To drive a black cab in London, you have to memorize 25,000 streets, 6,000 landmarks, and every shortcut between them — a feat so cognitively brutal that scientists at University College London found it physically enlarges the human brain. Now a Google-backed robot car with no driver, no hippocampus, and no apparent shame is gunning for those jobs. This is either the most compelling argument for human exceptionalism or the most depressing story about capitalism you'll read this week. Possibly both.
161 Years of Brain Damage, Now Threatened by an App
The Knowledge has been breaking people since 1865. That's when London first decided that anyone who wanted to drive a cab — horse-drawn, at the time — needed to prove they actually knew the city. What started as a sensible safeguard for Victorian passengers has evolved into one of the most demanding professional licensing exams on earth.
As CBS News reported in a 60 Minutes segment this past May, candidates must memorize 25,000 streets and 6,000 points of interest across London, then demonstrate the shortest route between any two of them on demand, verbally, to an examiner who is actively measuring the distance. Not close enough. Say it again. Steven Fairbrass has been trying to pass for eight years. Anshu Moorjani, five. These are not people who gave up easily.
The oral exam portion, called "appearances," requires candidates to dress up and recite turn-by-turn routes from memory like they're performing surgery on a map. One candidate in the CBS segment flubbed a route from Soho House to a hotel and failed on the spot. No partial credit. No rubric for effort. The Knowledge doesn't grade on a curve.
The Brain Science Is Real and Kind of Incredible
Here's the part of this story that should make every tech bro pause before writing another hot take about human obsolescence. University College London researchers found that black cab drivers' posterior hippocampi — the region of the brain most associated with spatial memory — actually grew larger the longer they drove. The Knowledge doesn't just test the brain. It reshapes it.
That's not a metaphor. That's measurable, peer-reviewed, biological change in response to years of deliberate cognitive effort. The human brain, when pushed hard enough for long enough, builds new architecture. Tom Scullion, who has been driving London's black cabs for 34 years, compared his knowledge to the difference between a hot dog vendor and Gordon Ramsay. He was talking about Google Maps. He was not wrong.
So when Scullion tells CBS News that passengers quiz him every single ride — "which way you going, mate?" and "Google says this" — and he still wins those arguments, that's not just cabbie bravado. That's a man with a structurally different brain telling you his GPS is better than yours. It is built into his skull.
Enter the Robot Car With No Remorse
Waymo, the autonomous vehicle company owned by Alphabet (Google's parent), is already running robotaxi services across 11 major American cities, with millions of rides logged every month. The company first launched paid rides in a Phoenix suburb in 2020. Now it wants London.
According to CBS News, Waymo hopes to be operational in London later this year. So does Wayve, a British startup backed by Nvidia and Microsoft. Neither company has received approval yet to pick up paying passengers in the city, but both are already running test vehicles on London streets. The future, as they say, is already here. It just hasn't gotten its license yet.
Waymo co-CEO Tekedra Mawakana made the case to CBS News that the safety argument is real and serious. Waymo claims its vehicles are five times safer than human drivers, pointing to data from its fleet, which the company says travels over two million miles every week. For context, Mawakana noted that the average human drives around 700,000 miles in an entire lifetime. The Waymo fleet covers nearly three of those lifetimes every seven days. Those are not fake numbers. They deserve to be taken seriously.
The Part Where the CEO Says 'There Is a Driver' and There Isn't
Mawakana told CBS News, with a straight face, that Waymo prefers to describe its cars as having a "Waymo driver." There is, to be absolutely clear, no driver. The whole point is that there is no driver. That is the product.
To be fair to Mawakana, this is a coherent branding position: the AI system is the driver in the functional sense, making all the decisions, reading the road, executing the route. Calling it a "driver" is a way of saying the car is not driverless chaos — it has a brain, just not a biological one. Fine. But the framing also does some heavy lifting in terms of making passengers comfortable with the fact that the thing keeping them alive in traffic is a software stack.
Mawakana also made a point that will resonate with anyone who has sat in a cab while the human driver behind the wheel sent a text: people are bad at driving while distracted. We want to send emails. We want to check on the screaming kid. We are, statistically, terrible at resisting the urge to do seventeen things at once while operating two tons of metal at speed. Waymo's counterargument is not unfair.
What Actually Gets Lost When the Cabs Go Away
Tom Scullion told CBS News that one of his regulars is an Irish wolfhound. The owner hands Scullion a piece of paper with the dog's destination, the dog climbs in, and Scullion delivers it. "Best customer I've got," Scullion says. "Never says a word, never complains."
Parents hail him to take their kids to school. Strangers trust him completely, on sight, without introduction, because the black cab and the Knowledge behind it represent something: a standard, a credential, a human being who earned the right to be trusted through years of demonstrated effort. That's not just nostalgia. That's a real social contract that took 161 years to build.
No one is putting their dog in a Waymo without a lot of thought about what happens if the software decides to update mid-route. And "take my kid to school" feels considerably more complicated when there's no adult in the vehicle. These are not trivial concerns. They are also not concerns that will ultimately stop the technology from arriving. They're just things we'll lose without much ceremony when it does.
The Dingo Take
Look, the safety data is real and it matters. A million people die in traffic accidents every year worldwide. If autonomous vehicles genuinely reduce that number, the moral case for them is serious and it deserves a serious hearing. Waymo isn't making up the five-times-safer figure for fun. If it holds up at scale, it will eventually outweigh a lot of the human costs of this transition, and we should be honest about that.
But here's the thing: the people getting displaced aren't abstract units of labor. They're Steven Fairbrass, who spent eight years studying for an exam that changed his brain, competing with a company that burns through three human lifetimes' worth of driving experience every single week and calls that a fair fight. The economics of that comparison are not a coincidence. They are the whole point. When tech companies promise their products are safer and smarter and better, they are also, always, promising a world where human expertise becomes worthless faster than humans can retrain for the next thing.
Waymo's co-CEO wants you to think of the car as having a driver. The London cabbie wants you to remember that his driver built his brain from the ground up over years of brutal study just to earn the right to be trusted on streets older than the United States. Both of those things are true. Only one of them is profitable. That's how this ends.
