The Assembled Workforce

TAM-RIM.6-06 · The Reimagined, Cluster 6: The Coordination · The Approximate Mind

Nina checks her morning assignments at 6:15, sitting on the edge of her bed in a one-bedroom apartment in Austin that she chose because it was equidistant from three of her most frequent work sites. She has a mug of tea that she makes the same way every morning regardless of what the day holds, black tea, too much sugar, a habit she picked up from her grandmother in Odessa and has never wanted to correct.

Three projects. Two she joined yesterday. One she will leave by Thursday.

The first is a commercial kitchen renovation in a restaurant that is rebranding. Nina is there for the electrical work, specifically the reconfiguration of the ventilation hood circuits, which requires someone who understands both the National Electrical Code and the particular way that commercial kitchen exhaust systems interact with the electrical panel in buildings constructed before 1990. She has this knowledge because she spent four years working for a contractor who specialized in restaurant buildouts, and the knowledge lodged in her the way all genuine expertise lodges: through repetition, failure, and the accumulating instinct for what a specific situation requires before the manual confirms it.

The second project is a residential rewiring in a house where the owners are adding a home office and discovered that the existing wiring was installed by someone who had creative opinions about grounding. Nina was matched to this project because the AI coordination system identified her specialty in older residential electrical systems and her availability this week.

The third project is finishing up. A small data center cooling system that needed an electrical interface redesign. She joined it ten days ago, worked alongside a mechanical engineer named Paul whom she had never met and will probably never see again. They worked well together. Paul understood airflow in a way that complemented Nina’s understanding of load distribution, and for ten days they operated as a unit, communicating in the shorthand that skilled people develop when the work is specific enough to create its own language. Thursday she will file her final documentation through the system and Paul will become a name she might or might not recognize if it appears on a future assignment.

She is excellent at what she does. She belongs to nothing.

The Model

#

The assembled workforce is not the gig economy. The distinction matters.

The gig economy as it currently operates is platform feudalism. The worker has nominal independence: no boss, no schedule, no obligation. The worker has actual dependence: on the platform that controls demand access, pricing, reputation scores, and the algorithm that determines whether work appears on the screen at all. The platform is the intermediary, extracting rent from the connection between the worker and the work. The gig worker is a toll booth economy participant who has been told they are an entrepreneur.

The assembled workforce inverts the dependency. The AI coordination layer does not belong to a platform. It belongs to the workers collectively, or to the project, or to the cooperative, or to a structure that the workers govern. It matches capability to need. It assembles teams dynamically for specific problems. It manages the logistics of multi-skilled collaboration: scheduling, sequencing, parts procurement, permitting, documentation. When the project is complete, the team dissolves. The workers return to the pool. The AI assembles the next team for the next problem.

No general contractor extracting a margin for the coordination function. No platform taking a percentage for the matching function. No permanent organizational overhead. Each worker participates in multiple projects, assembled and reassembled as needs arise, with the AI handling the coordination that used to require either a firm or a middleman.

The model works best in domains where the work is project-based, multi-skilled, and variable. Construction. Event production. Film and media, which has operated on an assembled model for decades and has something to teach the rest of the economy about its costs. Complex maintenance and renovation. Product launches that require a strategist for three weeks, a designer for two, a copywriter for one, and an engineer for four.

In each case, the traditional model was either a firm that maintained permanent staff for variable demand, which meant paying people during the gaps, or a general contractor who assembled teams manually, which meant paying the contractor’s margin and accepting the contractor’s judgment about who was good enough.

The AI coordination layer eliminates both costs. The assembly is dynamic, optimized, and owned by the people being assembled.

What Film Already Knows

#

Hollywood has been operating on the assembled model for nearly a century. A film production assembles a team of specialists, hundreds of them, for a specific project. The cinematographer, the gaffer, the key grip, the production designer, the editor: each one joins, performs their function, and leaves when the work is done. The next project assembles a different team. Some members recur. Most do not.

The film industry has had longer than anyone to discover the costs of this model, and the costs are specific and well documented.

The first cost is precarity. The assembled worker lives between projects. The space between is unpaid, unstructured, and anxious. Nina, between her three current projects and her next set of assignments, does not know what next week holds. The AI system shows her probable matches based on upcoming demand, but probable is not certain, and the uncertainty is a low hum that never fully resolves. Film workers describe this as the permanent audition: you are always proving yourself, always one bad project or one slow season away from a gap that eats your savings.

The second cost is the absence of development. A firm invests in its employees because it expects to benefit from the investment. The firm sends the electrician to training because the firm needs a better electrician. The assembled model has no such incentive. Nobody invests in Nina’s development because nobody employs Nina long enough to capture the return. Nina invests in herself, which means she pays for her own training, on her own time, from her own savings. Development becomes another cost the worker bears alone.

The third cost, and this is the one that the economic analyses miss, is belonging. Nina has colleagues on every project and coworkers on none. She works alongside Paul for ten days and they develop a functional intimacy, the shorthand of shared problem-solving, that dissolves when the project ends. She works alongside a plumber named DeShawn on the restaurant renovation and they eat lunch together in the half-finished dining room, talking about their kids, and on Friday DeShawn moves to his next assignment and Nina moves to hers and the lunch is over.

The assembled workforce provides work. It does not provide a workplace.

The difference is not semantic. A workplace is where you are known. Where someone remembers that you don’t eat peanuts. Where the morning has a rhythm that includes people who saw you yesterday and will see you tomorrow. Where the accumulation of small interactions over time produces something that is not exactly friendship and not exactly professional relationship but is a texture of daily life that humans seem to need in ways they do not always articulate until it is gone.

Nina has her apartment and her tea and her grandmother’s habit of too much sugar. She has her skills, which are genuine and valued. She has her autonomy, which is real and which she chose and which she would not trade for a permanent position at a firm where a supervisor told her what to do. She chose this.

She also chose a life where she is perpetually arriving and perpetually leaving, where every team is temporary, where the investment of learning someone’s rhythm, their jokes, their way of handling the mid-afternoon slump, is an investment she makes knowing it will be liquidated within days or weeks.

Film people have a word for the feeling that comes at the end of a production. They call it wrap grief. The project is done. The family you built is dissolving. Everyone hugs. Everyone promises to stay in touch. Most don’t.

Nina experiences wrap grief three or four times a month. She does not call it that. She calls it Thursday.

The Skill Development Problem

#

The assembled model has a circularity that the film industry has never solved.

The work requires expertise. Expertise develops through sustained practice, mentorship, and the accumulation of failure in a context where failure is survivable and instructive. Historically, this accumulation happened inside firms. The apprentice electrician worked alongside the journeyman for years. The junior developer was reviewed by the senior developer across dozens of projects. The associate learned from the partner through proximity and repetition.

The assembled model has no apprenticeship structure. The AI matches capability to need, which means it matches existing capability. The worker who already has the skill gets the project. The worker who needs to develop the skill does not get the project, because the project needs someone who can perform now, not someone who will be able to perform in six months.

This creates a closing loop. The experienced workers get work because they have experience. The inexperienced workers cannot get experience because they cannot get work. The pool of available expertise gradually ages and does not replenish. The system consumes the expertise that the old model created and does not produce new expertise to replace it.

Nina learned her specialty inside a firm. She spent four years with a contractor who tolerated her early mistakes because the contractor expected to benefit from her growing competence. The tolerance was an investment. The assembled model does not make this investment because there is no entity in the model whose time horizon extends beyond the current project.

Some solutions exist. Apprenticeship could be structured as its own project type: a senior worker is matched with a junior worker, the AI coordinates the pairing, and the cost of the apprenticeship is borne by the cooperative or the pool or the public or somebody. But “somebody” is the problem. In a model with no permanent employer, nobody has the structural incentive to pay for development. The cost falls to the worker, to the state, or to a cooperative governance structure that has enough other problems to solve.

When AI absorbs the routine work that was also the training ground, where does expertise develop? The assembled model makes the question concrete. The training ground was not just the work. It was the firm. The firm was where you were bad at something in the presence of someone who was good at it, for long enough that the goodness transferred. Remove the firm and you remove the transfer mechanism.

What the AI Sees and Does Not See

#

The AI coordination system that assembles Nina’s teams is good at matching. It knows her certifications, her specialties, her past project ratings, her availability, her geographic range. It can predict, with reasonable accuracy, which projects will need her skills in the coming weeks. It can optimize the assembly: putting Nina with Paul on the data center project was a good match because their complementary expertise produced a better outcome than either could have achieved alone.

What the AI cannot see is the human dimension of the assembly.

It cannot see that Nina and Paul worked well together not just because of their complementary skills but because of a quality of mutual respect that emerged in the first hour and that neither could have predicted or specified. It cannot see that Nina’s work on the restaurant renovation is slightly less sharp this week because the residential rewiring project is emotionally draining, not physically but in the way that encountering someone else’s dangerous incompetence is draining when your professional conscience compels you to fix it properly rather than minimally. It cannot see that the reason Nina chose this life, the autonomy, the variety, the freedom from organizational politics, is also the reason this life is wearing her down, and that the two things are not in tension. They are the same thing.

The assembled worker’s freedom is real. The assembled worker’s exhaustion is real. They come from the same source: the absence of a structure that holds you, that constrains you, that tells you where to be tomorrow, and that in constraining you also carries some of the weight of being a person with a working life.

I wonder whether there is a structure that provides the belonging without the constraint. Whether the cooperative model from the previous essays could be adapted: a pool of workers who own the coordination layer collectively, who govern themselves, who have the autonomy of the assembled model and the continuity of a shared enterprise. A home base that is not an employer but a commons. A place where Nina’s name is known, where her tea preference is remembered, where she returns between projects to something that is hers.

The film industry has guilds. The guilds provide some of this: health insurance, pension, a directory of members, a set of professional standards. But the guilds are defensive organizations, built to protect workers from the industry’s structural precarity. They do not provide belonging. They provide a floor.

What Nina needs is not a floor. It is a ceiling she chose and a floor she trusts and walls she can leave through. A room of her own in a building she co-owns.

Whether that room can be built is a question the next essay approaches from a different direction.

Thursday

#

The data center project wraps on Thursday as scheduled. Nina files her documentation. Paul sends a message through the system: “Good working with you.” She responds in kind. The system logs the collaboration rating. They will both receive a slight boost in their match scores for complementary projects.

She drives home. The apartment is quiet. The tea is the same. She checks the system for next week. Two probables. One confirmed. A hospital generator upgrade that starts Monday, which she is looking forward to because hospital work is complex and her attention sharpens when the stakes are real.

She has her skills. She has her freedom. She has her grandmother’s tea.

She does not have a workplace. She has a series of places where she works.

The distinction sits in the apartment with her, unresolved, on a Thursday evening when the project has ended and the next one has not begun and the hours between belong to no one and to nothing and to her.

This is the sixth essay in The Coordination, a cluster within The Reimagined examining what happens to the structure of the firm when AI can perform the coordination function. The previous essays traced the one-person firm (TAM-RIM.6-01), the zero-person firm (TAM-RIM.6-02), the inverted firm (TAM-RIM.6-03), the worker-owned factory (TAM-RIM.6-04), and the direct supply chain (TAM-RIM.6-05). This essay asks what happens when the workforce itself becomes fluid, assembled and disassembled by AI for specific projects. The essay that follows (TAM-RIM.6-07) asks what happens when everyone in the chain joins a single collective. This essay connects to the belonging gap in TAM-027 and TAM-028; to the connected loneliness in TAM-060, where identity dissolves when nothing requires anyone specifically; to the new apprenticeship crisis in TAM-TRF.6-02, where the developmental path that produced expertise has been automated; to the curation economy in TAM-033, where the assembled worker’s reputation is curated by a system; and to the friction-was-load-bearing insight, applied here to the firm as a social structure rather than an economic one.

References

#

Gig Economy and Platform Labor

De Stefano, Valerio. “The Rise of the ‘Just-in-Time Workforce’: On-Demand Work, Crowdwork, and Labour Protection in the ‘Gig-Economy.’” Comparative Labor Law and Policy Journal, vol. 37, no. 3, 2016, pp. 471-504.

Ravenelle, Alexandrea J. Hustle and Gig: Struggling and Surviving in the Sharing Economy. University of California Press, 2019.

Rosenblat, Alex. Uberland: How Algorithms Are Rewriting the Rules of Work. University of California Press, 2018.

Project-Based Work and Creative Industries

Bechky, Beth A. “Gaffers, Gofers, and Grips: Role-Based Coordination in Temporary Organizations.” Organization Science, vol. 17, no. 1, 2006, pp. 3-21.

DeFillippi, Robert J., and Michael B. Arthur. “The Boundaryless Career: A Competency-Based Perspective.” Journal of Organizational Behavior, vol. 15, no. 4, 1994, pp. 307-324.

Belonging and Workplace Community

Oldenburg, Ray. The Great Good Place: Cafes, Coffee Shops, Bookstores, Bars, Hair Salons, and Other Hangouts at the Heart of a Community. Paragon House, 1989.

Sennett, Richard. Together: The Rituals, Pleasures and Politics of Cooperation. Yale University Press, 2012.

Weil, David. The Fissured Workplace: Why Work Became So Bad for So Many and What Can Be Done to Improve It. Harvard University Press, 2014.

Apprenticeship and Skill Formation

Lave, Jean, and Etienne Wenger. Situated Learning: Legitimate Peripheral Participation. Cambridge University Press, 1991.

Wenger, Etienne. Communities of Practice: Learning, Meaning, and Identity. Cambridge University Press, 1998.