 |
| Tesla FSD Warning |
My wife and I have been regularly using Tesla Full Self Drive (Supervised FSD) software for several months now. It is a subscription costing us $108.16, including tax, per vehicle. Our experience has generally been positive. We set our destination in the navigation system and let FSD drive us. It pulls out of our driveway, takes us wherever we want to go, and parks on the street, in a parking lot, or in a garage. On occasion, though, we have had to take control when we felt it was necessary, often in response to Tesla warning to “take over immediately”.
 |
| Tesla Sensors. Source: Car and Driver Magazine |
Both of our Tesla Model Y vehicles use hardware version 3 (HW3). In addition to a computer, HW3 includes 7 cameras and four ultrasonic sensors. We get frequent Tesla software updates, downloaded over the air. As of today, we're on software version 2026.21.6, including FSD v14.3.7. The latest hardware version on new Tesla vehicles is HW4. Tesla recently outlined details on Full Self-Driving (FSD) v15, describing it as a major "step-change" in capability featuring a 10-billion parameter neural network. Tesla confirmed that current Hardware 4 (AI4) computers will natively support v15 and unsupervised features, targeting a late 2026 to early 2027 rollout window.
 |
| Tesla FSD Speed Modes. Source: Jowua |
Tesla FSD allows you to choose a driving mode. The choices include Sloth (least aggressive), Chill, Standard, Hurry and Mad Max (most aggressive). Here are more detail on each speed mode:
Sloth: The most cautious setting. It stays strictly at or slightly below the speed limit, favors slow right lanes on highways, and avoids proactive passing.
Chill: A relaxed, smooth profile. It generally targets the exact speed limit and makes fewer or less urgent lane changes.
Standard: The balanced, default setup. It typically drives roughly 5 to 7 mph over the speed limit (traffic permitting) and maintains an average commuting pace.
Hurry: A faster profile designed for quicker travel times. It allows higher offsets above the speed limit and executes more assertive lane changes to pass slower vehicles.
Mad Max: The most assertive profile, built for heavy gridlock or aggressive urban environments (like dense city traffic). It proactively navigates gaps, moves swiftly around traffic flow, and executes maneuvers with higher urgency.
We drive mostly in Standard or Hurry modes. The best part my wife likes is how smoothly it changes lanes even in high traffic conditions. One particular place where it has eliminated her stress is on Oakland Maze, where her route requires 5 quick lane changes to go from one freeway to another.
 |
| Waymo Robotaxi With LiDAR mounted on top |
There have been several instances where we have had to intervene to take control of the vehicle. These include loss of control when the front camera is blinded by direct sun at certain times of the day. This problem would probably not occur if Tesla vehicles were equipped with a LiDAR scanner that uses laser light pulses to measure distances and create accurate, 3D digital models of objects and environments. Robotaxis like Google's Waymo and Amazon's Zoox use LiDAR devices to achieve fully autonomous driving. However, like the rest of Tesla's vehicle lineup and Full Self-Driving stack, the recently announced Cybercab robotaxi also relies entirely on a vision-based camera system and onboard AI to interpret its surroundings.
In other instances, our Tesla car ran a stop sign that was blocked by a parked truck. It failed to read "STOP" painted on the road surface. At a local Safeway supermarket, it attempted to park in a shopping cart corral. A couple of times after pulling out of our driveway, it took a wrong turn but found its way back to the correct route after taking a detour.
Rapid development and deployment of AI applications like Tesla full self drive and various robotaxi services, including Waymo and Zoox, have broad social and economic implications. It is likely to make transportation safer, but it also threatens the livelihoods of tens of millions of taxi and truck drivers around the world. There have already been protests by drivers' unions in many major cities like London and New York.
Related Links:
Haq's Musings
South Asia Investor Review
My Family's Contribution to Climate Action
Pakistan Joins China in World AI Cooperation Organization
Pakistani-Americans Largest Foreign-Born Muslim Group in Silicon Valley
A Personal Story: When My Heart Stopped in San Francisco
In Memory of My Loving Mother Rafiqa Khatoon
Silicon Valley Pakistani-Americans
Pakistani-American Leads Silicon Valley's Top Incubator
Silicon Valley Pakistanis Enabling 2nd Machine Revolution
Pakistan EV Launches to Accelerate Energy Transition
Pakistani-American Ashar Aziz's Fire-eye Goes Public
Pakistani-American Doctor Honored By President Obama
Pakistani-American's Game-Changing Vision
Minorities Are Majority in Silicon Valley
Over A Million Pakistani University Students Enrolled in STEM Fields
I use the self driving option on a daily basis!
ReplyDeleteI drove all the way to Los Angeles from Danville and to my surprise it took the exit automatically at a random place and drove me to the charging station while selecting a slot to charge by backing and parking it. WOW👌
Right by your rear-view mirror, there is a camera that monitors your eyes to make sure that you're watching the road. Failure to do so, the warning comes to instruct you to take control.
I have noticed the camera on other vehicles also. For example in Lexus if you are driving in cruse mode and your eyes start wondering or are blocked then you get a message on the screen.
DeleteZamir
I enjoyed reading your experience on self driving Teslas.
ReplyDeleteVery interesting read.
ReplyDeleteI had similar experience with FSD about 4 months ago in Seattle. We inputted an address and after a while we changed the destination address. The system was not confident to take the three points turn and asked for human take over.
ReplyDeleteThank you for sharing, Riaz Bhai. I use FSD 99% of the time. I have a phobia of highway driving, and with FSD, I can go on the freeway myself. FSD has been a life-changer for me.
ReplyDeleteWedbush's Dan Ives and Requisite Capital's Bryn Talkington see Tesla Inc. (NASDAQ:TSLA) evolving beyond autos, with AI, autonomy, and robotics driving its long-term story while near-term execution remains under scrutiny.
ReplyDeletehttps://www.benzinga.com/analyst-stock-ratings/analyst-color/26/04/51992122/why-wall-street-is-betting-on-teslas-ai-future
He pointed to autonomous driving, robotaxis, and robotics as the core drivers of future value, noting that expansion into more cities, such as Houston and Dallas, will signal progress toward a broader network.
Ives added that achieving the next leg of valuation growth depends on scaling these AI-driven businesses, including projects like Optimus and Cybercab.
Earnings Still Matter As A Credibility Test
Talkington told CNBC that investors still need clarity on fundamentals, calling the earnings report a credibility test for a company in transition.
She emphasized the importance of margins and capital spending, questioning whether investments are focused on vehicles or on newer initiatives such as advanced manufacturing facilities.
Talkington also described Tesla as a "physical AI company," highlighting its advantage from real-world driving data compared with rivals relying primarily on external technology.
Execution, Timelines, And Stock Outlook
Ives stressed that Tesla must show realistic timelines for scaling production and expanding its autonomous footprint, alongside signs of stabilizing demand in key markets like China and Europe.
He said broader rollout across multiple cities will be critical to building a global network.
Talkington remained positive in the long term but cautioned that progress could take several quarters, noting that the stock faces technical resistance in the near term, even if stronger traction could eventually push it higher.
On Wednesday, Tesla reported mixed first-quarter results while highlighting progress in AI, autonomy, and future production plans.
Earnings Mixed as Revenue Misses, EPS Beats
Tesla posted first-quarter revenue of $22.71 billion, up 16% year-over-year but missing estimates, while adjusted earnings of 41 cents per share beat expectations. Automotive revenue rose 16% to $16.23 billion.
AI, Robotaxis and Production Outlook in Focus
The company pointed to growth in AI and autonomy, with robotaxi paid miles nearly doubling sequentially and FSD subscriptions reaching 1.28 million, up 51% year-over-year.
Tesla also expanded robotaxi rollouts to more cities and said Cybercab could eventually replace the Model Y as its highest-volume vehicle.
Looking ahead, Tesla expects volume production of the Cybercab and Semi this year and plans to begin operations at its first large-scale Optimus factory in the second quarter.
"This problem would probably not occur if Tesla vehicles were equipped with a LiDAR scanner that uses laser light pulses to measure distances and create accurate, 3D digital models of objects and environments. Robotaxis like Google's Waymo and Amazon's Zoox use LiDAR devices to achieve fully autonomous driving. However, like the rest of Tesla's vehicle lineup and Full Self-Driving stack, the recently announced Cybercab robotaxi also relies entirely on a vision-based camera system and onboard AI to interpret its surroundings. "
ReplyDeleteThere was an interesting BI article where Tesla Ashok Elluswami (Indian origin s/w head of tesla) managed this difficult feat
Zen: "There was an interesting BI article where Tesla Ashok Elluswami (Indian origin s/w head of tesla) managed this difficult feat"
ReplyDeleteWhat Ashok has accomplished with Tesla FSD so far is indeed a "difficult feat". It's now the key differentiator between Tesla and its EV competition.
What remains to be seen is the results of Tesla's "vision only" Cybercab safely delivering full autonomy required for robotaxis like Waymo and Zoox.
Tesla Fan Describes Riding the Cybercab for the First Time - Business Insider
ReplyDeletehttps://www.businessinsider.com/tesla-elon-musk-fan-robotaxi-event-experiences-cybercab-2024-10
Elon Musk showed off Tesla's robotaxi on Thursday night — and one of the first people to ride it told Business Insider what it was like to take a trip in the so-called Cybercab.
Tesla fan and investor Matthew Donegan-Ryan attended the glitzy launch on a film lot at the Warner Bros. Studio in Burbank, California, and said he was impressed by the ease with which the Cybercab dealt with obstacles.
He described it as "much smoother" and less jerky than the latest version of Tesla's Full Self-Driving assisted driving software.
The Cybercab, Tesla's first robotaxi, was revealed alongside a 20-person Robovan. It will have no steering wheel or pedals and will charge via wireless inductive charging.
Musk heralded the futuristic vehicle as a step toward an "age of abundance," but the launch left Tesla facing questions over regulatory scrutiny and how exactly its attempt to challenge Uber with a ride-hailing business will work.
Cybercab first impressions
Donegan-Ryan said the Cybercab was smaller than Tesla's Model 3 and had three buttons inside — one to turn on the inside lights and two to raise the door windows.
The only thing on the dashboard was a giant screen, on which attendees could select one of several destinations within Warner Bros. lot.
When the ride began, the Cybercab frequently encountered obstacles like cyclists and bystanders crossing the road, but Donegan-Ryan said the vehicle dealt with them smoothly.
"I was pleasantly surprised," he said. "The cars reacted flawlessly to them, and all took slightly different paths. It was kind of like watching a Formula 1 race."
Donegan-Ryan was also surprised by how few cameras the Cybercab had.
He said the robotaxi, which depends on computer vision to run its self-driving tech, appeared to have the same camera layout as a Model Y or 3. He said that he only saw one additional camera on the front bumper.
The Cybercab bears a stylistic resemblance to another Tesla vehicle, the Cybertruck, and Donegan-Ryan said that similarity was particularly apparent in person.
"All the Cybercabs that were on display were wrapped in a titanium stealth wrap. If you felt the edges and the curves, it felt just like a Cybertruck," he said.
A party atmosphere
The event was a glitzy affair, with drone light shows, pumping music, and Tesla's humanoid Optimus robots serving drinks. It's unclear whether these robots were human-assisted. When contacted by BI, Tesla did not respond to a request for comment on this point.
Donegan-Ryan said it had a festival atmosphere.
"There were many staff at the event, and they had a real sense of euphoria. They were all universally excited and pumped up," he said.
Staff were apparently mingling freely with the crowd. Donegan-Ryan told Business Insider that he spoke to Tesla's chief designer, Franz von Holzhausen, and its VP, Lars Moravy — and that both told him Tesla had no plans to release a version of the Cybercab with a steering wheel.
Cybercab's mixed reception
Ultimately, Donegan-Ryan came away impressed. He told BI that, while he wouldn't buy a two-seater Cybercab as he had three kids, he would consider one for his parents.
During the unveiling, Musk also said that the company intends to allow Tesla vehicle owners to rent Cybercabs out when they're not using them to be part of Tesla's ride-hailing network.
Donegan-Ryan said he might consider owning a Cybercab "to operate it in the fleet, as a revenue generator."
The unveiling of Tesla's Robovan and Cybercab, which Musk said would cost less than $30,000 and go into production before 2027, failed to impress everyone.
Some Wall Street analysts expressed disquiet over the lack of details on Tesla's business model and the absence of the more affordable EV the company is planning to ship next year, and Tesla's shares slumped as much as 10% in early trading on Friday.
Elon Musk’s robotaxi heterodoxy gets put to the test
ReplyDeleteThe occasional trillionaire is betting his company on a stripped down, camera-only, steering-wheel-free two-seater. How safe is safe enough?
https://www.theverge.com/transportation/987901/tesla-cybercab-launch-elon-musk-robotaxi-camera-lidar
Believe it or not, the Tesla Cybercab is coming.
Or will it? Musk has bet his company — and his trillion-dollar pay package — on a stripped down, barebones autonomous driving system that eschews multiple sensors used by all of his rivals in favor of a camera-only approach. The Cybercab, as a purpose-built autonomous vehicle without traditional controls, embodies this bet. Musk believes he can forge his own path to robotaxi success — and the Cybercab is his gilded vehicle of choice. We’ll soon know if he’s in for a crash.
The Cybercab will be using a modified version of Tesla’s Full Self-Driving (FSD). For the company’s millions of vehicle owners, FSD remains a Level 2 driver assist feature that requires human supervision. In the Cybercab, it will be monitored by remote operators who theoretically can step in if there’s a problem.
———
The light beam sensor that practically everyone views as an essential ingredient for self-driving cars is “a fool’s errand,” Musk said at the event in 2019. And “anyone relying on lidar is doomed. Doomed. Expensive sensors that are unnecessary. It’s like having a whole bunch of expensive appendices... you’ll see.” The crux of his argument is that humans use their own eyes to make driving decisions, so why not autonomous vehicles?
Tesla once relied on radar, in addition to cameras, as part of its perception system, only for Musk to do away with the sensor in 2022. The company’s own engineers tried to persuade Musk to keep radar. They worried that without radar, Tesla’s self-driving system would be susceptible to basic perception errors if cameras were obscured by rain or sunlight — which could lead to crashes. They were rebuffed.
The reason everyone else uses lidar underscores the limitations of Musk’s camera-only approach. Cameras only provide flat, two-dimensional images, requiring Tesla to use computer vision and neural networks to approximate depth and distance. A camera-only system leaves itself vulnerable to sunlight glare, while a multi-sensor system has additional data in case its perception is blinded. And cameras struggle in adverse weather, like snow or fog, while lidar and radar can often pierce through these conditions.
Few of Musk’s predictions have come true, including his condemnation of lidar. Waymo, which uses lidar in addition to cameras and radar, is operating a fleet of 4,000 fully driverless vehicles in up to 14 cities across the US. This is the largest, more fully realized deployment of self-driving cars in the US — and possibly the world. Lidar hasn’t slowed Waymo’s progress; in fact, its engineers strongly believe the sensor helps accelerate it.
In a recent talk at Y Combinator’s Startup School, Waymo co-CEO Dmitri Dolgov agreed that cameras have gotten good enough to closely approximate human driving. But if the goal is to exceed humans and build a system that truly excels at safety, then cameras alone aren’t good enough. If you’re targeting full autonomy and superhuman performance, Dolgov said, “you find that weak sensing just leads to a safety curve that flattens out way too early.”
But Musk claims that a camera-only system is more affordable to mass produce, and thus will enable Tesla to deploy Cybercabs more quickly than Waymo. There’s no arguing that the Cybercab is cheaper to produce than Waymo’s sensor-laden robotaxis, but the lack of lidar isn’t the only reason. Tesla’s also saving money by leaving out traditional controls like steering wheels, pedals, and mirrors. The company argues this helps create more passenger space, but the hope is that it will also make the vehicle more profitable to manufacture.