Enlarge /. John Krafcik, CEO of Waymo, speaks in 2018.
Justin Sullivan / Getty Images
Many Tesla fans consider the electric car maker to be a world leader in self-drive technology. CEO Elon Musk himself has repeatedly claimed that the company is less than two years away from perfecting fully self-driving technology.
In an interview with German manager magazine John Krafcik, CEO of Waymo, dismissed Tesla as a Waymo competitor, arguing that Tesla's current strategy is unlikely to ever produce a fully self-driving system.
"For us, Tesla is not a competitor at all," said Krafcik. "We're making a completely autonomous driving system. Tesla is an automaker that is developing a really good driver assistance system."
For Musk, these two technologies exist along a continuum. His plan is to gradually improve Tesla's autopilot software until it is good enough to work without human supervision. But Krafcik argues that that's not realistic.
"It's a misconception that you can just keep developing a driver assistance system until one day you can magically jump to a fully autonomous driving system," said Krafcik. "In terms of robustness and accuracy, for example, our sensors are orders of magnitude better than what we see from other manufacturers on the road."
This is not a new argument.
Tesla keeps missing Musk's predictions
Tesla is working hard to develop technology that is completely self-driving. Since 2016, Tesla has been selling cars with a range of cameras, radars, and other hardware that will be touted as ready for full self-driving in a future software update. That same year, Tesla began charging thousands of dollars for a "fully self-driving" software package that would later be delivered.
If Tesla can't deliver fully autonomous software, there will be a lot of disgruntled customers. Musk initially predicted that this technology would be available by 2018, but repeatedly pushed back the timeline.
Last October, Tesla finally released a beta version of its "fully self-driving" software for a select group of Tesla customers. While previous versions were limited to certain driving environments – mainly highways – the FSD Beta was designed for most common types of roads and intersections. It could stop for traffic lights, turn left, and navigate roundabouts.
But the software was clearly not ready for widespread release. I watched a few hours early unedited footage posted by Tesla owners who received the new software. The software made a number of errors, including two incidents where a Tesla appeared to be on the verge of colliding with another vehicle before the driver intervened.
Tesla is of course working to fix these issues and improve the software over time. Musk believes the software will shortly be good enough to operate without active human supervision. And not long after that, it will be good enough to work completely autonomously – without anyone sitting in the driver's seat or in the car at all.
Waymo rejected Tesla's strategy years ago
But Waymo's leaders have long doubted that premise. They believe lidar sensors will be essential to getting self-driving vehicles on the road early. They also believe that the transition from a driver assistance system to a completely driverless system is fraught with danger.
The Waymo team believe their own early experiences – when it was Google's self-driving car project – confirm this. In the early 2010s, Google developed a driver assistance system similar to today's autopilot and considered selling it to automakers. When they had Google employees test the software on public roads, they found that the drivers trusted it far too quickly. Drivers who were supposed to closely monitor the system instead spent their time checking their phones, putting on makeup, and other distractions.
The fundamental challenge is that the better a driver assistance system becomes, the more difficult it is to get drivers to pay attention and the less likely they are to be prepared if the software makes a mistake. The Google team didn't see a good solution to this problem, so they completely changed their strategy. They focused on building a self-driving taxi service where customers would never be in the driver's seat, and trusted that trained, professional safety drivers would oversee the software during the tests.
According to Krafcik, Waymo has largely completed the engineering work on its self-driving software and is now focused on scaling the technology. In this case, the company may be able to demonstrate the technical and economic feasibility of its approach in the next few years. Musk dismissed Waymo's approach as a "highly specialized solution" and questioned whether Waymo could scale it.
Despite the repeated failures of his earlier predictions, Musk continues to insist that Tesla's full self-driving technology is about to be released. "I am extremely confident that I will achieve full autonomy and pass it on to the Tesla customer base," Musk stated last month in 2021. Meanwhile, Krafcik believes that Tesla's approach is a dead end.
In the next few years we may finally find out whose theory is correct.