Bloomberg – Tesla’s ‘Full Self-Driving’ Struggled Soon After Leaving My Driveway
Editors note: this weak technology is supposed to be the basis for Tesla robotaxis? On August 8th?? To keep the stock price artificially inflated????
See original article by Edward Ludlow at Bloomberg
My Month Using Tesla FSD
Elon Musk has reoriented Tesla recently to, in his words, go “balls to the wall for autonomy.” For a little over a month, I tried out the product that’s convinced him Tesla ought to wager its future on robotaxis — even if it means sacrificing the $25,000 electric car he’s also promised investors.
In late March, I leased a Model Y, meaning I was one of the Tesla customers offered a 30-day free trial of what the company calls Full Self-Driving, or FSD. I drove several hundred miles with this system engaged, and while I lived to tell about it and was impressed at times, I can confirm that it doesn’t live up to the name. In fact, FSD tended to have trouble assisting me almost immediately after leaving my driveway.
My first trip on FSD was April 6, the day after Musk set a date to unveil a dedicated robotaxi in August. The system doesn’t make vehicles autonomous, and you the driver are required to have your hands on the wheel and eyes on the road at all times.
My test drives ranged from trips to the grocery store, jaunts to Napa and commutes back and forth between my home outside San Francisco and Bloomberg’s bureau on the Embarcadero. I was unable to make the more than 25-mile journey without at least a few disengagements per trip.
My journey from home to work most days entails taking the famed Golden Gate Bridge, with most of the time spent southbound on the 101 highway. I found FSD to be adept at maintaining pace with traffic. The system was able to change lanes to avoid slower cars ahead or speedy ones coming up behind. It could indicate and then get itself onto the freeway ramp, and capably handled turns in many cases.
On several days, however, FSD would somewhat inexplicably disengage. This was particularly common near to the Robin Williams Tunnel, maybe due to fog or sudden uphill-to-downhill roadway swings. On the other side of the bridge, FSD wasn’t able to handle the toll booths, which require you to go through a lane with a green, illuminated “Go” arrow, and to slow down to 25 miles per hour as you pass. There were times when I proactively disengaged out of concern I wouldn’t slow in time or pick the right toll-booth lane; other times, the system itself disengaged.
Another challenge for the system was a sharp right turn where the 101 rolls into San Francisco, on Lombard and Van Ness. The intersection requires you to pick one of two right-turn lanes, from which you make a protected turn at a right-arrow light.
On each of my journeys, the car would successfully pull into one of the correct turn lanes several blocks in advance. But when it actually came time to make the right, FSD was unable to grasp just how sharp a turn was required. On more than one occasion, the system disengaged on its own, or I did by tapping the brake, because I was straddling two lanes and pointing toward the center embankment with cars around me. The arc of the turn was completely off.
Toward the end of my trial, FSD was sort of able to handle the maneuver. So, there was some evidence of progress even in my abbreviated time using the system.
Other challenges FSD had were closer to home — literally. My driveway leads out to dual-lane roads in both directions, and the exit of the driveway is right-turn only. This means I have to go to the end of the block and come back to get to where I need to go.
Starting from my driveway, I would try using FSD to turn right and then move into left lane to do a U-turn at a four-way stop sign. It could never quite do this. At first, it wouldn’t move from the right lane to the left lane. On one occasion, with my wife in the passenger seat, FSD did manage to get us into the correct lane. However, the system’s attempt at the U-turn put us on a direct path toward the car facing us. I disengaged, and off we went — with me driving even more carefully than normal the rest of the way.
I haven’t yet decided if I’m going to pay the $99-a-month subscription for FSD. Tesla recently cut the price both of the subscription and option to pay for the suite of features outright. I do know that shelling out $8,000 for the system just isn’t realistic for me.
I did find the experience interesting, informative and fun. I learned a lot both about the technology’s progress and its limitations. I also managed to keep using FSD past the 30-day mark — perhaps the person tasked with deactivating it was one of the thousands of employees Musk dismissed in recent weeks.
In the end, though, the reality is that being required to have your hands on the wheel and eyes on the road at all times means that FSD is not, by definition, an autonomous or self-driving experience.
I’ve separately spent lots of time taking rides around San Francisco in Waymo vehicles, sitting in the back behind an empty front seat. Those rides can be jerky, at times tense and not for the faint of heart. But these vehicles actually are capable of taking passengers around for short journeys fully autonomously. I’ve yet to experience a disengagement in which Waymo has needed to come to the rescue for a car I’ve been in.
The experience with FSD did make me feel a part of a community that’s contributing to the development of the technology. Each time the car disengaged, you have the option to press the voice button and send a message to Tesla about what happened. Other users encouraged me to do so and regularly engaged with me about how it was going.
Musk has increasingly been posting on X, his social media platform, about forthcoming software updates and the enhancements they’ll deliver. My sources at Tesla are similarly encouraged by the neural networks in development to support a viable robotaxi. But Tesla owners paying for FSD are unable to see or experience that here and now.
“Try out Tesla supervised full self-driving and you’ll be amazed!” Musk wrote in one of his posts as my trial was coming to an end. After trying it out, I have a hard time seeing how Tesla will make the jump from a consumer car you own and must supervise at all times, to a robotaxi that entirely drives itself. Being part of the experiment was interesting, but not all that comforting.
I like driving myself, after all. We humans have an innate preference to be in control of our destiny. And as Musk himself once said, we’re underrated.
See original article by Edward Ludlow at Bloomberg
