Cruise, a division of General Motors, announced that they were offering driverless rides in San Francisco. The company started back in 2013 and raised more than $10 Billion from the likes of GM, Honda, Microsoft, Softbank and Walmart to name a few.
It seems finally the company is making robotaxi rides available to the general public. As per the website, Cruise promotes that passengers (or riders) can get in, crank up the volume and have a snooze. Gone are the days of having a human-supervised driver in the driver’s seat just in case, Cruise is clearly confident enough with their technology to accept rides.
In the latest video published to their YouTube channel, Cruise shows passengers get into Cruise vehicles in San Francisco to experience their first fully driverless ride.
Why passengers and not customers?
Those paying attention would note that these passengers are not customers. If you’re a company that’s spent more than US$10 Billion of other people’s money on building an autonomous car, wouldn’t you want to start charging for rides as soon as its ready?
This is a really important point. The people you see in the video above are making their testimonials, based on receiving a free ride. Obviously, you are likely to be fairly forgiving of the riding experience, when you’re taking a free ride, versus paying for an Uber.
So what’s going on here?
Cruise is running this service under a very restrictive test permit from the California DMV that allows them to test their autonomous vehicles, including members of the public, but are not yet able to charge for rides.
The vehicles are designed to operate on roads with posted speed limits not exceeding 30 miles per hour, during all times of the day and night, but will not test during heavy fog or heavy rain.
It turns out Cruise isn’t exactly the pioneer here, they were actually the fifth company out of seven that have now received a driverless testing permit. Competitors include Apollo, AutoX, Nuro, Waymo, Weride and Zoox.
There are also currently 60 companies that have an active permit to test autonomous vehicles with a safety driver which includes BMW, Honda, Mercedes Benz, Nissan, Tesla and VW to name a few.
In order to receive a driverless testing permit, manufacturers must certify they meet a number of safety, insurance and vehicle registration requirements, including:
- Providing evidence of insurance or a bond equal to $5 million.
- Verifying vehicles are capable of operating without a driver, meet federal Motor Vehicle Safety Standards or have an exemption from the National Highway Traffic Safety Administration, and is a SAE Level 4 or 5 vehicle.
- Confirming vehicles have been tested under controlled conditions that simulate the planned area of operation.
- Notifying local governments of planned testing in the area.
- Developing a Law Enforcement Interaction Plan that provides information to law enforcement and other first responders on how to interact with test vehicles.
- Continuously monitoring the status of test vehicles.
- Training remote operators on the technology being tested.
- Driverless testing permit holders must also report to the DMV any collisions involving a driverless test vehicle within 10 days and submit an annual report of disengagements.
Credit where credit is due, the riders appear to get into the vehicle and are taken to their destination without issue, which is a big achievement for the autonomous vehicle industry. The problem is that this service is radically limited in its geofenced area of operation.
If you’re familiar with San Francisco, this is a tiny percentage of the city and if you got into a Cruise and asked it to take you where you wanted to go, a majority of the times, it could not. You get out and would likely call an Uber and never use the service again.
Of course, Cruise is expected to grow this coverage area over time, but the technology stack they’re relying on is a combination of Lidar and HD Maps. These are fragile and as soon as something changes could create issues for the car. Sure, in a well-mapped environment, the car can successfully navigate its way through the city streets, but for commercial service to make sense, customers need to have the flexibility of being driven almost anywhere they wish.
If we compare this approach to Waymo, we find the technology stack is very similar, but compare it to the approach of Tesla and its night and day.
To many, if you watch these videos, you may think they’re well ahead of what Tesla is offering, given you can’t go to sleep in a Tesla today. I’m not sure how much sleeping you can get done on a 5-minute ride exactly but Tesla’s approach is far more generalised.
Tesla doesn’t have a coverage map like Waymo or Cruise, their coverage area is the entire United States right now and soon to be Canada and when the software is ready, will roll across all regions Tesla sells vehicles in. Tesla doesn’t use Lidar to scan the environment around the vehicle. They don’t rely on HD Maps as ground truth, instead, they use computer vision to determine the environment around them and identify the driveable space to plot a safe route through.
Ultimately this looks like Cruise and others are in the lead and if you look at it with blinders on, they are, but what’s likely to happen over the next 12-24 months is that these small test areas are rapidly going to be consumed by a competitor playing a very different game.
If Tesla can land FSD with the ability to navigate city streets and highways, then they setup up a robotaxi network like we’ve never seen. They have the chance to scale to dozens of cities overnight, while others have taken decades to reach their first.
Back on January 28th, Cruise released a video of GM’s CEO Mary Barra, taking what was promoted as her first ride in a driverless vehicle. That’s a little hard to believe that she’d never ventured inside one before but fine, we’ll assume that’s true, take a look at the video below.
Something that stuck out to me was the fact these vehicles are not finding a safe parking space to pull over and let passengers exit, they simply stop in the middle of the street, something that’s completely unacceptable when there’s serious traffic around.
I hope autonomy is a race between multiple companies and we as consumers end up with a price war, similar to what we see with Uber and Lyft, but that seems unlikely with such vastly different approaches to the problem.