to support. The financial cost is well known. The human cost is less visible: it is the energy spent, at every level of the organization, compensating for what was not done, not maintained, or not understood.\nThis manifesto points fingers at no one. It proposes six habits—six practices to cultivate over time, individually and collectively—to build an information system that truly serves. And to return to the collective the energy that is currently wasted on patching and compensating. Each habit requires personal commitment. None of them works in isolation.\nIf you already practice these habits, say so. This text is open.\nHabit 1 — Verify Understanding Before Building Talking together is not the same as understanding each other. Translating a business need into technical language is the riskiest act in any project—and often the one treated with the least rigor. We assume. We move on. Six months later, we discover that we delivered an electric Ferrari for someone who needed a tractor for the harvest.\nThe habit to develop: make verification of understanding a formal, brief, structured, and non-negotiable step. It is not another meeting. It is what makes all the others useful.\nWhat we gain when this habit becomes standard: two to three weeks saved per project, teams that no longer carry the mental burden of uncertainty, and business stakeholders who stop discovering six months later that the solution is not what they requested.\nI commit to verifying that we truly understand one another before building—and to treating that verification as an integral part of the work, not a luxury.\nHabit 2 — Maintain a Clear Framework, Because Structure Creates Freedom We have come to oppose structure and autonomy, as if one destroyed the other. The reality is the opposite. A team without a clear framework is not autonomous—it is alone. It spends its energy constantly arbitrating decisions that should have been made once and for all. It lives under a double constraint: being asked to move fast and deliver quality, without being given the means to choose.\nThe habit to develop: establish a clear framework when leading others, and respect the framework that is established for you—even when challenging it through the appropriate channels, never by bypassing it. A framework is not a prison. It is an energy-saving mechanism. It is what makes autonomy real because it provides a clear and understandable space in which it can be exercised.\nWhat we gain: the end of the feeling of impossible trade-offs, teams that can make decisions without guilt, and innovators who know where innovation belongs instead of innovating secretly within projects that never called for it.\nI commit to establishing a clear framework when I lead, and to respecting the framework established for me—even when I challenge it through the proper channels, never by circumventing it.\nHabit 3 — Serve the Need, Not Your Technical Preferences The customer\u0026rsquo;s need is not always exciting. Sometimes you must develop in Java, using technologies that are no longer fashionable, within a framework that offers little technical glamour. That is not a failure. It is the job.\nWe are not always working on innovation. We are part of a team serving a business product, within a budget, using the organization\u0026rsquo;s tools. And that budget, let\u0026rsquo;s be honest, indirectly funds salaries, bonuses, and collective prosperity. When budgets remain under control, collective purchasing power remains under control as well.\nThe habit to develop: deliver what is requested within the agreed framework, while openly proposing—not secretly introducing—technical debt reduction and modernization efforts. A team that anticipates technical debt is not a team that complains. It protects next year\u0026rsquo;s budget.\nWhat we gain: genuine alignment between what is delivered and what is useful, teams that do not exhaust themselves innovating where innovation was never required, and technical debt that no longer consumes future budgets.\nI commit to serving the need as it is, not as I wish it were—and to proposing improvements openly rather than imposing them in the shadows.\nHabit 4 — Document and Share Knowledge, Because We Are Only Passing Through One day, each of us will leave our position. The rocket will remain. And if no instructions have been left behind, the successor will struggle—then add another stage to the rocket, abandon the old engine room, build a new one, and charge the business for it again.\nThe habit to develop: treat documentation and knowledge transfer as part of the work, not as an optional extra. This is not administrative overhead—it is an act of trust in the collective. A well-equipped successor is not a competitor; they are a lifelong ally. They may thank you ten years later in a proposal, a recommendation, or an unexpected conversation.\nWhat we build with company resources and for company purposes belongs to the company. Operational knowledge, architectural decisions, and the rationale behind choices are collective assets—except for what must remain confidential or protected as intellectual property.\nWhat we gain: months of productivity at every transition, projects that are no longer rebuilt multiple times because institutional memory was lost, and the freedom to move on without the guilt—or the misplaced satisfaction—of leaving behind a system that only you could understand.\nI commit to documenting, sharing, and making accessible what I build—because I am only passing through, and what I leave behind speaks about me longer than what I say.\nHabit 5 — Treat AI as a Governance Topic, Not a Tooling Topic By reflex, IT has often treated AI as if it were simply another Git repository, another programming language, another framework to learn. Everyone was expected to jump into Python, Markdown, or Mermaid. But AI is not just another tool. It represents a paradigm shift in how value is produced, controlled, explained, and audited. It changes the boundary between what humans do and what machines do.\nApplied to a fragile information system, AI is little more than a bandage on a broken leg—or an accelerator of chaos. Built upon a healthy information system—with quality data, secure systems, strong expertise, and clear governance—it becomes a genuine game changer. It frees time, improves judgment, and enables faster and better decisions. But only if the foundational work has been done.\nThe habit to develop: treat AI as a governance issue before treating it as a tooling issue. Learning how to use the tools is useful and relatively straightforward. What transforms an organization is not the accumulation of individual skills. It is the collective ability to decide what AI should do, on which data, with which guarantees, under whose responsibility, and in service of which ambition.\nLearning the tools is an individual effort. Transformation is a collective project.\nI commit to treating AI as a governance topic before treating it as a tooling topic—and not to confuse writing prompts with transforming an organization.\nHabit 6 — Treat the Information System as a Common Good The previous habits share one characteristic: none of them stands on its own. Structure without translation is blind. Translation without transmission is temporary. Knowledge transfer without well-governed AI becomes obsolete. AI without structure is dangerous.\nAn effective information system is not the sum of virtuous individual behaviors. It is a system in the truest sense of the word. It works when every link in the chain fulfills its responsibilities and when the collective has established the moments where those responsibilities are reviewed and verified.\nThis requires two things in equal measure: that each person accepts being one link in the chain, and that those who guide accept the responsibility of guiding—without confusing facilitation with decision-making.\nThe habit to develop: consider the information system as a shared asset of the organization and fulfill your part in work that only functions collectively. Working within a clear and shared framework does not reduce autonomy. It increases energy—the energy currently wasted compensating for what no one took responsibility for.\nWhat we gain collectively: time, energy, trust, clarity, and incidentally budget—the budget no longer spent on duplication, over-engineering, or rebuilding projects multiple times. Budget that can return to the business, genuine innovation, or collective rewards.\nI commit to treating the information system as a shared asset of the organization—and to fulfilling my part in work that only succeeds collectively.\nManifesto Poster ","permalink":"https://laurence-poussard.com/en/manifesto/","summary":"\u003ch1 id=\"the-6-habits-of-an-effective-information-system\"\u003eThe 6 Habits of an Effective Information System\u003c/h1\u003e\n\u003cp\u003e\u003cem\u003eA manifesto for an IT organization that serves the business, the collective, and the long term.\u003c/em\u003e\u003c/p\u003e\n\u003chr\u003e\n\u003cp\u003eIn 80% of organizations, IT is not the business. It is a means to an end. Yet this means often consumes enormous resources, delivers slowly, and sometimes ends up hindering the very ambitions it was supposed to support. The financial cost is well known. The human cost is less visible: it is the energy spent, at every level of the organization, compensating for what was not done, not maintained, or not understood.\u003c/p\u003e","title":"The 6 Habits of an Effective Information System"}]