Waymo already solved the self-driving problem. They have Level 4 self-driving taxis in Arizona. So, why are they not expanding their self-driving services to other cities, states, or countries? The answer to this question will show us why Waymo will fail.
Waymo will fail because it relies on technology that is not scalable and expensive, such as Lidar and HD 3D digital maps.
Let us explain.
“Building a car is difficult. But making a car that works well with Waymo technology is far more complicated.” It is not our words. Waymo CEO John Krafcik said it.
Waymo relies on 3D digital maps:
We use Google Maps daily. It’s a great map — however, it’s not Google’s success. Hundreds of thousands of people around the world contributed to Google Maps. Many of them still do. However, you would always find outdated or wrong information in Google Maps.
Before a Waymo car can drive in a city or town, Waymo needs to map every road in 3D beforehand. For this reason, they use sophisticated mapping systems. Moreover, they also pour thousands of working hours to label those maps. To digitally map a city is an expensive and labor-intensive process. That’s why you would see even Waymo taxi service is geofenced within a particular area of Phoenix, Arizona.
Now, consider how to scale this HD mapping to the entire Arizona state, then the whole US, and the entire world. Do you think it’s doable? It may be possible if Waymo invests billions of dollars to 3D map every road of this country.
Many people claim that Waymo can accomplish it because Google has a successful Streetview program. But even Google can’t keep their Streetview data updated. On occasions, it is 3 to 4 years old. Moreover, Streetview data is only 360-degree photos, whereas Waymo maps are 3D and need to be precise up to in cm level and must be labeled before it can be used.
Fun Waymo Facts: Every time you encounter an image re-Captcha, where it asks you to find a sidewalk, fire hydrant, traffic lights, cars, busses, etc. you are training the Waymo image dataset.
But the problem does not end there. How to keep up to date these road maps? It will be painful for Waymo to keep every road data updated. Waymo has a solution. Every vehicle will report back to Waymo central data server if it finds any road changes. Then Waymo engineers would validate the data and notify each car in the area about the changes. Nevertheless, it’s easier said than done.
As you can see, even Google can’t keep up to date its StreetView, then how Waymo would be able to keep their crucial 3D maps updated all the time.
Waymo is dependent on Lidar:
A self-driving technology dependent on Lidar does not mean it will fail. Lidar has many benefits. However, certain drawbacks will cripple Waymos ambition to expand. For example, the Lidar is expensive. Though Google now builds its Lidar and has reduced the price by 90%, it is still costly. According to multiple reports, these Lidar costs around $8,000 each. Waymo needs multiple Lidar because it is unreliable.
Moreover, it performs poorly during rain or snow. That’s why Waymo started its service in Arizona, where the weather is year-round good. Furthermore, Lidar can’t detect colors and read signs. For this, you need cameras. So even if you use Lidar, you need to have cameras too.
Waymo is GPS dependent:
Waymo is also GPS dependent. It uses GPS data to position itself in the 3D HD maps.
The reliance on GPS makes Waymo technology vulnerable to hacking. Someone can jam the GPS signal even with a homemade device. Though it’s not legal in the US, it will not stop criminals from doing it for fun. Moreover, GPS does not perform very well around tall buildings, inside the tunnel, and around bridges.
Internet connectivity issues:
Waymo depends on 3D HD maps. Therefore, the vehicle needs to have onboard downloaded maps at all times. Waymo must download these maps in realtime, or you have to download these 3D maps on your Waymo vehicle before you visit a new Area. Hence, a Waymo car must have to have a fast internet connection at all times. Our current cellular network can’t support these heavy data use by a Self-driving car. Moreover, cellular data is not cheap.
No viable business strategy:
Waymo’s primary business strategy is to sell its technology to the vehicle manufacturer. It is similar to Google’s Android business. Google develops the Android OS, and 3rd party phone manufacturer uses the Android in their phones. Though Waymo has taxi services in several areas, it is only to develop, validate, and demonstrate its technology. So, unless a major OEM comes forward and licenses Waymo’s self-driving hardware and technology, Waymo will have a hard time scaling its product throughout the US, let alone throughout the world.
Self-driving is only a single feature of a vehicle. A car also needs to be good looking, quick and efficient, and have a modern infotainment system. Waymo will thrive only if a car company can thrive with Waymo technology. Everything in a package attracts buyers. Only Waymo self-driving technology is not enough.
Building a car is difficult. But making a car that works well with Waymo technology is far more complicated – John krafcik.
Simulated driving is not real-world driving:
Data is king. The more real-world data Waymo collects, the better. So far, Waymo collected tens of millions of miles of real-world data for their self-driving cars. However, it’s not enough. Moreover, they collected their data from a limited number of cities in the US, let alone outside of the US. They frequently say that their Waymo driver software has driven billions of miles in simulation. But a simulation can’t mimic everything of the condition of our roads.
Every corner case in the real world is a bug for Waymo software. For an engineer, it’s impossible to simulate every real-world corner case. To be successful, Waymo has to work very well not only in the USA but also in Canada, Mexico, Europe, China, India, the Middle East, and Africa. Good luck for those Waymo engineers to simulate every road condition of this world and program Waymo software.
Taxi driving is not easy:
Taxi driving is only profitable in big cities where we have more customers, and payment is substantial. So, if Waymo One wants to be profitable, they have to operate in big cities.
Many people say that Waymo will soon replace taxis in our cities. But we know that’s not going to happen anytime soon. We have talked to numerous Uber and Lyft drivers and asked them what challenges they face daily. Here are a few challenges these drivers face:
- Where to park for the next customer: Riders are not available at all times. There are downtimes when a taxi driver parks their car at a safe and permitted location for the following riders’ call. A human driver struggles to find a parking spot in big cities. A self-driving taxi would struggle more to find a parking spot.
- Where to park and wait for customers: After a driver reaches a customer pickup location, they need to wait for customers to arrive at that location. Finding a perfect spot to park and wait is not easy. For example, Some roads have one lane, especially in suburban areas, where it is impossible to park without obstructing traffic flows.
- Wrong GPS location: GPS and our mobile technology have improved a lot. Still, there are times where the maps would lead you to the incorrect address or GPS data. In self-driving Waymo taxis, it would be a frustrating experience for customers.
- A human driver can follow visual and auditory instruction from law enforcement or other personnel in a construction zone or an emergency. A self-driving tech also has to understand and follow these instructions.
- How is the Waymo One going to handle a drunk or puking passenger?
There are other challenges too. But, we think you can already imagine the complexity of the taxi driving situation.
In short, Waymo would not fail because they depend on Lidar or its price. They would fail because of the complexity and difficulty related to scale up and maintain Waymo technology.