One of the things I continually hear when people talk about “the gig life” is the assumption that because an app operates somewhere, there must be enough work there for everyone who signs up.
That simply isn’t how markets work.
A small town does not suddenly develop the transportation demand of Boston because DoorDash, Uber, Lyft, or another platform allows people there to create driver accounts.
There are only so many rides.
There are only so many deliveries.
There are only so many customers willing to pay for them.
And there are only so many drivers a particular market can reasonably support.
That seems painfully obvious, but it’s something the gig economy has spent years avoiding saying out loud.
Not Every Market Can Be a Full-Time Job
I’ve been doing rideshare for three years, and I’ve done delivery apps as well. One thing you learn quickly is that every market is different.
A large metropolitan area can potentially support a tremendous number of drivers because it has population density, airports, hotels, hospitals, businesses, entertainment, commuters, tourists and people constantly moving around.
A small community doesn’t have those numbers.
Maybe that town can support three drivers having pretty good days.
Put thirty drivers there and suddenly everybody thinks the platform is dead.
The platform may not be dead at all.
You’ve simply put too much supply into a market that doesn’t have enough demand.
Then add seasonal changes.
Add school calendars.
Add tourism.
Add weather.
Add changes in the economy.
Add established drivers who have already been working that market for years.
Then add another wave of people who were told they could download an app and start making money whenever they wanted.
That’s when the promise of gig work runs headfirst into basic economics.
You can certainly make gig work a full-time occupation. People do it.
That does not mean every geographical market can support every person who wants to make it a full-time occupation.
There eventually has to be some mechanism balancing supply with demand.
Whether the rideshare companies eventually call that scheduling, driver tiers, priority access, reservations, geographic limits or something else, the underlying problem remains the same:
You cannot indefinitely keep adding vehicles to a market when the number of customers isn’t growing with them.
And now the transportation companies want to add another source of supply.
Autonomous vehicles.
Now Add Cars That Don’t Need Drivers
Robotaxis aren’t entering an empty transportation system.
They’re being placed into markets where taxi companies, limousine operators, public transportation, rideshare drivers and other transportation providers already exist.
Waymo and Zoox are continuing to expand into additional American cities, meaning this isn’t some distant theoretical discussion anymore. (Reuters)
That means the supply-and-demand question becomes even more important.
If a city needs 10,000 vehicles to meet passenger demand and there are already enough human-operated vehicles available, adding another 2,000 autonomous vehicles doesn’t magically create another 2,000 vehicles’ worth of customers.
Somebody loses utilization.
But there’s an even larger problem with the autonomous-vehicle discussion.
We’re acting as though the only thing a rideshare driver does is operate the steering wheel.
Anyone who has actually transported the public knows better.
People Are Nervous About This — and That’s Not Irrational
There are people who genuinely find the idea of climbing into a vehicle without a driver unsettling.
That isn’t some tiny fringe reaction.
A Pew Research Center survey conducted in February 2026 found that 71% of American adults said they would be either not too comfortable or not at all comfortable riding in a driverless car. Only 7% said they would be extremely or very comfortable. (Pew Research Center)
So when people tell me they’re nervous about getting into a vehicle where nobody is sitting behind the wheel, I don’t dismiss that as people being afraid of technology.
It is an unusual experience.
Human beings have spent generations understanding transportation as something where another person—or historically another living creature—is responsible for getting you there.
There’s an old joke about horse-drawn transportation that actually illustrates the difference wonderfully.
There have been real cases involving intoxicated buggy operators. In one Ohio incident in 2022, police encountered an allegedly intoxicated Amish buggy driver who appeared to be passed out while his horse continued pulling the buggy. (https://www.cleveland19.com)
People immediately make the joke:
The horse knows the way home.
And, frankly, if I had to choose some completely unattended form of transportation, there’s something strangely reassuring about the old horse.
That animal has instincts.
It knows its surroundings.
It knows where home is.
It knows something is standing in front of it.
It can become frightened.
It can decide something is wrong.
That doesn’t make a horse-and-buggy an autonomous transportation solution. The Ohio incident itself demonstrates why—police ultimately had to intervene.
But it does illustrate something we’ve always understood about transportation:
Something has to be responsible for the situation, not merely the route.
And that’s where robotaxis raise questions that have almost nothing to do with whether their computer can successfully make a left turn.
The Driver Is On-Site Customer Service
I once watched somebody get out of my vehicle, walk several steps away and fall down.
At that point, I wasn’t “driving” them anymore.
The trip was effectively over.
But I was still there.
I could see what happened.
I could determine whether that person got back up.
I could call for help.
I could tell emergency responders what I had witnessed.
That’s part of transportation that doesn’t appear on the mileage report.
Now imagine that same situation with an autonomous vehicle.
Passenger exits.
Passenger collapses.
Does the vehicle recognize the fall?
Can it distinguish somebody sitting on the ground from somebody experiencing a medical emergency?
Does it leave?
Does a camera notify somebody?
Does a remote support employee have to review what happened?
Who calls 911?
Who remains there until help arrives?
What happens if the passenger is unconscious and therefore cannot push the big HELP button?
California’s own regulatory material for autonomous passenger transportation has had to contemplate remote operators and communications links because these passenger-service questions don’t disappear simply because the steering has been automated. (California Public Utilities Commission)
And that’s only an emergency.
Consider what drivers deal with during completely ordinary rides.
A family arrives with a small child but no appropriate child restraint.
Too many people attempt to get into the vehicle.
Somebody is dangerously intoxicated.
A passenger becomes sick.
Two passengers start fighting.
Someone leaves a phone behind.
Someone damages the interior.
Someone leaves something dangerous or disgusting in the back seat.
The passenger getting into the next ride doesn’t care that the autonomous-driving software successfully navigated seventeen traffic lights.
They care that somebody vomited on the seat.
Waymo’s own passenger rules demonstrate how much of this exists outside the driving task. Its vehicles have a maximum passenger capacity, children under eight require appropriate child restraints, seat belts are required, and riders can be charged cleaning fees for things including smoking, vomit and other significant messes. (Google Help)
So clearly those problems still exist.
Removing the driver didn’t remove human behavior.
It removed the human being who was standing there to deal with it.
We’re Not Hotel Rooms
Then we get into the behavior people exhibit when they believe they’re alone.
People already joke about the Mile High Club.
Do we seriously believe somebody isn’t eventually going to climb into an autonomous taxi and think:
“Well, honey, we’ve got some privacy.”
Except they don’t.
Waymo explicitly tells passengers that interior cameras are on and recording during rides, while microphones are generally off unless the passenger is connected with Support. It also provides 24-hour support through the vehicle or application. (Google Help)
That produces an interesting contradiction.
The absence of a driver makes the vehicle feel private.
But maintaining safety without that driver can require additional surveillance.
Now the company needs to identify vandalism.
Smoking.
Drug use.
Medical emergencies.
Improper occupancy.
Damage.
Bodily fluids.
Potentially sexual activity.
And whatever other creative behavior human beings eventually discover once somebody gives them what feels like a private room on wheels.
So who’s monitoring that?
A computer?
A remote employee?
When does automated monitoring become human monitoring?
What gets recorded?
How long is it retained?
Who gets to watch it?
Suddenly the driverless taxi isn’t quite as private as people imagined.
Who Replaced the Person?
This is the question I think regulators should be asking.
Not merely:
Can the car drive itself?
Ask:
Who replaced the driver?
Because those aren’t the same question.
The driver was operating the vehicle.
But the driver was also enforcing passenger capacity.
The driver was checking whether the child had a car seat.
The driver was recognizing medical problems.
The driver was observing passenger behavior.
The driver was protecting the vehicle.
The driver was checking the back seat before the next passenger arrived.
The driver was handling unexpected situations at pickup and drop-off.
The driver was the witness when something happened.
The driver was the person a passenger could immediately speak to.
In other words:
The driver was on-site customer service.
A remote support center may replace the telephone customer-service department.
It does not automatically replace the human being physically present at the vehicle.
Technology may eventually provide good solutions to many of those problems.
Fine.
Show us.
Demonstrate them.
Test them.
Regulate them.
Establish who is responsible when they fail.
But don’t tell the public that you’ve successfully replaced the driver because the vehicle managed to stay between the lines.
You replaced the steering function.
That’s only part of the job.
Transportation Has Always Been More Than Moving the Vehicle
And that’s where these two apparently unrelated gig-economy conversations ultimately meet.
A small town doesn’t need unlimited drivers simply because unlimited drivers are willing to work.
And a city doesn’t automatically need unlimited autonomous cars simply because manufacturers are capable of putting them there.
Transportation is a service responding to actual demand.
And transportation involves considerably more than moving a machine from Point A to Point B.
We should be asking how many vehicles a community actually needs.
We should be asking what happens to workers when platforms deliberately oversupply those markets.
We should be asking what responsibilities transportation companies assume when they deploy autonomous fleets.
And most importantly, we should stop pretending that removing the person from the driver’s seat removes the jobs that person was performing.
Because when something unusual happens—and anyone who has spent enough time transporting the public knows that eventually something will happen—the question isn’t going to be:
Did the computer know the route?
The question is going to be:
Who is here to deal with this?
And sometimes, apparently, even the horse had an answer.