Autonomous autos – Faulty steering.

Removing the safety driver can lead to a misinterpretation of reality. 

• Cruise has received permission to run its vehicles without safety drivers being present which may say more about the quality of its remote pilot system rather than the autonomous solution itself.
• Cruise does not have a bad solution for autonomous driving (although it is way behind on its initial promises (see here)) as RFM rates it as No. 3 in its rankings behind Waymo and Baidu.
• However, as far as the regulator is concerned, the biggest difference between having a safety driver in the vehicle or at a remote location is going to be the quality of the remote piloting system, not the autonomous driving system itself.
• Obviously, the autonomous driving system has to be at a certain level of performance before the regulator will consider allowing an empty cockpit, but the primary concern will be for the safety of other road users.
• Hence, the ability of a remote driver to quickly take control of a vehicle should go wrong will be the most important factor being considered when issuing a licence to operate without a safety driver in the vehicle.
• As is typical in this sector, this is being framed as “we are so good at driving, that there is no safety driver” when in reality the safety driver is still there but located at some remote location.
• This is the normal progression from having a safety driver to having the vehicle run around on its own and there are many intermediate steps.
• At the outset, it would typically be one safety driver at the remote location watching only one vehicle all of the time.
• (This may be why Cruise is only running 5 vehicles initially).
• As the system improves, one remote driver could then monitor more than vehicle and so on until the safety driver is no longer needed.
• There is no evidence anywhere to suggest that we are even close to a stage where the drivers can be removed completely and so I think it will still be around 2028 before this becomes a reality.
• To be fair to Cruise, this is definitely a step forward because it will continue to run its vehicles in downtown San Francisco which is an environment that is far more complex than Mountain View or Phoenix.
• However, it is also worth pointing out that the average speed in San Francisco is much lower than Mountain View meaning that the vehicles have more time to react when a hazard is detected.
• The company that I think is furthest down this route is Yandex which claims it is ready to run vehicles in Russia with no dedicated safety engineer in the vehicle or in a remote centre.
• However, the locations where Yandex is launching this are in the middle of nowhere and Google Maps reveals what must be one of the simplest and quietest driving environments available anywhere in the world (see here).
• The net result is that Cruise is justifying its place among the leaders of this industry, but I do not think that the grant of this licence pulls it ahead meaningfully of any of its competitors.
• There are still way too many players in this market and given RFM’s expectations that there will be no meaningful revenues before 2028, I continue to expect substantial consolidation.
• The bloodletting is not over.