Was this essay forwarded to you? Join for Free Here

"Oi, tudo bem?"

My driver was floored. I don't speak Portuguese. I had picked up the casual greeting years earlier on a trip to Brazil.

Orlando is a frequent stop for conferences, and I speak there often. This one event had arranged a driving service to pick me up from MCO, and that is how I met Renato. We talked about Guaraná, a Brazilian soft drink I enjoy, and about real churrascarias, and he told me about a hole-in-the-wall place in Orlando most tourists never find.

Ever since, when I book him, he has my candy and an ice cold Guaraná waiting for me.

Sadly, that human experience is being traded for a machine. Just two weeks ago in Orlando, I saw more Waymos than Black Cars. And that is the direction we are headed with autonomous vehicles.

In August, I predicted driver's ed would become optional for this year's kindergartners. In this essay, I'll go further. Driver's licenses will be optional by the time they turn 16, and no longer necessary at all for Gen Beta. Hallelujah, for no more DMV lines.

To be honest, I'm not worried about autonomous vehicles. I'm worried about what we lose when we trade driving.

Maybe the most iconic job for a driver is getting behind the wheel of one of those infamous NYC yellow cabs. As Will Ferrell (playing Buddy) in Elf famously quipped, “The yellow ones don’t stop” as he recalled getting hit by a cab in his green tights. There may be some comical truth to that.

In 1938, New York City capped the number of taxi medallions at 13,595. A medallion cost just $10 (City Limits, 2019). Today, there are just eight fewer (NYC TLC, 2025).

By 1950, medallions averaged $5,000 (City Limits, 2019).

By 2014, one sold for $1.3 million (Reuters, 2017).

Securing a taxi medallion was one of the more valuable assets in New York City, passed down from generation to generation, right down to the city rule that gives an heir 180 days to qualify before the medallion has to be sold (35 RCNY § 58-48). It created generational wealth. It created access. It created exclusivity. Until it was disrupted.

Then, by May 2021, the average sale of a medallion plummeted to a mere $79,105 (NYC Taxi News, 2022).

I've ridden in plenty of those cabs. On a flat-rate ride out to Melville on Long Island, an already grumpy driver got so angry that I was paying by card that the hotel doorman tapped on the window and intervened. He sent me inside, and the two of them yelled it out on the curb. I paid him. I never learned how that argument ended.

Then the app showed up. Uber launched in San Francisco in 2010 and reached Raleigh four years later (TheStreet, 2020; Uber, 2014). I still remember watching that first Uber drive up, questioning: Do I get in the back? Is this really like a cab? Do I talk to the driver?

The shift from raising your arm or whistling from the curb to opening an app happened fast…very fast. In Manhattan, Uber rides rose by 3.82 million from April to June 2015 compared with a year earlier, while yellow cab pickups fell by 3.83 million (Stempel, 2015). Nearly every ride Uber gained was traded for a cab ride. Medallion owners and their lenders sued the city. In 2017, a federal judge dismissed the case, ruling that hailing a ride on an app is not the same as hailing a cab on the street (Reuters, 2017).

What happens to the people who drive for a living, like the black car driver who waited for me at JFK holding a sign with my name or the driver with the Brazilian candy? We're already seeing it with the yellow cabs. New York's TLC alone licenses about 179,000 drivers (TLC, 2024). Nationally, taxi, rideshare, shuttle, and chauffeur work accounts for about 447,900 jobs at a median wage of $36,660 (Bureau of Labor Statistics [BLS], 2025a). That is roughly everyone in Minneapolis (Minnesota State Demographic Center, 2026). Add 2.2 million heavy and tractor-trailer truck drivers on top of it (BLS, 2025b).

BLS projects that taxi, shuttle, and chauffeur work will grow 9% through 2034, much faster than average (BLS, 2025a). Those are extremely optimistic growth charts. And frankly, I don't agree with them at all. I think they are way off. We already watched this happen once. In a single year, nearly every ride Uber gained in Manhattan came straight out of the back seat of a cab.

Robotaxis will not move quite as fast as Uber did, but the trajectory is very aggressive. Going from a taxi driver to an Uber driver was moving from one human being to another with a different title, and moving from a yellow car with a roof light to everyday cars that were already parked in driveways. Going to no driver at all is a paradigm shift. It takes new hardware on the road, and it takes the mental shift of a rider willing to get into a car with nobody in it.

That is a real hurdle. But people are accepting that driverless option far more quickly than you may guess. Waymo went from 50,000 paid rides a week in May 2024 to 500,000 by March 2026, a tenfold jump while its fleet grew about sixfold (TechCrunch, 2026b). The rides are growing faster than the fleet can carry them, which tells you the appetite is already there. For scale, 500,000 rides a week is still only about half of what New York's TLC-licensed fleet does in a single day (TLC, 2024).

And the purpose-built cars have arrived—meaning a car built solely for the sake of autonomously shuttling people from point A to point B. Tesla's Cybercab has two seats, gull-wing doors, no steering wheel, and no pedals (Associated Press, 2026). It opened to the public in Austin in September 2026, days before I was in Orlando watching Waymos outnumber the black cars.

So the adoption curve will be a little slower than rideshare. However, it will likely not be so slow that there will still be a 9% increase in drivers as the BLS projects.

It's not just robotaxis and Waymos. It's truck driving and our whole infrastructure of ground transportation. Aurora's driverless trucks already run 1,000 miles from Fort Worth to Phoenix without stopping, farther than federal rules let one human driver go without a break (TechCrunch, 2026a).

I recently had a rental with full self-driving capability. The screen told me something like 88% of my miles had been self-driven.

Tesla's Autopilot logged 3 billion miles before Full Self-Driving even launched (Lambert, 2020). Full Self-Driving went from about 6 million miles in 2021 to 4.25 billion in 2025 (Teslarati, 2026), and it crossed 14 billion total this August. Elon Musk had said it would need about 10 billion miles to drive safely on its own (Electrek, 2026).

What human being has ever driven 10 billion miles? Or even 1 billion? A person who drives from age 16 to 72 covers about 744,000 miles in a lifetime (Federal Highway Administration, n.d.). Fourteen billion miles is roughly 18,800 lifetimes of driving.

That is the Velocity Gap. The technology is moving faster than we can adapt to it.

Waymo, with no one behind the wheel, has driven more than 220 million miles, with 94% fewer serious-injury crashes than human drivers in the same cities (Waymo, 2026). And in San Francisco, a Tesla robotaxi ride averaged less than half the price of an Uber (Axios, 2026).

Are robotaxis going to take over? Absolutely. It is the future. I bet on that. It is going to be the main way we get around.

A machine does not get tired, distracted, or angry. It will not get mad that you are paying by card. It will not be high out of its mind, like the Uber driver who picked me up in Los Angeles before an early flight home.

A self-driving car sees every car in every direction at once and decides in an instant. That is safer, as long as the computer is functioning. The human computer, the brain, often is not.

They are still learning, though. In Santa Monica this summer, I watched a Waymo in a one-way alley rock back and forth while a man on the curb leaned forward every time it moved just to mess with it.

North Carolina separated the driver from the license back in 2017. The operator of a fully autonomous vehicle does not need a license, and a 12-year-old can ride in one alone (N.C. Gen. Stat. § 20-401).

We did not learn to drive overnight. Horses were our first autonomous vehicles. My wife and I once rode horseback along a narrow cliffside path in the Santa Ynez Mountains. My horse set hoof in front of hoof, and when rocks gave way at the edge, it kept its balance and kept me upright.

Cars are smarter than horses now. But a horse decides in real time. A car's hardest choices, like what to hit and who to protect, were made in advance by someone writing code. Last November, a Waymo carried a passenger through an active LAPD felony stop, past a suspect lying face down with officers' guns drawn (ABC7, 2025). It knew that was a person. It did not ask whether that person was dangerous. That wisdom will die with a generation.

There are real benefits. The horror stories of rideshare assaults, by drivers and against them, go away. But you're now trusting machines more than other humans to get you safely from point A to point B.

And you trade the conversations. In San Francisco, as the war in Ukraine broke out, my driver had a Ukrainian flag in her car. Several members of my software team were in Ukraine, and we talked the whole ride. She was moved that I knew her country.

So does taking the wheel away give our kids more freedom, or less? Is it like handing an 8-year-old the reins and sending them down to the market alone?