The gig worker staring at a phone screen at 2 a.m., having chosen to accept one more ride, one more delivery, one more batch of groceries, is the living contradiction of capitalist freedom. This worker is nominally independent—no boss demands overtime, no time clock punches their card. Yet they work sixty hours. They work seventy. They work until the algorithm stops pinging, then they refresh the app and wait. The paradox is not a quirk of platform design. It is the logical outcome of capital pressing both levers of surplus value extraction simultaneously, reverting to methods Marx identified in the 1860s while dressing them in the smartphone interface of the 2020s.
Marx Distinguished Two Roads to Surplus Value—Capital Uses Both
In Volume I of Capital, Marx defined surplus value as the difference between the value a worker creates and the wages they receive. Capitalists extract this surplus by either lengthening the working day—absolute surplus value—or by compressing the labor time necessary to produce the worker’s own subsistence through technological intensification—relative surplus value. The first method is crude and finite: there are only twenty-four hours in a day, and workers must sleep, eat, and recover. The second method is more elegant, raising productivity so that the worker reproduces their wages in fewer hours, leaving more of the day as unpaid labor for capital.
Classical industrial capitalism leaned heavily on absolute surplus value: the twelve-hour shift, the six-day week, the outright theft of meal breaks and rest periods. The labor movement’s great achievement was to force capital to limit the working day, establishing a legal boundary beyond which exploitation had to become more efficient rather than more extensive. Relative surplus value became the dominant mode of extraction in the twentieth century—factory automation, assembly line speed-up, scientific management. The length of the working day stabilized, but its intensity increased.
The gig economy shatters this historical compromise. The platform does not employ workers, the legal fiction says, so no clock ever starts. No working day has a fixed length because no working day exists as a legal category. The worker is never on shift, but they are never off call. Capital has found a way to combine absolute and relative surplus value extraction by eliminating the boundary between work and life, then intensifying every minute that crosses that erased boundary.
Algorithmic Scheduling Extends the Working Day Without a Clock
When Marx analyzed the struggle over the working day, he documented factory inspectors tabulating hours, employers fining workers for lateness, and the slow legislative imposition of the ten-hour day. The central antagonism was visible and measurable: capital wanted the laborer at the machine for as many hours as possible, and the laborer wanted to go home. The gig economy replaces the factory whistle with the ping of a notification, but the underlying dynamic is identical.
Uber’s algorithm does not schedule a shift; it creates conditions under which the driver chooses to work longer. Surge pricing, peak hour bonuses, and consecutive trip incentives are all designed to manipulate the driver’s calculation of whether the next hour is worth staying online. A 2018 MIT study found that after accounting for vehicle depreciation, fuel, and maintenance, the median Uber driver earned $9.21 per hour—below the federal minimum wage in real terms. Yet drivers continue working because the algorithm presents each ride as a discrete decision: this ride pays $8 and takes twenty minutes, which looks acceptable in isolation. The cumulative effect of thousands of these micro-decisions is a working week that stretches to sixty, seventy, or eighty hours.
Amazon Flex deploys a different mechanism. Workers are assigned delivery blocks—typically three- to four-hour windows—but the app does not guarantee a minimum number of blocks. Drivers must compete for available slots, refreshing the interface at release times, accepting blocks at any hour of the day or night. The worker who refuses a 4 a.m. block today may find no blocks available tomorrow. The platform does not command the worker to labor; it creates a scarcity of earning opportunities that compels the worker to accept whatever time slots the algorithm offers. The working day becomes whatever the app determines the market will bear.
DoorDash takes the logic further. The company’s peak pay system offers higher base rates during high-demand periods, but the algorithm adjusts these rates dynamically based on real-time supply of drivers. If too many dashers log on, the peak pay disappears. If drivers log off, the algorithm raises the incentive. The worker is locked into a constant negotiation with an opaque pricing system that rewards availability but never guarantees return. The result is that UK gig workers surveyed in 2021 reported averaging fifty-five hours per week, with a substantial minority exceeding seventy hours. These are not outliers; they are the systemic product of a system that extracts absolute surplus value by eliminating the concept of a standard working day entirely.
App-Mediated Speed-Up Intensifies Labor Without Hiring More Workers
If absolute surplus value in the gig economy is achieved through the endless extension of available work hours, relative surplus value is achieved through the algorithmic compression of each paid task. Marx described relative surplus value as arising from the revolution in the productive forces: when a capitalist introduces a machine that doubles output per hour, the value of labor power falls, and the portion of the day devoted to reproducing wages shrinks. The gig economy achieves this without any physical machinery beyond the smartphone itself. The app is the machine, and its productivity gains are extracted not by speeding up a factory line but by optimizing the routing, timing, and bundling of service tasks.
Consider the Amazon Flex block again. A driver assigned a three-hour block in 2018 might have delivered thirty packages. By 2022, algorithmic route optimization squeezed forty-five packages into the same nominal block. The driver does not work longer hours; they work faster, carrying more weight, navigating tighter schedules, and absorbing the fatigue of a compressed labor process. The app tracks every minute: time spent at each stop, deviation from the planned route, breaks that extend beyond the invisible threshold. Drivers who fail to maintain pace are deactivated, their access to future blocks terminated without explanation. This is speed-up as pure information control, unmediated by supervisors or stopwatches.
DoorDash and Uber Eats deploy a similar logic. The app surfaces multiple orders simultaneously, stacking deliveries that a human dispatcher would have considered impossible to complete within the promised time. Dashers report being offered three orders from three restaurants with pickups in opposite directions, all bundled into a single acceptance decision. The refusal penalty is ambiguous but real: decline too many stacked orders, and the algorithm reduces the frequency of future offers. The worker must either accept the intensified labor or accept reduced earnings. There is no third option.
The Gig Worker Absorbs All Risk While Capital Captures All Surplus
The synthesis of absolute and relative surplus value extraction in the gig economy rests on a single legal and material fact: the worker bears all the costs of production. The Uber driver owns the car, pays for insurance, covers fuel, handles maintenance, and bears the depreciation that makes the vehicle worth less with every mile. The DoorDash dasher supplies the vehicle, the phone, the data plan, and the insulated bag. The Amazon Flex driver provides the vehicle, the loading equipment, and the physical stamina to haul fifty-pound boxes up apartment stairs. Capital provides only the platform—a software interface that costs pennies per transaction to maintain—and takes a commission that typically ranges from twenty-five to forty percent of each transaction's gross value.
The 2018 MIT study that estimated $9.21 per hour in net earnings for Uber drivers excluded the cost of health insurance, Social Security contributions, and workers' compensation—all costs that a traditional employer would bear. When those costs are factored in, effective hourly earnings drop closer to $6 or $7 per hour. The driver is not merely underpaid; they are subsidizing the platform's profitability out of their own declining asset base. Every mile driven reduces the value of the car. Every hour worked accelerates the physical deterioration of the driver's body. The platform captures the immediate revenue while the worker shoulders the deferred costs.
The rhetorical defense of gig work hinges on flexibility and autonomy. The worker, it is claimed, chooses when and where to labor. They are entrepreneurs of their own time, empowered by technology to craft a schedule that fits their life. This argument has a moment of truth: the gig worker does not face a supervisor demanding overtime, and they can technically decline an offer without being fired. But the steel-man version of this case collapses on the material question of power. Flexibility means nothing when the alternative to accepting a low-paying ride is earning nothing. Autonomy is empty when the algorithm controls the flow of work opportunities and can throttle access arbitrarily. The choice between accepting exploitation and accepting poverty is not freedom; it is the coercion of the market made invisible by the friendly interface of an app.
The counter-argument's deeper error is that it treats the isolation of the gig worker as liberation from the collective discipline of the factory. In reality, the individualization of risk is a more thorough form of subordination. The factory worker knows who their boss is; they can organize, bargain, strike. The gig worker faces not a person but an algorithm—a black box that adjusts incentives, modifies pay scales, and issues deactivations without explanation or appeal. The flexibility celebrated by platform defenders is the flexibility of the sidewalk laborer in 1850, not the flexibility of the craftsperson in full control of their trade.
The image that should close this analysis is not a statistic or a slogan. Picture a concrete loading dock in the gray light before dawn, the fluorescent glow of a convenience store, a cargo van idling in the lot. The driver is sorting packages by street address, checking a phone mounted to the dashboard, calculating how many deliveries can fit into the next three hours. The phone buzzes with a new offer. The driver could decline. The driver could turn off the app and go home. The driver stays. The algorithm does not command, but the driver stays. That is not freedom. That is capital having learned, after one hundred and fifty years, that the most efficient form of control requires no foreman, no whistle, no clock at all—only a screen, a ping, and the quiet desperation of the person who must choose to work or starve.
