Ajay Agrawal0:02
If you want to have a license to drive a taxi in the city of London in England, you have to go to school for three years to learn what's called the Knowledge. So in your first year, you study maps. That's what they're doing here, they're studying maps. Second year, maps and mopeds. Third year, mopeds, and then you write an exam. So it takes three years, and if you pass the exam at the end, then you get a license to drive a taxi.
Now, of course, today you or I can fly into London and we can rent a car, and even if we've never been to the city of London, we can navigate the city as well as a pro. So for example, if you're asked to drive from the Churchill War Rooms to the Royal Botanical Gardens, you and I can drive that, navigate with a navigational AI just as well as a person who's been to school for three years. The navigational AI predicts the best route between one point and another.
So think about this. Imagine we have the navigational AI. How can we use it? We can use it as a point solution. For example, we can give it to taxi drivers, and the taxi drivers can use this to be better drivers. How much better would they be? Some of our colleagues spent a year and a half studying this question in Tokyo, and their punchline was they found that the inexperienced drivers, the new drivers who got the AI, got a 7% productivity gain. The more high-skill drivers got a 0% gain. They already were very good at knowing how to navigate the city. That would be a point solution. The rest of the taxi system stays the same, and we just increase, we drop in a prediction tool for the driver, but the rest of the taxi system stays the same. That's a point solution.
The system solution, I pick this because everybody here knows it, everyone's seen the system-level solution. There are many different companies that do this in the United States. The predominant one is Uber. Uber used the navigational AI to unlock a total system redesign. So for example, in the United States before Uber, there were about 200,000 people that drove limousines and taxis. 200,000. Today with Uber, there's about 4 million people that drive for Uber. Now you can think about 4 million people. Let's say the average driver brings a $25,000 car onto the road. $25,000 car, 4 million drivers, that's a hundred billion of capex brought into the transportation system like that. Now imagine what would happen if the first time you ordered an Uber, a stranger pulled up in front of your house, you got into their car, and 3 minutes into the drive they pulled over to the side of the street to take out a map to figure out how to get you where you're going. You would probably never take another Uber. The navigational AI allowed anyone to drive like a professional, to navigate like a professional, and so it unlocked a total system redesign. So it was not just giving drivers the AI but keeping the rest of the system the same. Uber redesigned the system. There's no central, the whole thing is done through an interface. It was a system-level redesign. So when you look at these two pictures together, I think it gives us a bit of a foreshadowing of what's coming down the pipe in every industry. And it's not lost on me that when you look at these pictures, there's both good and bad. There's good and bad, and I think we are going to have to navigate as businesses and as society both parts of this.
