Tradition as the Infrastructure of Modernity

At the start of this essay, I tried to say why I came back. Why a path that had taken me from Belfort to Paris, and from Paris into corporate roles in several major European capitals, ended up bringing me back to Haute-Nendaz. I did not know, at that point, whether the answer I was about to give would hold up over the length of the chapters that followed. I return to it here, to answer it differently, in light of what has since been examined.

The personal answer is simple, and it commits no one but me. I came back because it had become possible, and because what had become possible suddenly revealed what had never been: that one could, without giving up the productivity of the present, live in a place to which one belongs both by descent and by choice. The AI transformation gave me far more than technological comfort — a reconciliation between what I wanted to do and where I wanted to do it, one that had no solution in the world before, and that has one now.

This personal answer holds for a few thousand people, perhaps a few tens of thousands, across Western Europe who could, in the decade ahead, make a journey comparable to mine. A mass movement? No, and it never will be. But it shifts the terms for the territories capable of taking these people in. And it is this movement, aggregated at the scale of the decisions that steer it, that can make the difference between a Valais that becomes what this essay has called a showcase, and a Valais that remains an actor in its own right.

What the preceding chapters have tried to say

I set out, in the early chapters, a diagnosis in two movements. The AI rupture, far from reducing to one more wave of automation, partly dissolves a geographic advantage that had defined the world economy for a century. This dissolution opens a rare territorial window, one that will not stay open indefinitely and does not open for everyone: only territories that position themselves actively will catch this movement, and Valais is among those that can, provided it decides to.

To this dissolution is added, and this is what the economic chapters set out to make visible, a competitive shift of a different nature. Well-orchestrated AI simultaneously flattens the differentials that had worked against Swiss territories for twenty years in the economy of skilled services: the cost differential with offshoring, the price differential with generic software solutions, the rigidity of standardised solutions faced with specific needs. A qualified local team, equipped with the right tools, can now produce under economic conditions it had not enjoyed since the industrialisation of services. This shift does not trigger itself. It requires, as this essay has restated chapter after chapter, the professional experience accumulated by seniors — the kind that knows how to architect a problem, orchestrate production, guarantee the quality of the final output. It is this experience, more than AI itself, that becomes the strategic lock on the new cognitive economy.

I then wanted to show that the canton holds, for this transition, a rare institutional and cultural capital. Not merely an endearing heritage: a grammar of governance (bourgeoisies, consortages, alpine-pasture associations, Swiss federalism) that answers questions urban modernity still struggles to state clearly, and one that has earned, through Törbel and Elinor Ostrom, a worldwide scientific recognition now drawn on by thinkers of the digital commons. This grammar is completed by an intangible and linguistic heritage of rare density (bilingualism, Walliserdeutsch, know-how listed by UNESCO) — a non-relocatable asset whose relative value grows as generic cultural output becomes infinitely reproducible by machines.

I then followed this transformation to where it is already unfolding, in every Valais trade, from the most ancestral to the most contemporary, each time in a form specific to it yet coherent with the whole: narrative multiplication for wine, instrumented signature for the dairy sector, personalisation at scale for hospitality, access to bespoke service for fiduciaries, demographic compensation for valley medicine. Each chapter argued this in detail.

For every one of these domains, the stakes are political long before they are technological: who gets to decide on the tools, on control of the data, on the burden of transmission, on funding the equipping of the workforce already in place — and, more precisely, the priority equipping of the experienced seniors who know how to architect and orchestrate, the ones who unlock everything else.

In the chapters of the fourth part, I wanted to go all the way down to the decisions that follow from this. A realistic digital-sovereignty policy, sized to a canton that is neither Bern nor Zurich, distinguishing strategies of substitution (doomed to fail) from strategies of complement (which can succeed). A training scheme for the working population that takes on a counter-intuitive priority: equipping the seniors who orchestrate rather than multiplying courses for young graduates. A demographic policy that thinks retention, attraction and transmission together, and recognises that the canton's new strategic target — mid-career workers capable of architecting and orchestrating — differs from the target of traditional approaches in this area.

None of these proposals is revolutionary. That is probably their flaw with readers looking for something spectacular, and probably their merit with those who have to decide.

The false dilemma, revisited

I stated in the first chapter that the Valais debate about the future often remains trapped in a dilemma that is not one — the one that pits tradition against modernity. This point can now be restated in light of what the chapters have clarified.

Valais tradition, taken seriously, is not the opposite of modernity. In its institutional, cultural and linguistic dimensions, it is what makes modernity governable at all. A bourgeoisie — one of these civic corporations holding forests in common — that has held its woodland for seven centuries carries a long temporality that becomes precious in an age of ephemeral platforms. A consortage that allocates water among rights-holders under rules set down in writing in the sixteenth century carries a grammar for managing common goods that today inspires thinkers of the open digital commons. And a patois that dies out takes with it a way of naming landscape and gesture that cannot be reconstituted, and that only active policy, backed by the very tools that threaten regional languages, can still preserve.

This tradition, taken seriously, is also what makes possible the reading that structures this essay. Generative AI acts, in the cognitive field, the way a bisse would act at continental scale — one of those centuries-old irrigation channels that capture water where it is abundant and carry it to where it is scarce. It captures a resource where it is abundant, carries it to where it is lacking, makes it available to territories that had been deprived of it. The canton that invented the grammar of the stone bisses seven centuries ago is, whether by chance or by destiny, particularly well placed to understand the cognitive bisse now opening, and to govern it with the same institutions that have proved themselves on water and alpine pasture.

Conversely, a modernity governed by no living tradition produces interchangeable territories, folkloric showcases, economies that quietly empty out from within. The linguistic standardisation driven by large language models, the erosion of the regional professional fabric by global platforms, the Disneyfication of festivals, the second home asleep eleven months out of twelve: in all of these one should see, not tradition resisting modernity, but a modernity never linked to tradition, reduced, for want of that link, to absorption into a uniform current whose beneficiaries sit elsewhere.

This essay's wager is that Valais can do otherwise, and that wager owes nothing to nostalgia, to conservatism, or to distrust of technology: it rests on a clear-eyed reading of the strengths the canton has at its disposal. The window is open. The institutions are alive, the intellectual and linguistic assets exist, the technology has matured, migration flows are favourable. All that is missing is the decision that ties these elements together.

A political question, at human scale

From all of the above emerges a central question, and I want to put it as directly as possible. It is neither technological, nor economic in the narrow sense. It is political, in the original meaning of the word: who decides, here, and when.

In Valais, this question is never asked in a vacuum. It is put to municipal councils, to a Council of State, to the presidents of bourgeoisies, to heads of institutions, to business leaders, to federal elected officials, to professional associations, to families deciding whether or not to send their children to study outside the canton, to workers deciding whether or not to return, to newcomers who will or will not get involved. This dispersal of decisions is, against all appearances, the canton's greatest asset: it is consistent with the institutional grammar that has made it distinctive since the Middle Ages. It is also its greatest difficulty, because dispersal can turn into dilution if nothing coordinates it.

What this essay has tried to sketch is less a policy in the centralising sense of the word than a compass — a handful of bearings that scattered decisions could take as their reference, so as to produce, over ten years, an overall coherence. Resident, skilled human capital is the central question, and it conditions all the others. The competitive shift AI is triggering can only be captured by equipping the experienced seniors who know how to architect and orchestrate. Practical sovereignty over everyday uses matters more than large abstract infrastructure. A specialisation deliberately built on what Valais does better than others beats a scattering of efforts with no horizon. Rather than one more institution, what is needed is a hand to hold together what already exists. Conscious transmission, in the age of large language models, becomes the very condition of cultural continuity. And no one should be left outside this transformation — not out of ideology, but because without that, nothing holds together.

None of these proposals is, strictly speaking, my invention. They have circulated, fragmentary and scattered, in Valais public debate for years. What these pages have tried to do is bring them into a coherence that mobilises more than the sum of their separate statements.

And what if I am wrong?

Every forward-looking essay runs the risk of being wrong. This one will not escape that risk, and I would rather name up front the three serious objections one can raise against it, because ignoring them would be both dishonest and strategically weak. If the argument holds, it should withstand its critics; if it does not hold, better to know it now.

A first objection: what if generative AI were merely a bubble, its real economic reach overestimated the way that of cryptocurrencies and NFTs was in their time, or that of virtual reality? The objection rests on real and documented cycles of technological enthusiasm, and I take it seriously. Short of absolute certainty, the concrete uses of generative AI are already measurable; they rest less on a promise than on productivity gains actually observed internally, within my own group and among my clients. The trajectory of model improvement, over the past three years, has followed a curve few observers anticipated even in 2022, and nothing suggests it is levelling off in the near term. Above all, a bubble-and-collapse scenario would still leave behind the uses already embedded, just as the dot-com bubble of 2000 left behind the infrastructure and the practices that made the following decade possible. The canton that positions itself now takes a limited risk; the one that waits takes a greater risk — that of missing the window if the current trajectory holds.

A second objection follows: what if, contrary to what this essay argues, the real dynamic led not to a territorial dilution of cognitive capital, but to an even harsher re-concentration in existing urban hubs? One can argue, and some do, that AI tools require concentrations of capital, talent and infrastructure found only in major metropolises, and that they will concentrate further rather than redistribute. My answer is that the two movements are not mutually exclusive. Urban hubs will likely continue to capture the bulk of cutting-edge innovation, funding, and the youngest talent; what they will no longer be able to keep is the exclusive claim on the capacity to produce skilled work of medium-to-high quality. That exclusivity was what sustained the metropolitan advantage within the real economy. If it dissolves, as everything suggests it will, the relative position of non-metropolitan territories improves, without any of them becoming a hub in its own right. A canton that misses this nuance risks resigning itself wrongly, or deceiving itself in the opposite direction.

A third objection, finally: what if European and Swiss AI regulation, through excessive caution, slowed the adoption of these tools to the point of sterilising the advantages one might expect from them? This one calls for the greatest clear-sightedness, because it rests less on conjecture than on the observation of trends already under way. The European AI Act, sector-specific restrictions in health and finance, constraints around personal data can slow deployment, and will probably slow certain uses. That said, Switzerland is not in the position of EU member states, and the canton of Valais even less so. The room for regulatory manoeuvre, at Swiss and cantonal level, remains greater than the continental debate suggests. And should regulation genuinely slow deployment massively at the European scale, Valais, like any Swiss territory, could draw from that a relative advantage that territories bound by the AI Act would not enjoy. The risk is real. It cuts both ways.

None of these objections invalidates the thesis this essay defends; all of them must stay present while the decisions get made. Serious foresight builds its risks into the robustness of its own proposals, rather than ignoring them. I believe the proposals made in these pages hold, in their great majority, even should one of these three objections turn out to be partly confirmed. Were all three to hold at once and in full, then yes, this essay would have overestimated the opportunity. But given the current state of the technological, economic and regulatory debate, the window I describe is more likely open than closed.

For those who might come

This book is addressed, as I said from the first chapter, primarily to decision-makers in Valais. I want, though, to turn for a moment to another reader: the one who, in some large European city, might read these pages wondering whether a path like mine could be theirs too.

This reader is not looking at Valais alone. Other Alpine and peri-Alpine regions of Europe have launched comparable policies, and the canton is acting within a regional landscape that is actively reshaping itself. Austrian Tyrol and Vorarlberg have built, around their areas of specialisation (precision industry, energy, technical know-how), a dense economic fabric that already draws in skilled workers. French Haute-Savoie combines its proximity to Geneva with a resettlement dynamic already well under way, though one running into severe housing pressure that Valais would do well to study closely. Italian Trentino has invested for a long time in its institutional autonomy and in its research fabric built around the University of Trento, and shows that a coherent regional policy can produce measurable effects over ten to fifteen years. Further afield, along the Atlantic arc, the Spanish regions of Asturias and Cantabria have pursued, for a decade now, rural-revitalisation policies well worth studying closely, for what they teach about the patience required and the resources that must be mobilised⁴⁶. None of these regions occupies exactly the position Valais does, which holds assets these neighbours do not share: community institutions of rare density, hydropower as an asset of energy and digital sovereignty, a central position along the Alpine arc, structural bilingualism. But none of them is standing still, and the window this essay describes is opening for them just as it is for Valais. The canton that acts will not be acting into a void: relative advantage, in this landscape, is being built decision by decision.

To this reader, I owe honesty. Valais is no El Dorado. It has its administrative heaviness, its circles that can feel closed, its demanding winters, its short summers, its distance from the major cultural stages, its rising property prices, its languages and dialects that can throw a newcomer off balance. It also has a quality of life that no statistic captures but that one senses within the first few months — a light unlike any other, a landscape that reminds you at every walk how small you are, institutions one can approach without abstraction, neighbours one comes to know, a community that, if you accept its rules, can become your own.

It is not for me to say that one must come, only that it has become possible. The decade now opening is, to my mind, the first in recent history in which a path like the one I followed becomes realistic for a significant number of people. And the territories that measure this not in speeches but in concrete decisions about their connectivity, their tax policy, their housing, their institutions, will capture a momentum that others will merely watch go by.

A letter to the future

I close these pages thinking of those who will read them ten years from now. By then, the transformation this essay has tried to describe will have largely played out, one way or the other. Either Valais will have become an active territory: it will have captured a significant share of the dynamic of skilled resettlement, equipped its workforce and its institutions for the new cognitive economy, passed on what needed passing on while taking in what needed taking in. Or it will have become a showcase, still beautiful, still visited and photographed, but one that let slip the window that opened in the mid-2020s, watching its young people leave for the urban hubs and its traditions slide into folklore.

This fork in the road is not already written. It is being decided now, in decisions that will never make headlines but that, taken together, will trace the slope the canton follows. No chapter of this essay offers a miracle solution, because there is none. All of them propose concrete decisions, at achievable scales, with means compatible with the canton's resources. Taken separately, none of these decisions changes the fate of Valais. Taken together, and pursued patiently over ten years, they can significantly bend its trajectory.

What the bourgeoisies of Valais have done for seven centuries with their forests and their alpine pastures — managing a collective inheritance over time without turning it into merchandise or handing it to some distant state — is exactly what the canton must now do with its digital, linguistic, institutional and demographic inheritance. This continuity is a political instruction far more than a poetic metaphor. It requires accepting that one should not wait for solutions to arrive from elsewhere, taking seriously what one already has, deciding at whatever scale one can, and passing on to the next generation what must be passed on.

I told, in the first chapter, of the fibre-optic cable running a few metres above my house, alongside the bisse Vieux — one of the canton's old stone-built water channels — and of the inattention that means the two channels are never looked at together. Learning to look at them together: that is this book's programme, and it no longer belongs to me. It lies in the decisions that others, not I, will have to make in the months and years ahead. These pages have tried to give them a few arguments to work with.