All points
2286 words
- London’s black-cab industry faces a new technological competitor. Artificial intelligence-powered autonomous taxis are preparing to compete with traditional London black cabs. Companies argue that their vehicles are safer and smarter than human drivers, but London's cabbies remain determined to defend their profession.
- Black cabs are deeply connected to London's identity and history. One driver describes them as icons of the city and claims they predate buses and trains. Their importance extends beyond transportation: generations of Londoners associate them with the character and heritage of the capital.
- The Knowledge examination dates back to 1865. Originally introduced for London's horse-drawn cabmen, the examination requires aspiring drivers to memorize approximately 25,000 streets, thousands of landmarks and businesses, and the shortest routes between destinations. It has remained a defining qualification for London's licensed black-cab drivers for 161 years.
- Black-cab drivers build an unusual level of trust with passengers. Tom Scullion, who has driven a London black cab for 34 years, describes customers who send their children to school with drivers they have never met before. He also recalls a regular customer who sends an Irish wolfhound to travel alone in his cab, with instructions identifying where the dog lives. He regards this trust as one of the distinctive features of the profession.
- Passengers routinely test cabbies' geographical knowledge. Scullion says customers frequently question which route he intends to take, sometimes comparing his decisions with Google Maps. He believes a trained London cabbie's knowledge can outperform directions supplied by Google.
- The Knowledge is tested through demanding oral examinations. At Transport for London, aspiring cabbies attend formal examinations called appearances, dressed in their best clothes. Examiners give candidates two randomly selected destinations and require them to describe the shortest route between them, while measuring the distance to determine whether the answer is correct.
- The examination tests detailed route knowledge rather than general familiarity with London. One candidate was asked to travel from Soho House, 40 Greek Street, to the Chancery Rosewood. He had to identify the streets and turns in sequence, including Greek Street, Shaftesbury Avenue, Great Windmill Street and Haymarket. He failed that examination.
- The Knowledge can physically change the structure of the brain. Research from University College London found that black-cab drivers' posterior hippocampi, a brain region associated with memory, increased in size during their careers. The findings demonstrate the neurological effects of years of intensive geographical memorization.
- The official study guide contains approximately 6,000 points of interest. Aspiring drivers must learn the locations of hotels, restaurants, pubs, businesses, cultural institutions and other destinations that paying passengers might request. The material includes places such as the National Gallery, Law Society Hall and Londoner Hotel.
- The Knowledge requires continuous updating as London's businesses change. Candidates must learn individual restaurant and pub locations, remember new business names when establishments change ownership or branding, and adapt when businesses close or new ones appear on the official list. The task therefore extends beyond memorizing a fixed collection of streets.
- Some candidates spend many years attempting to qualify. Steven Fairbrass had been studying for eight years, while Anshu Morjaria had spent five years preparing. Their experience illustrates the extraordinary time commitment required to obtain a London black-cab licence.
- Steven Fairbrass failed his twentieth examination attempt. During his appearance, he was asked to provide a route to Wadenhoe Cafe on Great Titchfield Street. He successfully began describing the journey but lost track of a street name after Portland Place. Despite being encouraged to calm down and continue, he could not complete the route and had to try again.
- An aspiring cabbie named Anshuman Johnny expressed uncertainty about the future of the profession. He acknowledged that artificial intelligence is affecting professions across the economy and admitted that entering a career without knowing what it might become in the near future weighs on his mind.
- Autonomous taxis have not yet been approved to carry passengers in London. Nevertheless, several companies are already testing their vehicles on London's streets in preparation for potential commercial operations.
- Wayve and Waymo are the two principal autonomous-driving competitors featured in the investigation. Wayve is a British startup backed by Nvidia and Microsoft. Waymo belongs to Alphabet, Google's parent company. Both companies hope to begin operating their autonomous taxi services in London later in the year.
- Waymo claims its autonomous vehicles are five times safer than human drivers. Co-CEO Tekedra Mawakana says the company has data supporting that comparison. She argues that autonomous vehicles could help reduce traffic fatalities, noting that approximately one million people die in road accidents worldwide each year.
- Waymo has already established a substantial commercial presence in the United States. It began offering customer rides in a Phoenix suburb in 2020. By the time of the interview, millions of passengers were taking rides in its vehicles across 11 major American cities each month.
- Waymo presents autonomous transportation as a way to give passengers back their time. Mawakana argues that people frequently struggle to complete everyday tasks while driving, such as sending emails or checking on their children. Driverless transportation would allow passengers to handle those activities while the vehicle takes them to their destination.
- Waymo describes its artificial intelligence as an experienced driver despite having no human behind the wheel. Mawakana says its fleet travels more than two million miles each week. She compares that with approximately 700,000 miles driven by a human over an entire lifetime, arguing that the fleet accumulates nearly three human lifetimes of driving experience every week.
- Waymo's vehicles learn collectively from the experiences of its entire fleet. Mawakana confirms that information gathered from one vehicle's journeys contributes to the broader system's experience. This allows the company's AI to benefit from driving encounters across its vehicles rather than relying exclusively on an individual car's experience.
- Waymo also trains its AI through billions of simulated driving miles. These simulations expose the system to unusual and potentially dangerous situations that may rarely occur in real life, including snow on the Golden Gate Bridge and an elephant blocking traffic.
- A demonstration in San Francisco showed how Waymo's autonomous vehicles operate without a human driver. The investigation's reporter travelled in one of the company's taxis with product manager Chris Ludwig. The reporter initially found the experience unsettling, particularly because there was nobody behind the steering wheel, but the journey began to feel surprisingly ordinary after several minutes.
- Waymo's stated objective is to make autonomous journeys feel uneventful. Ludwig describes the company's goal as making the experience "blissfully boring," meaning passengers should feel comfortable rather than constantly aware that an AI system is driving.
- Waymo vehicles use an extensive array of sensors. The demonstrated car contained 29 cameras, six radars, five microphones and five lidar sensors. These systems continuously collect information about distances, surrounding objects and people, with detection capabilities extending as far as three football fields.
- Passengers can see what the autonomous vehicle is detecting. An internal display presents the car's interpretation of its surroundings, including intersections and vehicles that may not be visible to the passenger. During the demonstration, the reporter noticed that the system could identify an intersection approximately 300 feet ahead despite other vehicles obstructing the human view.
- Waymo argues that sensor placement gives its vehicles superior awareness compared with human drivers. The company's sensors can observe areas beyond a driver's normal field of vision, including through gaps and around surrounding traffic. Ludwig says this expanded awareness is one reason the system can outperform human perception in certain situations.
- The demonstration showed the vehicle responding to a pedestrian crossing unexpectedly. A woman talking on her phone walked directly in front of the car. The reporter observed how quickly the vehicle reacted to subtle changes in her movements, including when she appeared to change her mind about crossing. Ludwig explained that the AI had learned to recognize these small behavioral signals and respond accordingly.
- Waymo's autonomous vehicles have also made serious mistakes. In Los Angeles, one vehicle drove through an active police scene. Other incidents involved vehicles obstructing emergency responders and illegally passing stopped school buses.
- Waymo's safety problems have resulted in a software recall and a federal investigation. These incidents demonstrate that extensive training, sophisticated sensors and claims of superior safety do not eliminate operational failures. The company must still address situations in which its vehicles behave inappropriately around emergency services and school buses.
- Waymo uses detailed three-dimensional mapping when entering new cities. Its vehicles have been driving around London to construct a detailed 3D map of the city before commercial operations begin. The company treats this mapping process as a standard part of preparing to operate in a new geographical area.
- Wayve follows a different approach to autonomous driving. CEO Alex Kendall says his company does not need to map an entire city before its vehicles can drive autonomously there. Instead, its AI learns general driving concepts from millions of hours of experience across different parts of the world.
- Wayve compares its AI training process with how humans learn to drive. Kendall explains that a person does not need to memorize every traffic light in the world. Learning how traffic lights work in a few locations allows a human to understand them elsewhere. Wayve applies a similar principle, training its AI to recognize and interpret situations rather than depend on a complete map of every destination.
- Wayve believes its approach will allow autonomous vehicles to adapt to unfamiliar environments. The company says its AI can understand what is happening around it and make decisions in real time, even when travelling somewhere it has never encountered or mapped. Its vehicles use built-in sensors to interpret their surroundings.
- Wayve's autonomous taxis were still undergoing testing and were not available to the public. Kendall demonstrated the technology in London, where the company had been training its fleet since 2019.
- A Wayve demonstration took place around Westminster, one of London's most challenging driving environments. The area contains historic landmarks, tourists, busy roads, multiple vehicle types and complicated intersections. Kendall used the demonstration to show how the AI responds to unpredictable traffic conditions.
- Wayve's AI controlled the vehicle during the Westminster demonstration. It operated the steering, speed, indicators and brakes while Kendall sat in the driver's seat without touching the pedals or controls. However, a human safety driver must remain present until the government approves autonomous taxis for passenger operations.
- The Westminster demonstration exposed the variety of situations an autonomous system must handle. The vehicle encountered a driver moving into its lane, changing lanes again and a cyclist whom it needed to wait for before making its own lane change. Kendall argued that roads contain an enormous range of unpredictable events and that AI could have an advantage in processing them.
- Wayve's central argument is that AI can learn to handle situations humans may never anticipate. Kendall believes the technology's ability to interpret its surroundings and make real-time decisions will be particularly valuable in complex urban environments.
- Steven Fairbrass remained confident that human drivers would always be necessary. He rejected concerns about autonomous vehicles replacing cabbies, arguing that the human brain would remain the strongest tool. He imagined a passenger standing in the rain trying to hail a driverless vehicle that simply passes because its sensors fail to recognize the request.
- Fairbrass believes autonomous vehicles cannot fully replace the judgment and flexibility of human cabbies. His concern is not simply whether a vehicle can follow roads safely, but whether it can understand unusual human requests and respond appropriately to circumstances outside its programmed expectations.
- London's black-cab industry has suffered a substantial decline over the previous decade. The number of licensed black-cab drivers has fallen from approximately 25,000 to 16,000, a reduction of 9,000 drivers, or roughly 36%.
- Black-cab drivers' incomes have also declined. The transcript attributes growing pressure on the industry to Uber and other ride-hailing companies, which have taken business away from traditional taxis. The economic difficulties therefore predate the potential arrival of autonomous taxis.
- Hundreds of people still register to study The Knowledge every year despite the industry's decline. The continued interest shows that the qualification retains considerable appeal, even as established drivers face falling numbers and reduced earnings.
- Steven Fairbrass's twentieth failed attempt illustrates the personal cost of pursuing the profession. After eight years of study, he was still trying to obtain his licence, demonstrating how demanding the qualification remains even for determined candidates.
- Anshu Morjaria finally passed The Knowledge after five years and 41 attempts. During his successful examination, he was asked to describe a route to Ladywell Station, beginning on Brockley Road and continuing through Adelaide Avenue, Ladywell Road and Railway Terrace. The examiner eventually confirmed that he had passed, allowing him to obtain his black-cab licence.
- Passing The Knowledge was a significant personal achievement for Morjaria. His reaction included thanking the examiner repeatedly and celebrating the successful completion of five years of preparation and 41 examination attempts.
- The attraction of becoming a black-cab driver goes beyond finding a quicker route into employment. When asked why someone would spend years studying instead of becoming an Uber driver more quickly, Morjaria pointed to the history and cultural significance of London's black cabs.
- Black-cab ownership and driving are associated with a sense of London identity. Morjaria compared the idea of London without black cabs to London without a king or queen, arguing that the city would lose something essential to its character. His remarks explain why aspiring drivers continue pursuing The Knowledge despite the profession's economic challenges and uncertain future.