The claim: The "job" — trading your time to a large organization in exchange for a salary — is roughly 300 years old. It is a historical anomaly, not the natural state of human economic life. AI is pulling it apart from both ends at once: it lets organizations run with fewer people (the push, which is automation displacement), and it lets one person work at a scale that used to require a company (the pull, which is the augmentation thesis). We are not entering a post-work future. We are entering a post-job future — work returning to its old default of self-directed agency, augmented by AI instead of organized by corporations.
The core of it
- The job is a recent invention. For nearly all of human history people were artisans, farmers, merchants and makers who directed their own labor. The corporation as an employer is about 300 years old. That makes it an experiment, not a law of nature.
- This isn't nostalgia. It is a structural observation. The same technology that makes a large organization ask whether it still needs as many employees is the technology that lets an individual operate at a scale that previously required an organization.
- It applies everywhere, not just to knowledge work. A carpenter in Guatemala using AI-assisted design to sell straight to customers. A nurse in Nigeria building a telemedicine practice (treating patients remotely). A farmer in India improving yield and selling without a middleman. A mechanic in Detroit diagnosing with AI and running his own shop instead of working at a dealership. The plumber doesn't need a plumbing company. The teacher doesn't need a school district. The accountant doesn't need a firm.
- Because the reasons companies existed are now supplied elsewhere. Coordination, access to information, scale, and credentialing (proof that you are qualified) are increasingly provided by AI at the level of one person. The organizations aren't bad. Their reason for being is being handled by something cheaper.
- The middle of the organization goes first. Coordination, administration and reporting — the layer between the people doing the work and the people funding it — is exactly what AI replaces earliest. That layer is largely white-collar work.
- Whether this reads as freedom or as wreckage is the open fight. The hopeful version is techno-optimism. The version where independence doesn't pay enough to live on is post-labor economics. And there is an irony worth naming: organizations built the AI that makes organizations less necessary — the tools of their own demise.
Four eras of work
| Era | What work looked like |
|---|---|
| Pre-industrial (until ~1750) | Work was personal, local and self-directed. Guilds, apprenticeships, family enterprises, farming, trade. You made things, grew things, or provided services directly. There was no employer in the modern sense — you had patrons, customers, or the land. |
| Industrial (~1750–1970) | The factory created the job: standardized labor traded for wages. Corporations emerged as a way to coordinate complex production. The deal was give up autonomy, get stability — salary, benefits, pension. Middle management was invented as the coordination layer between capital and labor. |
| Corporate (~1970–2020) | Knowledge work extended the factory model to information. Same deal: time and autonomy traded for salary and benefits. The organization became identity — "I work at a big firm" rather than "I am a consultant." Health insurance, retirement and career ladders were all tied to the employer. |
| AI (~2020–?) | AI collapses the coordination cost that justified organizations. One person plus AI does what used to need a team. The middle of the organization is the first thing to go. And the deal breaks: organizations offer less stability while asking for more flexibility. |
Not the gig economy
The gig economy — Uber, TaskRabbit — is agency without leverage. You are still serving a platform that captures the value. This is a different thing:
- Gig economy. The platform owns the customer relationship, sets the price and controls the algorithm. You are a contractor, not an agent.
- Postmodern agency. You own the relationship, set the terms, and use AI as your operating infrastructure. The platform is a tool, not a boss.
The difference is who holds the leverage. A solo consultant with Claude, a website and real expertise is closer to a 1700s master craftsman than to an Uber driver.
Who argues it
- David Shapiro — bilateral dependence (each side needing the other) is breaking. States and corporations no longer need people the way they used to, so the individual has to build new leverage.
- Marc Andreessen — the techno-optimist case: individual builders are the engine of progress, and institutions get in the way.
- Peter Thiel — zero-to-one thinking. The unit of innovation is the lone founder or builder, not the organization.
- Salim Ismail — exponential organizations: small teams using technology to operate at 10x scale (ten times what their headcount alone would allow).
- Emad Mostaque — sovereign AI for individuals and small nations. AI as the thing that evens out the advantage of organizational scale.
- Peter Diamandis — the abundance case. Technology lets individuals do what only governments and corporations could do before.
- Nikhyl Singhal — practitioner evidence from inside a big tech company that AI amplifies individual product work.
- Jim Rutt — Game B sovereignty. An individual who can make sense of things and make choices without leaning on an institution is the base unit of a different kind of civilization. This supplies the systems framing for what self-directed agency scales into.
- Jordan Hall — Game B co-originator, and the deepest version of sovereignty as the base unit: the individual capacity to perceive, make sense and choose without institutional authority. Pairs that primitive with collective intelligence.
Who pushes back
- Daron Acemoglu — institutions matter. Historically, individual agency without institutional support produced exploitation, not freedom.
- Yanis Varoufakis — techno-feudalism (platforms as the new landlords). The agency is an illusion if the platforms capture the surplus.
- Kate Crawford — the supply chain behind "individual" AI use still runs on extractive labor somewhere else.
- Shoshana Zuboff — surveillance capitalism (your activity resold as data). Your agency generates value that someone else collects.
- Carl Benedikt Frey — the historical transition from artisan to factory was violent and took generations. The trip back may be too.
- Pope Leo XIV — subsidiarity (decide things at the most local level that works) supports local agency, but he warns against atomization with no community around it.
Adjacent
- Tyler Cowen — "average is over." The highly skilled thrive as independents, but the median worker may not have enough leverage for real agency.
- Erik Brynjolfsson — racing with the machine requires complementary skills. Not everyone can be an independent agent.
The evidence
What follows is what has actually shown up between 2024 and 2026.
- The one-person business. A growing number of solo founders and tiny teams are building $1M+ businesses with AI — indie hackers shipping products with Claude or GPT, Pieter Levels among the named examples. It isn't only tech: Etsy sellers using AI for product design, freelance translators using AI to 10x their throughput (get ten times as much work done), independent tutors scaling with AI-assisted curriculum.
- The coordination layer is being shed. The "Turbulent Transition" framework describes firms paying twice during the changeover — once for the legacy organization, once for the AI-native edge. The firms that survive drop the coordination layer. The people in that layer either join a smaller firm or become agents.
- The entry-level work crisis. Stanford data showing a 16% employment decline for 22-25 year olds in AI-exposed occupations (jobs AI can already do parts of). If the first rung disappears, the path into organizations narrows — which pushes more people toward independent work by necessity rather than by choice.
- Location stops being the gate. A developer in Lagos has access to the same Claude or GPT as a developer in San Francisco. For the first time the tools of agency don't require being near the capital centers. Mobile-first AI access in developing economies could leapfrog the corporate employment phase entirely, the way mobile banking leapfrogged branch banking in Kenya.
- Health insurance is the lock, in the US. The single biggest structural barrier here is health insurance tied to the employer. Countries with universal healthcare have already removed that lock, which makes independent work less risky. The US may be the last major economy where the job persists mainly because of the benefits rather than because of the work itself.
- The infrastructure is being built either way. The PIIE/Anthropic/UVA paper documents $250B in AI GDP growing at 2,600% quality-adjusted (counting capability gained, not only dollars spent). That infrastructure doesn't only serve corporations. It serves anyone with access. Whether cheaper AI labor ends up creating more demand for human agency or less is the Jevons paradox question.
Open questions
- Is this available to everyone, or only to the already-skilled? Cowen's "average is over" critique applies: this may be a privilege of the knowledge class. The master craftsman had years of guild training behind him. AI compresses how long it takes to get good — but does it compress it enough? Can a first-generation college student in rural Brazil reach agency as readily as a senior professional in a major city?
- What replaces everything else a company provides? Companies supply more than paychecks — identity, community, mentorship, structure, healthcare (in the US), retirement. Loneliness, no mentor and loss of identity are real costs of leaving.
- What happens to bargaining as a group? Unions emerged to give workers leverage against organizations. If there are no organizations, who do you bargain with? The platforms? The AI providers? Could AI-era guilds appear — networks of independents sharing standards, reputation, training and collective negotiation?
- Is this a choice or a push? If organizations are shedding workers faster than people are choosing independence, this is not a return to agency. It is displacement dressed up as freedom. A craftsman who chose independence and a factory worker who was laid off may end up doing the same work, and the difference between them is enormous.
- Who builds the infrastructure, and who controls it? This requires AI access, internet, banking, legal frameworks for independent work, and portable benefits (insurance that follows the person, not the job). In countries where those exist, agency is possible. Where they don't, the thesis breaks.
- Where does judgment come from? If organizations no longer train people, and guilds don't exist yet, how does the next generation develop the judgment that agency requires? AI can execute. Someone still has to know what is worth executing, and the apprenticeship that taught that used to live inside organizations.
Rich's take
Rich is testing this thesis in real time through his own semi-retirement plan and through Agency Lab. But the theory stands independent of any one person's application of it. It is a claim about a structural shift in how humans organize work — not about any particular discipline or skill set.
All theories · clickbridge.com · Rich Price · last reviewed September 2026 · People linked here have their own pages at people.clickbridge.com