---
title: "A Platform for the Portfolio Economy: Creating a Talent Commons"
authors:
  - "Kevin Frazier"
essay: "08"
part: "Part II: The Intelligent State"
source: "https://futurecraft.thefai.org/essays/a-platform-for-the-portfolio-economy/"
audio: "https://futurecraft.thefai.org/audio/a-platform-for-the-portfolio-economy.mp3"
narration: "AI voice (ElevenLabs “Statesman”)"
publisher: "Foundation for American Innovation"
year: 2026
rights: "© 2026 by the Foundation for American Innovation. All rights reserved."
---

# A Platform for the Portfolio Economy: Creating a Talent Commons

Kevin Frazier

Essay 08 · Part II: The Intelligent State

![An applicant is interviewed at the Texas State Employment Service in Corpus Christi, 1940. Matching workers with jobs required institutions that could make their abilities legible to employers.](https://futurecraft.thefai.org/img/essay-08-wide-1200.jpg)

An applicant is interviewed at the Texas State Employment Service in Corpus Christi, 1940. Matching workers with jobs required institutions that could make their abilities legible to employers. Photograph by Russell Lee / Library of Congress.

## Summary

The Fair Labor Standards Act, signed in 1938, sorted workers into rigid categories and built wage, hour, and benefit protections around them. Those categories were hard to administer then, as even the officials charged with enforcing them admitted. Transformative AI now threatens to render them unworkable. As capabilities and consumer tastes shift, employment will grow more fluid: shorter engagements, varied forms of compensation, and workers who log hours for several employers at once. Labor and employment law—specifically, the FLSA—was not built for that economy. This essay proposes a Talent Commons, a federally chartered platform through which workers maintain portable skill profiles, complete recurring AI-enabled assessments, identify work they are qualified to perform, and route pay and benefits across employers into accounts they control. Done well, it would help more people turn proven skill into work, bargaining power, and economic security.

Transformative AI will change how workers qualify for job opportunities as well as how they are compensated. Museums will include exhibits with punchcards, W-2s, and newspaper clippings discussing the 9-to-5 grind. The centerpiece of the exhibit will be the original copy of President Franklin Delano Roosevelt’s speech to Congress in which he called for the Fair Labor Standards Act (FLSA), which he later signed in 1938.[^1] President Roosevelt urged, “The time has arrived for us to take further action to extend the frontiers of social progress.”[^2]

The FLSA established labor protections based on rigid worker classifications in response to widespread exploitation. Within months, lawmakers realized that administering the FLSA—specifically, determining who qualified as an employee and therefore received additional protections—would be harder than anticipated. Officials today are still attempting to draw that line.[^3] This futile effort comes at great expense to workers and employers, who commonly find themselves litigating arbitrary definitions.[^4]

Muddled worker classifications and compensation schemes that were out of date in 1938 will not survive in the transformative AI economy. Rapid shifts in AI capabilities will force firms and workers to quickly and recurrently adjust; shorter, flexible arrangements with varying types of compensation will become far more common.

The incoming portfolio economy—one characterized by fluid employer-worker arrangements, changes in the demand and supply of various skills, and, by extension, an ongoing need for workers to develop and improve a varied set of capabilities—requires a new legal infrastructure: one that allows workers to prove their skills, present such proof to potential employers, and collect pay and benefits across whatever arrangements follow. Three principles ought to guide Congress in establishing this overhauled approach.

The first principle is portability. Pay, overtime-equivalent premiums, capability credentials, and whatever benefits are offered in the future must be transferable across all jobs a worker completes. Imagine a month in which someone works 120 hours for employer A, 30 hours for employer B, and three hours each for employers C, D, and E. While this may have seemed radical to FDR, it is a foreseeable schedule for a worker in 2027. Calculating the benefits and safeguards for such a schedule would be a nightmare. Whatever legislative framework follows the FLSA must make it easier for all kinds of work—gig, part-time, seasonal, etc.—to bolster the worker’s economic security.

The second principle is optionality. The FLSA and related labor and employment laws act as though workers have monolithic views around the maximum hours they would like to work, when, and subject to which minimum financial compensation package. If, as predicted by Alex Imas, the jobs of the future are increasingly relational—therapists, excursion leaders, care providers, etc.—then people may find themselves working odd schedules and seeking compensation in diverse forms.[^5] Workers should be able to bargain over a menu of lawful compensation options. A shared platform would make those terms easier to express, compare, and carry across engagements.

The third principle is clarity. Whatever agreement is reached by the worker and employer ought to be specified in plain language and upon the opportunity for meaningful review by the worker or an AI agent of their choice, such as one offered by a worker’s collective. The same should be true of capability assessments: workers should understand what a test measures, what roles it unlocks, and how long the credential lasts.

Applied together, these principles point toward a labor law framework less concerned with preserving old classifications and more concerned with worker autonomy.

Federal legislators keen to apply these principles should start from three premises about the portfolio economy. First, the FLSA’s classification architecture will strain as workers move among tasks of varying sophistication and oversight. Second, portable benefits must become table stakes as more people work more jobs. And, third, regardless of what Congress does, new ways of assessing skill and knowledge will develop, such as VR simulations and AI-generated work trials.

Congress should anticipate those developments and build around them. A labor market based on recurring assessments of a worker’s capabilities could open more work to more people than one organized around degrees and licenses. Yet the assessment market could easily fragment. Every employer could build its own test. Every trade association could defend its own credential. Workers would face a familiar burden with new technology: prove yourself here, then prove yourself again somewhere else, then hope the job, pay, and benefits follow.

A Talent Commons would supply the missing infrastructure for the portfolio economy. Congress should create a federally chartered platform through which workers can maintain portable skill profiles, complete skill assessments, identify work, and route pay and benefits into a single account. This Commons would differ from existing job and career platforms. Existing platforms help workers search for work, describe their experience, or acquire training. The Commons would perform a different function: it would create a trusted recognition layer for the labor market. It would define what counts as a valid assessment, what must be disclosed about that assessment, how credentials travel across employers and states, how often they must be refreshed, and how workers can challenge errors. Its value would come from making capability signals interoperable, benefits portable, and labor matching less dependent on stale proxies.

Consider a hypothetical: a person with emerging plumbing skills—albeit no formal license or degree—looking to earn work. Under the Commons, this person may not need to become a master plumber to qualify for every available plumbing-related task. An industry group could develop an AI-enabled simulation testing leak diagnosis. Passing that assessment would not erase licensing rules where public health and safety require them. It would allow participating employers to identify workers who can perform discrete tasks for which such additional training is unnecessary. Our hypothetical novice plumber would then use that assessment to find more work across the whole market.

Over time, the Talent Commons would lead to consolidation around best practices. Employers would gravitate toward assessments that reliably identify workers who can perform defined tasks. Bad assessments would wither from minimal use.

If successful, labor matching would greatly improve and further encourage workers to seek out and earn new credentials. The portfolio economy will create more discrete forms of work: supervising AI agents, performing embodied tasks, providing judgment, and carrying out narrow technical functions with AI assistance. The Talent Commons would help firms find people qualified for those tasks without pretending every task must map onto an inherited occupation.

Portable benefits are the other half of the system. Once a worker’s capability profile helps unlock an engagement, the same platform should help route the required benefit contribution into an account the worker controls. Health coverage, retirement contributions, training credits, and other benefits should accumulate across qualifying engagements.

The Department of Labor should help build the worker-side tools needed to make the Commons usable. It could launch a contest for unions, workers’ collectives, community colleges, professional associations, trade schools, or related nonprofits to develop AI agents that represent workers during contract formation and capability development. These agents should help workers understand which assessments unlock which opportunities, which providers offer credible pathways, how a given engagement affects pay and benefits, and what steps would make the worker eligible for more valuable work.

The Talent Commons would not repeal the FLSA’s ambitions so much as redeem them, giving workers a way to carry proven skill, fair pay, and earned benefits across whatever arrangements the AI economy invents next. Roosevelt asked his generation to extend the frontiers of social progress; ours can answer by building the infrastructure that lets all capable workers prove what they can do and profit from it.

[^1]: Fair Labor Standards Act of 1938, Pub. L. No. 75-718, 52 Stat. 1060 (1938).

[^2]: Franklin D. Roosevelt, “[A Fair Day’s Pay for a Fair Day’s Work](https://publicpolicy.pepperdine.edu/academics/research/faculty-research/new-deal/roosevelt-speeches/fr052437.htm)” (speech to Congress, May 24, 1937), Pepperdine University School of Public Policy.

[^3]: William G. Whittaker, [*The Fair Labor Standards Act: A Historical Sketch of the Overtime Pay Requirements of Section 13(a)(1)*](https://www.everycrsreport.com/reports/RL32088.html), RL32088 (Congressional Research Service, 2005).

[^4]: Daniel V. Dorris, “[Fair Labor Standards Act Preemption of State Wage-and-Hour Law Claims](https://lawreview.uchicago.edu/sites/default/files/76_3_Dorris.pdf),” *University of Chicago Law Review* 76, no. 3 (Summer 2009): 965.

[^5]: Alex Imas, “[What Will Be Scarce?](https://aleximas.substack.com/p/what-will-be-scarce)” *Ghosts of Electricity*, April 14, 2026.

## About the author

Kevin Frazier directs the AI Innovation and Law Program at the University of Texas School of Law and serves as a senior editor at *Lawfare*. He also co-hosts the *Scaling Laws* podcast on AI policy.

---

From *Futurecraft: The Day After AGI*, Foundation for American Innovation, 2026. <https://futurecraft.thefai.org/essays/a-platform-for-the-portfolio-economy/>

© 2026 by the Foundation for American Innovation. All rights reserved.
