The Wrong Problem
The architecture of the social internet taught us to mistake attention for relationships.
Every era leaves behind a few essays that continue to illuminate the decades that follow. They possess the unusual quality of becoming more relevant with time rather than less. Kevin Kelly’s 1,000 True Fans is one of those essays.
The argument was almost disarmingly simple. A creator does not need millions of followers. A musician, writer, teacher, or artist needs only a relatively small number of people who genuinely value the work, return to it consistently, recommend it to others, and willingly support it over many years. Kelly’s insight quietly shifted the question from scale to relationships. The internet, he suggested, was valuable not because it could reach everyone, but because it could finally connect creators with the people who cared deeply about what they made.
Looking back nearly two decades later, I think the idea was brilliant.
The architecture that followed, however, failed to honor it.
Rather than helping creators cultivate enduring relationships, the social internet gradually optimized for something much easier to measure. Follower counts became public. Likes became currency. Reach became the objective. Every improvement in recommendation systems, every refinement in algorithmic prediction, every reduction in friction made platforms increasingly effective at distributing attention across enormous populations. The internet became extraordinarily good at introducing strangers to one another.
Somewhere along the way, however, we quietly began behaving as though attention and relationships were interchangeable. They are not. Attention may create the first encounter, but relationships emerge through an entirely different process. They require repeated encounters, sustained conversations, shared interests, memory, patience, and the quiet confidence that develops when someone continues to be thoughtful, reliable, and worth returning to. Trust is simply the name we give to a relationship that has endured long enough to become dependable.
The distinction matters because the internet solved one problem while gradually obscuring another. It became a remarkable machine for distributing attention, yet remained surprisingly indifferent to the slower architecture through which relationships are actually formed. That realization eventually led me to ask a different question, one that had far less to do with algorithms than with architecture itself.
What kind of digital world would we build if our objective were relationships rather than attention?
Architecture as Destiny
If relationships and attention are fundamentally different, then the obvious question is why the internet became so much better at one than the other.
The answer, I think, lies less in psychology than in architecture.
Human beings respond to incentives, but incentives are themselves products of design. Every architecture quietly rewards certain kinds of behavior while making others less likely. Roads determine how cities expand. Classrooms shape how students learn. Markets influence how companies compete. We often attribute these outcomes to culture or human nature, forgetting that culture itself is frequently downstream of the environments we create.
We usually think of architecture as something that organizes buildings or software. Some of humanity’s most enduring institutions begin by answering a simple question: how do people come to know, trust, and learn from one another? Universities determine which scholars encounter one another. Markets determine which strangers learn to trust. Scientific communities determine how criticism and collaboration unfold. The design changes, but the underlying question remains remarkably constant: who meets whom, under what conditions, and for how long?
The architecture of the modern social internet rewarded broadcasting over conversation because broadcasting scales more easily. It rewarded reaction over reflection because reactions are immediate while reflection unfolds slowly. It rewarded prediction over curiosity because algorithms can optimize for what we are likely to click, but struggle to cultivate what we have not yet learned to seek. Above all, it rewarded visibility because visibility is measurable, while relationships resist measurement almost entirely.
None of this required malice. Platforms optimized for the objectives they were given.
If success is measured through engagement, architecture naturally evolves toward engagement. If growth is measured through reach, architecture naturally evolves toward reach. Once those objectives are established, many of the design decisions that define modern social media begin to feel almost inevitable. Public follower counts, infinite scrolling, frictionless sharing, algorithmic recommendations, and viral distribution are not random features. They are logical consequences of a system designed to maximize attention.
The interesting question is what happens when we change the objective.
Suppose the goal were no longer to maximize attention but to cultivate relationships. Suppose success were measured by whether conversations continued months later, whether people returned to one another’s work, whether intellectual friendships emerged, and whether communities developed shared memory instead of endlessly consuming novelty.
Not because our values had changed, but because our objective had.
That realization eventually led me to a simple conclusion. The principles that define digital spaces are never arbitrary. They are consequences of the question we ask at the beginning. Change the question, and the architecture changes with it.
The Small Web began as an attempt to ask a different question.
Deriving The Small Web
Once I began thinking about digital spaces in terms of relationships rather than attention, something unexpected happened. The principles of what I had been calling The Small Web no longer felt like a collection of interesting product ideas. They began to feel like consequences. If the objective became relationships rather than attention, they followed almost inevitably.
If the objective is maximizing attention, certain design choices become almost inevitable. Public metrics encourage status competition. Infinite scroll minimizes friction. Recommendation systems optimize for prediction. Virality becomes a feature rather than an accident because every additional interaction is treated as evidence that the system is succeeding.
If the objective changes, however, the architecture begins to reorganize itself.
Suppose relationships become the primary outcome. Suddenly, many of the assumptions that shaped the modern social internet become less obvious. Public follower counts no longer seem essential because relationships do not become stronger simply because they are publicly visible. Endless feeds become less attractive because relationships are deepened through conversation rather than continuous consumption. Friction, so often treated as the enemy of good design, begins to acquire a different meaning. Not every interaction should be effortless if the goal is to encourage thoughtfulness rather than impulse.
The same is true of discovery. Modern platforms have become extraordinarily good at predicting what we are likely to engage with next. Curiosity, however, does not always follow prediction. Some of the most important books we read, the most influential teachers we encounter, and the most enduring friendships we form arrive unexpectedly. They are discovered rather than predicted.
Even virality begins to look different. We have come to regard it as an unquestioned virtue because it accelerates distribution. Relationships, however, rarely become meaningful at viral speed. They develop through continuity rather than explosion, through return rather than novelty. An idea that quietly shapes a hundred minds over ten years may ultimately matter far more than one that briefly captures the attention of a million people.
Viewed from this perspective, the principles of The Small Web almost derive themselves. Human-scale communities become more valuable than massive audiences because people can actually know one another. Deliberate friction is a way to improve conversations rather than reducing engagement. Discovery through curiosity becomes more important than prediction because curiosity expands our world while prediction often reinforces the one we already inhabit. Reach becomes something earned rather than automatically granted because relationships flourish through trust, not exposure.
These principles are neither nostalgic nor anti-technology. They follow from beginning with a different question.
This is the architecture that naturally begins to emerge if our objective is relationships instead of attention
The Architecture of Relationships
The more I thought about The Small Web, the more I realized that it was never really an argument about social media. Social media merely happens to be the place where the consequences of digital architecture are most visible. Looking back across history, I began to notice that institutions as different as workshops, laboratories, seminars, salons, and The Republic of Letters were all solving remarkably similar problems. The Republic of Letters, for example, connected thinkers separated by geography but united by correspondence. Its achievement was not simply the exchange of ideas but the creation of relationships that allowed ideas to mature over decades. Like workshops, laboratories, and seminars, its real achievement lay less in the ideas it produced than in the relationships that allowed those ideas to mature over time.
The modern social internet inverted that sequence. It became extraordinarily effective at distributing information while paying comparatively little attention to the slower processes through which understanding, judgment, and community actually grow. Visibility gradually displaced familiarity, audiences displaced communities, and engagement displaced correspondence, not because these outcomes were inevitable, but because they were logical consequences of an architecture designed to maximize attention at scale.
Seen in that light, The Small Web is neither a proposal to return to an earlier internet nor an argument for abandoning technology. It is an attempt to recover a much older insight: relationships have an architecture of their own, and whenever we neglect it, we should not be surprised if the communities we build become increasingly rich in attention and increasingly poor in trust.
Every transformative technology eventually confronts the same question. After the initial excitement fades and its capabilities become familiar, attention shifts from what the technology can do to the kind of society it quietly encourages. The printing press did not merely reproduce books; it transformed the circulation of knowledge. The postal system did not merely carry letters; it enabled communities of correspondence that eventually became The Republic of Letters. The internet, in turn, collapsed the cost of publishing, communication, and discovery so dramatically that almost anyone could participate in the creation and exchange of ideas.
That achievement should not be underestimated. Never before had it been so easy for a writer to find readers, a teacher to reach students, or a community to form across continents. Yet every technological expansion creates a new scarcity. Once publishing became abundant, attention became scarce, and an extraordinary ecosystem emerged to allocate it. Recommendation systems grew increasingly sophisticated, platforms optimized engagement with remarkable efficiency, and social media became exceptionally good at distributing ideas across vast networks of people.
In solving the problem of distribution, however, we gradually forgot that distribution had never been the destination, because whether an encounter matures into a relationship depends upon an entirely different kind of architecture.
That is why The Small Web continues to occupy my thoughts. It is less an argument for a smaller internet than an attempt to imagine its next evolution. The first generation of the web solved access. The social web solved distribution. The next challenge is to design digital spaces that help relationships endure, because every lasting community is built not on moments of attention but on years of accumulated trust.
The Next Internet
Every transformative technology eventually confronts the same question. After the initial excitement fades and its capabilities become familiar, attention shifts from what the technology can do to the kind of society it quietly encourages. The printing press did not merely reproduce books; it transformed the circulation of knowledge. The postal system did not merely carry letters; it enabled communities of correspondence that eventually became The Republic of Letters. The internet, in turn, collapsed the cost of publishing, communication, and discovery, and now almost anyone could participate in the creation and exchange of ideas.
That achievement should not be underestimated. Never before had it been so easy for a writer to find readers, a teacher to reach students, or a community to form across continents. Yet every technological expansion creates a new scarcity. Once publishing became abundant, attention became scarce, and an extraordinary ecosystem emerged to allocate it. Recommendation systems grew increasingly sophisticated, platforms optimized engagement with remarkable efficiency, and social media became exceptionally good at distributing ideas across vast networks of people.
In solving the problem of distribution, however, we gradually forgot that distribution had never been the destination. Attention creates the possibility of a relationship, but relationships themselves develop according to a very different rhythm. They require continuity in place of novelty, correspondence instead of reaction, and enough shared history for trust to emerge naturally over time. Those qualities cannot simply be layered onto an architecture designed for perpetual engagement because they depend upon a different set of incentives from the very beginning.
That, I think, is why The Small Web continues to occupy my thoughts. It is less an argument for a smaller internet than an attempt to imagine the next stage in its evolution. The first generation of the web solved the problem of access. The social web solved the problem of distribution. Perhaps the challenge before us now is to design digital spaces that become places where relationships deepen, conversations accumulate, and communities develop the shared memory from which culture has always emerged.
If Kevin Kelly’s insight was that creators need only a thousand people who genuinely value their work, the question that naturally follows is no longer economic but architectural.
What kind of internet helps those thousand people find one another?
Perhaps that is how civilizations have always advanced. Their most enduring achievements are not only the technologies they invent but the ways they enable people to think together, trust one another, and build knowledge across generations. That may have been the deeper insight hiding inside Kevin Kelly’s essay all along. The next chapter of the internet will not be written by helping ideas travel farther, but by helping relationships last.