The incumbent is overmeasured.
He has an address, a title, a job description, a deed, a credential, a professional association, a neighborhood group, a before-and-after story. When growth threatens something around him, the loss arrives in high resolution: the blocked view, the changed block, the automated task, the cheaper competitor, the weakened status relation. The loss is local, narratable, and represented.
The beneficiary of growth often has no address yet. She has not moved into the apartment. Her child has not grown into the person the city made possible. The customer has not bought the cheaper service. The gardener has not become the operator of a robot-assisted landscape business. The new market has not been named. The gain exists as a possibility surface before it exists as a constituency.
Scarcity politics counts the incumbent because the incumbent is visible. Abundance politics has to count the path before the person walking it can appear.
The wrong unit is the person already holding the scarce position.
A familiar San Francisco housing argument treats the homeowner as the customer. The owner has the house, the view, the block association, the legal standing, and the time to show up. The proposed building becomes a subtraction from him. His view is blocked. His street changes. His neighborhood becomes less like the thing he bought.
The loss is real. The question is whether that loss is the unit the city should optimize around.
A city does not have a fixed stock of views. Growth does not merely obstruct scenery. It creates more positions from which the world can be seen, used, loved, and improved. The view from Pacific Heights exists because earlier layers of growth created streets, sewers, electricity, water systems, schools, hospitals, businesses, parks, law, and public order. The same urban expansion made Berkeley, Oakland, Tiburon, Sausalito, and the rest of the Bay legible as places rather than distant terrain. Development multiplied view-bearing positions.
The homeowner whose asset appreciated through urban density holds crystallized public growth. He is not wrong to value it. Accumulated value is real; the question is whether defending it at the margin is worth the path-production it forbids. He is wrong if he treats that crystallized gain as a veto against the next person's access to the process that created it.
The Victorian apartment owner makes the point cleaner than the sentimental single-house homeowner. Many owners of old flats, subdivided houses, and small apartment buildings already understand San Francisco's value. They like the city. They are not confused about density, scarcity, rent, or location. They may even prefer to live in Sausalito while the San Francisco property keeps producing income. The problem is not that they fail to see urban abundance. The problem is that the existing permission regime lets them capture a prior layer of abundance while treating the next layer as a threat.
Michael Saylor's Manhattan analogy for Bitcoin is useful here in reverse. His point is that scarce property in an economic capital can be worth buying and holding across cycles because the surrounding economy keeps thickening around it. Some San Francisco incumbents behave as if they already own that kind of asset. They do not need to sell because the city is valuable. They do not need more growth because the existing scarcity helps monetize what they already have. The incumbent is not always anti-city. He may be very pro-city, as long as the city's next unit of value accrues to his deed.
This is the test for "neighborhood character." Some neighborhood character is real and worth protecting: street life, architectural texture, safety, human scale, local memory. Some of it is the incumbent's preference for a scarce arrangement whose scarcity raised his asset price. The distinction is whether the claim protects the conditions under which more people can live good lives there, or whether it protects the incumbent's possession of a position made valuable by earlier public permission.
If San Francisco had frozen itself at any earlier layer, many of the people now invoking preservation would not have the thing they are preserving. The preservationist's position is often an artifact of prior non-preservation.
The homeowner is a stakeholder. He is not the boss.
The person trying to get off the street is closer to the boss. The mother trying to stabilize her life, study, earn, date, raise a child, and give that child enough room to become someone unpredictable is closer to the boss. Her child may become ordinary or extraordinary; the city cannot know. The city does know that paths require footholds, and housing is one of the first footholds.
This does not require pretending all preferences are equal. They are not. The marginal value of preserving an incumbent's current view is not the same as the marginal value of giving a family a path into stability. The marginal value of protecting a high-asset owner's neighborhood stasis is not the same as the marginal value of letting a worker live near opportunity.
Not "from each according to ability." Something softer, more practical, and harder to evade: to each path according to the margin it opens. The person with options can share margin. The person without options needs the margin to become a path.
The customer is all of us, but the next unit of public value usually does not come from polishing the incumbent's already-owned option. It comes from opening the path that is currently closed.
AI displacement anxiety repeats the same accounting error.
The sympathetic sentence writes itself: protect the truck driver, the gardener, the bookkeeper, the paralegal, the radiologist from having the job taken. The worker is visible. The job has a name. The task has a wage. The harm can be narrated.
The created work is harder to see. It may not use the old job title. It may not sit inside the old employer. It may require access to tools, customers, insurance, training, and platform position. It may become a service that was previously too expensive to exist.
A gardener is the clean case because the task layer and the purpose layer are visible. The task layer is mowing, edging, trimming, hauling, irrigation checks, planting, diagnosis, quoting, and scheduling. The purpose is maintaining and improving a living exterior space for someone who wants beauty, food, shade, order, status, memory, or pleasure. Robots can attack the task layer without exhausting the purpose.
If robot mowers, planting machines, cheap design software, sensor diagnostics, and scheduling agents make high-quality landscaping one-tenth or one-hundredth the price, the result is not necessarily fewer gardeners. It can be more gardened world. A household that never paid for landscaping may pay for seasonal plantings. A normal suburban facade can become a designed object. A renter can maintain a balcony garden. A block can afford coordinated street trees. Tiny fountains, odd flower palettes, bonsai, edible walls, and custom yards become service categories instead of rich-person whims.
The worker does not disappear in that world. The worker moves toward specification, supervision, taste, exception-handling, trust, and customer relationship. The number of humans doing the work can rise if the lower price expands demand faster than automation removes task-hours.
That is Jevons in service form. Efficiency lowers the effective price. If demand is elastic, total use can rise. The scarce activity stops being scarce enough to remain a luxury, and a larger market appears around it.
Jensen Huang's radiology example is the same distinction at professional scale. AI became strong at image interpretation, but Huang's point is that the purpose of radiology was never "look at pixels." It was diagnosing disease and helping patients and clinicians decide what should happen. Faster scan interpretation can increase scan volume, shorten bottlenecks, and increase demand for radiologists because the purpose remains larger than the automated task.
The example will not generalize to every profession. Some demand is inelastic. Some tasks are close to the whole job. Some institutions will use automation to reduce headcount rather than expand service. But the first cut is still correct. A job title is often a bundle of tasks wrapped around a purpose. Automation prices the tasks down. The abundance question is whether the worker can climb toward the purpose.
The answer depends on access. Without access, the abundance story collapses into platform capture. The robots are owned elsewhere. The customers route through a marketplace that takes the margin. The worker becomes a thin contractor handling exceptions at lower status. The homeowner gets cheaper landscaping; the gardener gets less life.
With access, the worker can become the operator of the cheaper capability. He can supervise machines, sell design, carry local trust, and serve customers who could not afford the old labor stack. The policy question is not how to protect the old job from the tool. It is how to give the worker enough claim on the tool-mediated market that cheaper capability expands his agency rather than routes around him.
Compensation begins after the path closes. Access asks who gets to walk the new path.
The incumbent frame wins because it has better evidence.
The homeowner can photograph the view. The worker can name the task. The neighborhood can point to the parcel. The union can point to the contract. The professional association can point to the credential. The harm is local, narratable, and represented.
The apartment resident who would have existed cannot attend the hearing. The child whose life would have bent differently has no standing. The customer who would have bought the cheaper service does not know she wants it yet. The gardener who would have built the robot-assisted business has not seen the tool stack. The market that would have formed has no trade association.
The future is structurally underrepresented because it is not yet addressable.
A planning process, labor policy, or AI-governance frame that simply asks "who is visibly harmed?" will overweight the people already close enough to the scarce good to have a harm. The better question is: what new positions become possible if the system grows, and what conditions determine whether those positions are broadly accessible or privately captured?
The incumbent's loss is evidence. It is not sovereignty.
Growth is not magic. It can fail every test this piece depends on.
Housing growth without infrastructure can create crowding instead of city. Density without safety, transit, public space, utility capacity, and beauty can degrade the life it claims to expand. In that case, the homeowner's objection may be misframed but still tracking a real system failure.
AI automation without access can create cheaper services while degrading workers. If the tools, customers, reputation systems, and capital are owned by the platform, the worker does not become an operator. He becomes residue. In that case, displacement is not the wrong concern; it is the result of an access failure.
Demand may be inelastic. Some services do not expand much when price falls. If the total market does not grow, automation really can mean fewer humans needed. The Jevons shape is conditional, not a law that rescues every profession.
The future entrant can be used as cover. Developers, labs, and platforms can invoke the excluded future customer while capturing the gains for themselves. That is the oldest trick in growth politics: speak for the diffuse beneficiary, route proceeds to the concentrated actor.
These failures matter. They do not make the incumbent the right unit. They define the conditions under which abundance is real: growth must create more positions, cheaper capability must expand demand, and access must let non-incumbents become operators rather than only consumers.
The city is not for preserving the current view. It is for creating more lives from which the world is worth seeing. The economy is not for preserving the current task. It is for creating more work in which human purpose can climb as capability gets cheaper.
An incumbent loss may be real. It is not automatically the thing society is for.
Sources: William Stanley Jevons, The Coal Question, "Of the Economy of Fuel"; Michael Saylor's CNBC "cyber Manhattan" argument as reported in "MicroStrategy's Michael Saylor says bitcoin is 'cyber Manhattan'"; Jensen Huang in Lex Fridman Podcast #494 transcript and NVIDIA's Davos writeup, "Largest Infrastructure Buildout in Human History".