Liquid Adulthood Is Not the Disease… It Is the User Interface

Leave a Reply

Liquid modernity did not begin with the internet, but the post-LLM web is turning social fluidity into computational infrastructure. This essay argues that the fragmentation of identity, work, relationships and institutions is no longer simply a social phenomenon. Through the Asymmetric Integration Model (AIM), it proposes that these forms of human fluidity are increasingly being instrumented, optimised and integrated into computational systems, signalling the emergence of a new phase in the evolution of the web and its relationship with society.

Contents

Table of Contents

1. Introduction

Kian Bakhtiari’s recent Forbes article, What Happened to Adulthood? Welcome to Gen-Z’s Liquid Adulthood, is an excellent piece of sociology. It describes the symptoms with clarity. Identity is fluid. Careers are temporary. Housing is increasingly unattainable. Relationships are mediated through apps. Institutions command less trust. Adulthood is no longer defined by a stable sequence of work, homeownership, marriage, family, and community.

I agree with almost all of the observations.

Where I diverge is in the explanation.

These phenomena did not begin with the internet, and they cannot be reduced to technology alone. Housing policy, wage stagnation, deindustrialisation, changing gender roles, declining institutions and cultural individualisation all predate the post-LLM web. But they are now being gathered, accelerated and made mutually reinforcing by the same computational architecture.

2. From Liquid Modernity to Computational Integration

Liquid adulthood is not the underlying disease. It is the visible user interface.

Bakhtiari builds on Zygmunt Bauman’s concept of liquid modernity. In Liquid Modernity (2000), Bauman described the shift from solid to fluid forms of social organisation. Solids, he wrote, “have clear spatial dimensions and they neutralise time by resisting deformation. They are stable, predictable.” Liquids, by contrast, “cannot easily hold their shape.” They flow, spill and leak. In liquid modernity, social bonds, institutions and identities lose their fixed character. Individuals are increasingly required to construct their own biographies without the inherited frameworks that once supplied continuity and meaning.

This process began long before social media, smartphones or generative AI. The internet did not invent liquid modernity. What it did was make liquidity measurable.

Platforms turned fluid behaviour into data. Recommendation systems turned that data into optimisation. Post-LLM systems are now beginning to participate directly in the construction of identity, judgment, and choice.

That is the additional layer described by the Asymmetric Integration Model.

The Asymmetric Integration Model, or AIM, describes the transition from systems that merely attract human attention to systems that continuously ingest human behaviour, identity and judgement as inputs into machine optimisation. The integration is asymmetric because the individual becomes increasingly legible to the system, while the system’s own reasoning, incentives and interventions remain largely illegible to the individual. The machine learns from us continuously. We understand it only partially.

3. Stability Was Never Free

For much of the twentieth century, adulthood rested upon a number of relatively stable institutions. A career. A mortgage. Marriage. Community. Religion. Professional identity. These were not merely cultural artefacts. They were mechanisms for reducing uncertainty. They allowed individuals to externalise much of identity formation because society had already solved much of the problem.

A person did not need to reinvent themselves continuously. Their occupation, location, family, class, community, and social role supplied much of the answer to the question: “Who am I?”

That solidity often came at a price. It could be restrictive, hierarchical and exclusionary. It could trap people in roles they had not chosen. The erosion of those institutions therefore brought genuine freedoms. But it also transferred the burden of coherence from society to the individual. The individual now has to construct, maintain and continuously revise a life narrative under conditions of economic insecurity, social fragmentation and constant comparison.

That is where computational systems enter the picture.

4. Liquid Identity Is Computationally Useful

Bakhtiari describes identity becoming something people continually construct rather than inherit. I would go further.

Modern digital systems benefit from identities that are continuously revised. Every change creates a new behavioural signal. Every experiment produces additional preference data. Every shift in values, language, style, affiliation or aspiration becomes another optimisation input.

Identity becomes less like a passport and more like a live telemetry feed.

This does not mean platforms deliberately destabilise identity in any simple or conspiratorial sense. They do not need to. The optimisation environment already rewards novelty, experimentation, responsiveness and repeated engagement. A stable identity produces relatively predictable behaviour. A fluid identity produces variation. Variation creates data. Data improves prediction. Prediction improves intervention.

However, the important transition is not simply that platforms observe identity. It is that identity itself is becoming computationally useful.

In the early days of the web, people used technology as a tool. Search engines helped us retrieve information. Social media helped us communicate. Recommendation systems helped us discover content. The human remained the principal author of their own identity.

The post-LLM web is different.

Increasingly, systems participate in identity construction itself. They suggest what to learn, whom to follow, how to present ourselves, what opportunities appear achievable and even which versions of ourselves receive the greatest social reinforcement. AI systems are beginning to act less like libraries and more like collaborators in the continuous production of the self.

This creates a subtle but profound shift.

Identity is no longer simply something that humans express through technology.

It becomes something that technology helps to assemble.

Every interaction simultaneously serves two purposes. It assists the individual while also improving the system’s model of that individual. Every conversation, correction, preference, hesitation and decision contributes to a progressively richer computational representation of the person.

This is the asymmetry at the heart of AIM.

Humans experience these systems episodically, one conversation or recommendation at a time. The system experiences the individual longitudinally, accumulating years of behavioural signals into an increasingly sophisticated predictive model. The individual remembers selectively. The system remembers systematically.

The consequence is that identity becomes both human narrative and machine artefact. We continue to tell stories about who we are, but increasingly those stories are being interpreted, categorised, predicted and, in subtle ways, influenced by computational systems that operate continuously in the background.

This is where AIM departs from earlier theories of the digital economy.

  • Bauman explained why society became liquid.
  • Zuboff explained how that liquidity became economically valuable.
  • AIM explains how it is becoming a computational infrastructure.

5. Liquid Relationships Create More Optimisation Opportunities

The same logic applies to relationships.

Dating apps are often criticised for commodifying intimacy. That criticism is true, but incomplete. Traditional relationships eventually stabilise. Stable relationships generate fewer new choices. Fewer choices generate fewer behavioural signals.

By contrast, repeated searching, comparing, selecting, rejecting, messaging, ghosting and returning to the platform produce a rich stream of information. Every swipe is a signal. Every pause is a signal. Every match is a signal. Every rejection is a signal. Every conversation is a signal. Every return to the app is a signal.

Dating platforms do not need to deliberately prevent successful relationships. Their instrumentation simply extracts more information from prolonged searching, repeated comparison and recurring participation than from permanent exit. This is not necessarily the same as saying that dating apps cause relationship instability. Housing pressures, economic insecurity, changing attitudes towards marriage and greater personal autonomy all matter. But platforms introduce an additional incentive gradient. They make indefinite optionality easy, measurable and commercially useful. They lower the friction of comparison. They increase the visibility of alternatives. They transform relationships from bounded social encounters into continuously ranked possibility spaces.

6. Liquid Work Is Not Merely Economic

Technology does not replace “till death do us part” with “let’s see what happens” on its own. It simply makes “let’s see what happens” easier to sustain at scale.

Bakhtiari is also right to emphasise insecure employment, fragmented careers and anxiety about AI.

A job was once one of the principal anchors of adult identity. It supplied income, routine, community, status, progression and a story about the future. That model had already begun to weaken through deindustrialisation, outsourcing, deregulation, the decline of unions and the rise of flexible labour. AI accelerates the process.

The industrial economy valued repeatability. The knowledge economy increasingly values adaptability. The post-LLM economy may value continuous reconfiguration. The most valuable worker is no longer necessarily the person who performs one role exceptionally well for forty years. It is increasingly the person who can repeatedly reconstruct themselves as tools, markets and organisational requirements change.

That has obvious economic advantages. It can also become psychologically corrosive. A person who must constantly reskill, rebrand, update, reposition and remain visible is never entirely allowed to settle. Professional identity becomes provisional. Work becomes another optimisation environment.

Platforms such as LinkedIn, freelance marketplaces, learning systems and recruitment engines make professional identity machine-readable. They reward activity, recency, responsiveness, visibility and continuous self-presentation. The worker does not simply do the work. They must continuously signal their suitability for future work. The labour market increasingly rewards liquid identity because liquid identity is easier to redirect.

7. AIM Is Not a Theory of Everything

Housing unaffordability was not created by recommendation algorithms. Neither were wage stagnation, deindustrialisation, delayed family formation, reduced union membership or the weakening of organised religion. Many components of liquid adulthood have been developing for half a century or more.

Any attempt to explain all of them through technology alone would be crude technological determinism.

AIM makes a narrower claim. Once these conditions exist, computational systems render them observable, segmentable and optimisable. They convert social fluidity into machine-readable behavioural variation. They create incentives that reward continued adaptation over settlement, recurring choice over finality and measurable activity over private stability.

Technology is therefore neither the original cause nor a neutral observer. It is an accelerant, an organiser, and, increasingly, an active participant.

Bauman described the liquefaction of institutions. Shoshana Zuboff, in The Age of Surveillance Capitalism (published by Penguin), described the extraction of behavioural surplus, the raw material harvested from human experience and used to improve prediction products. AIM describes the next stage, in which human behaviour is not merely harvested for prediction but integrated into adaptive intelligence systems that increasingly influence the environments in which future behaviour occurs.

The sequence is important. First, society becomes liquid. Then, platforms instrument the liquidity. Then, machine systems begin to participate in its direction. Bauman described the liquefaction of institutions. Shoshana Zuboff described the extraction of behavioural surplus. In my earlier work on Societal Evolution, I argued that societies evolve by repeatedly externalising human capabilities into technology and institutions. AIM extends that argument by proposing that AI represents a new phase in which cognition and identity themselves become partially externalised into computational systems.

8. The Web Became Liquid First

This connects directly with another theme I have been exploring in The Web Unbundled. The visible web is losing persistence. Blogs disappear from discovery. Search increasingly returns summaries rather than sources. Public conversation fragments into private messaging, closed communities and platform-specific feeds. Links are replaced by screenshots. Archives become inaccessible. Ownership gives way to rental. Open publication gives way to algorithmic distribution.

Knowledge becomes flow rather than accumulation. The web itself has become more liquid. Pages matter less than feeds. Destinations matter less than interfaces. Authorship matters less than synthesis. Discovery matters less than recommendation.

The web was once a collection of places. It is becoming a stream of computationally assembled moments. Society is now following the same pattern. Or, perhaps more accurately, the web and society are liquefying together. Each makes the other more fluid.

9. Identity Infrastructure Will Make the Liquid Web More Rigid

Here is where the next stage becomes especially interesting.

In my work on the age-gated internet, I have argued that governments and platforms are embedding identity into the web’s infrastructure. Age verification is only the most visible current use case. Once identity infrastructure exists, it can support far more than age: jurisdiction, reputation, entitlement, employment status, professional qualification, financial standing, risk classification, political eligibility, and social trust.

At first glance, this appears contradictory. Society becomes more fluid. Infrastructure becomes more rigid. But these trends are complementary.

The more uncertain human behaviour becomes, the more valuable a stable identity becomes to computational systems. The system requires a persistent subject precisely because everything else about that subject is changing. Fluid behaviour is most useful when it can be attached to a stable identity. Without persistence, there is only noise. With identity, there is a longitudinal model.

This is the paradox at the centre of the next web. Individuals become behaviourally flexible. Infrastructure becomes increasingly deterministic. The person is encouraged to experiment. The system remembers everything.

10. Liquid Adulthood Is Becoming Computational Adulthood

Perhaps the most important sentence in the Forbes article is this:

“The burden of building a coherent life and identity falls entirely on the self.”

I suspect even that now understates what is happening.

The burden is no longer falling solely upon the individual. It is increasingly being shared with machine systems: recommendation engines, AI companions, algorithmic feeds, digital identity systems, behavioural prediction models, large language models, automated advisers.

These systems increasingly participate in the construction of identity itself. They suggest what to watch, what to buy, what to learn, whom to meet, how to speak, how to present oneself and what future possibilities appear available. They help people narrate their experiences. They mediate emotional reflection. They support decisions about work, education, relationships and health. They become external cognitive structures through which the self is interpreted.

This is not necessarily dehumanising. AI may help individuals navigate complexity better than many of the institutions it replaces. It may expand access to knowledge, advice and reflection. It may help people escape inherited roles that were restrictive, unjust or simply wrong for them.

The danger lies not in mediation itself. The danger lies in asymmetry. The system knows more about the individual than the individual knows about the system. The system can model the person. The person cannot easily inspect the model. The system can shape the environment. The person often cannot see the intervention. The system can remember. The person cannot easily forget. The system can optimise. The person may not know what objective is being optimised.

This is the transition from human-authored adulthood to computationally mediated adulthood.

11. Human Beings Still Seek Solid Ground

None of this means that people have ceased to value stability. They have not.

Platform fatigue, renewed interest in physical communities, resistance to identity verification, demands for AI transparency and the search for more durable forms of work and relationships are not anomalies. They are counterforces.

Humans require enough continuity to build trust, responsibility and meaning. A society of permanent optionality cannot sustain commitment. A labour market of permanent reinvention cannot sustain mastery. A relationship culture of endless comparison cannot sustain intimacy. An information environment of constant flow cannot sustain memory.

Systems that maximise fluidity may eventually undermine the people and institutions upon which they depend.

There are commercial counterforces too. Advertisers want stable audiences. Governments want predictable citizens. Employers want reliable workers. Families require continuity. Communities require memory. Individuals eventually tire of being treated as unfinished experiments.

The future will therefore not be entirely liquid or entirely computational. It will be contested. The real struggle will concern where solidity is restored, who controls it and whose interests it serves.

12. The Future May Become Less Liquid, Not More

Liquid adulthood may be a transitional phase rather than a permanent destination. Periods of high uncertainty rarely persist indefinitely. Eventually, societies build new institutions.

The question is what those institutions will look like.

They are unlikely to resemble the institutions of the twentieth century. They may instead be built around identity infrastructure rather than anonymous participation, persistent digital reputation rather than local social memory, AI-mediated guidance rather than independent judgement, algorithmic governance rather than institutional discretion, continuous verification rather than assumed trust, and human-machine integration rather than human autonomy.

That would not be liquid modernity. It would be a form of computational solidity. The individual would appear flexible on the surface while being increasingly fixed beneath it by persistent records, identity layers, behavioural histories and machine-generated predictions. A person may be free to reinvent themselves. The system may still remember every previous version.

13. Conclusion

Bakhtiari describes the condition accurately. Adulthood has become less institutionally prescribed, less economically secure and more dependent upon continuous acts of self-construction. That process did not begin with the internet. Bauman saw it before social media, smartphones and generative AI. Housing policy, wage stagnation, deindustrialisation, individualisation and changing family structures remain essential parts of the explanation.

What has changed is the architecture surrounding that instability.

Digital platforms made liquid behaviour measurable. Recommendation systems made it optimisable. Post-LLM systems are beginning to participate directly in the construction of identity, judgement, relationships and life choices.

This is the additional claim made by the Asymmetric Integration Model. The individual becomes increasingly legible to the machine. The machine becomes increasingly involved in the individual. Yet the system’s purposes, interventions, and accumulated knowledge remain largely invisible and difficult to contest.

Liquid adulthood is therefore not merely the product of computational systems. It is the social condition into which those systems have arrived, the behavioural environment from which they learn and the instability through which they are becoming integrated into human life.

The future is unlikely to remain entirely liquid. Human beings require continuity, trust and belonging. New forms of solidity will emerge.

The decisive question is who will build them.

Will they be institutions through which people recover agency, commitment and collective power? Or will they be identity infrastructures, reputation systems and AI intermediaries that stabilise the individual primarily so that the individual can be more accurately predicted, managed and governed?

The question is no longer simply what happened to adulthood.

It is a question of what kind of adulthood will remain possible when the systems through which we construct our lives are simultaneously learning to model, predict, and shape us.

14. References

Related reading