Having accepted that Waymo has beaten Tesla to rolling out actual driverless robotaxis, Tesla fanboys have taken refuge in the notion that Tesla Full Self-Driving is better that Waymo’s autonomous driving software because it is able to drive anywhere in North America, while Waymos are geofenced into small areas. They rely upon the argument that the areas in which Waymos are able to operate is currently restricted, and if you dropped one of their vehicles in an area outside of their geofence, they would either not be capable of navigating their new surroundings, or would unleash chaos on the roads in which they found themselves.
Tesla fanboys argue that Tesla’s Full Self-Driving software is able to drive anywhere at any time, which they say presents Tesla with a natural advantage in the race for autonomy. Unlike other self-driving manufacturers such as Waymo and Cruise, who test their vehicles in carefully mapped and geofenced environments, Tesla has deployed their self-driving technology to “anyone in North America”, according to Elon Musk.
Let’s examine whether Tesla Full Self-Driving’s ubiquity on North American roads puts them ahead of their geofenced competitors, and whether it constitutes the competitive advantage in the race for autonomy as their fanboys and Elon Musk claim. Elon claims that Tesla is so far ahead in autonomous technology that he can’t “see a second place with a telescope“. In fact, he can’t see the competition because Tesla is so far behind.
Firstly, Waymo and Cruise vehicles are in fact able to drive anywhere with the presence of a safety driver:
With a safety driver Waymo & Cruise aren’t geo-fenced. They can drive anywhere for tens of thousands of miles without human intervention. That’s hundreds of times better than @Tesla Full Self-Driving on its best day, on the easiest route, in the best weather, with no traffic. https://t.co/1wcGGguRoj
— Dan O'Dowd (@RealDanODowd) July 30, 2023
Autonomous vehicle manufacturers in California are obliged to report disengagement statistics for their autonomous fleet, which are made publicly available on the California DMV’s website. These statistics show that supervised Waymo and Cruise vehicles on average travel 95,000 and 17,000 miles respectively between interventions, which is vastly superior to Tesla Full Self-Driving, which is lucky to go 17 miles without needing human help. Unlike Tesla, Waymo and Cruise vehicles are able to drive safely in geofenced areas without constantly requiring an attentive driver to intervene:
Tesla’s Full Self-Driving software lags far behind real autonomous vehicle manufacturers in the average number of miles the software can travel between disengagements. Guidehouse Insight’s analysis of autonomous vehicle companies’ strategy and execution ranked Tesla as a “follower”, and dead last in the race to full autonomy.
Though it is not just actual autonomous robotaxis that Tesla lags behind the competition. In fact, even as an Advanced Driver Assistance System (ADAS) Tesla is ranked in tenth place by Consumer Reports:
Tesla is the laughing stock of the autonomous vehicle industry. Ten years ago, Google threw into the dumpster a better self-driving system than FSD is today, because they knew it would never be safe enough for general release. Google decided not to bring AutoPilot to market, because their employees did not maintain attention as required during their commutes using the system no matter how often Google warned them. Then they started over. Now, Waymo, owned by Google’s parent company Alphabet, has many autonomous vehicles on the road today in many major cities, driving tens of thousands of miles between incidents.
Elon Musk realized that if he could get to market first with a self-driving car, owning any other car “would be like owning a horse”. This would enable him to sweep away the legacy auto companies. Time to market was the number one priority, so Elon Musk built his Autopilot using the same approach that Google had rejected for safety reasons.
But now, having re-evaluated their approach, Waymo has autonomous robotaxis, with no driver in the car. It is also expanding its driverless robotaxi service in Phoenix to include freeways, but only for their employees for the moment, as well as announcing that it will still commence operations in Los Angeles.
LA, this one’s for you: We’re pleased to share that the @CA_DMV has granted a modification to @waymo’s Driverless Testing Permit, expanding our fully autonomous geography to include LA – our next ride-hail city. 🤖🚘🌴 pic.twitter.com/N2BAOnJESF
— Waymo (@Waymo) December 22, 2022
When these systems are up and running they can be connected, further leaving Tesla’s failed self-driving project in its wake as Waymo continues to obtain real autonomous driving data in more and more complex environments.
In the meantime, Cruise has scaled from one major city to seven in less than a year:
Today we are announcing Nashville 🎸as our next robotaxi-enabled city!
You’ll see driverless @cruise AVs there in a few months.
This brings us from 1 city to 7 in about a year – and there are still more to come…
— Kyle Vogt (@kvogt) July 27, 2023
While Tesla continues to collect zero miles of autonomous driving data every day, Cruise and Waymo are scaling exponentially and collecting thousands of autonomous miles. As these manufacturers collect more and more data, each subsequent cities becomes easier to navigate.
Tesla Full Self-Driving, as it is today, is a worthless pile of trash as a level 3 or 4 autonomous vehicle. Even its staunchest defenders will tell you that it will “randomly try to kill you”:
If you're trying Tesla Full Self-Driving Beta for the first time, it's important to remember that it will at some point randomly try to kill you. This is a when, not an if.
As such you must keep your hand on the wheel and be ready to takeover at any time. You must pay attention
— Whole Mars Catalog (@WholeMarsBlog) November 22, 2022
The fanboys argue that drivers using FSD must remain alert and be ready to take over at any time. They further claim that the combination of an alert human driver and FSD renders FSD safer than the products of Tesla’s competitors. However, there are countless examples of FSD Beta testers posting videos where they commit critical driving errors that they would have never committed without FSD being activated. For example, during a fifteen minute video posted by one tester, a driver using Full Self-Driving v11.4.4, the software ran three stop signs, slammed on the brakes unnecessarily, went over the yellow line repeatedly, tried to take the wrong exit, and stopped at the wrong stop sign, mistakes that a driver without FSD activated would never commit. FSD and a human is demonstrably worse than a human driver, and in fact drives like a drunk driver.
Yesterday @TeslaTruckClub posted the below video and claimed that Tesla Full Self-Driving “is safer than a human. FSD beta + human is better than just human”.
In fact, FSD + Beta Tester is the worst driver on the road. For instance, in less than 15 minutes of his driving footage…
— Dan O'Dowd (@RealDanODowd) July 5, 2023
The narrative that Waymo’s geofenced restrictions leave it behind Tesla in the race for autonomy is entirely false. It is now obvious that Waymo and Cruise are technological leaps and bounds ahead of Tesla. They have driverless cars on the road today that go thousands of miles without needing any human intervention, while Tesla’s “Full Self-Driving” car requires a human driver to intervene every few miles to prevent it from crashing into something or someone. Tesla’s ‘self-driving’ software needs more attention and supervision than a 16-year-old new driver.
So, summing up, Waymo and Cruise thrash FSD at everything, everywhere.