Reality is Hitting Tesla
Even the bulls are beginning to realise that the emperor is naked.
In 2019, Musk promised that millions of driverless Tesla robotaxis would be on the road “next year”. For the past six years, Musk has been repeating this same promise like a drug-addled parrot in a painfully obvious attempt to pump up Tesla’s volatile yet stagnant stock price. Over the weekend, Tesla finally and tentatively rolled out its first-ever robotaxi service. But it’s a limp impression of what was promised and lays all of Tesla’s self-driving flaws bare.
There are more than a few elephants in the room when it comes to this “launch”.
Firstly, the launch is “invite only”, which isn’t surprising, as there are only a handful of vehicles included in this fleet. But this is far from the rapid roll-out Musk has been hinting at for years. However, because the only people invited seem to be Tesla employees, I’m not sure how this qualifies as a “launch”. Particularly because Tesla has been testing Model Y robotaxis with Tesla employees on the streets for months now.
These vehicles also aren’t the promised “Cybercab” and are instead standard Model Ys with the word “Robotaxi” written on the side in the same font as the Cybertruck logo. Why does this matter? The current FSD software and hardware have proven unable to deliver safe self-driving. So, many hoped that the Cybercab would introduce a far more substantial and robust system. These bog-standard Model Y robotaxis show that this isn’t the case. Musk is still attempting to get the highly flawed vision-only FSD system (read more here) to function correctly without attempting a significant overhaul.
This also means that the Robotaxi has the same limitations as FSD, such as having trouble accurately “seeing” the world around it, over-relying on its AI (which can never be fully accurate), and being unable to work in the rain, as most of the cameras it uses to sense the world around it don’t have wipers to clear their vision. In other words, harsh lighting conditions can cause it to veer off the road, and even a little rain can completely brick these robotaxis.
FSD’s limitations also explain why they opted for the Model Y rather than a Cybercab. Each Robotaxi currently has a safety driver in the passenger seat, ready to intervene at a moment’s notice. If they used the two-seater Cybercab, that means there would only be one free seat per ride. What’s more, the pedal-less and steering wheel-less design of the Cybercab would severely limit a safety driver’s ability to intervene or fully take over from FSD if it fails or shuts off due to a light drizzle.
Compare this to Waymo. They used safety drivers for over a decade before transitioning to a truly operatorless system in 2020. That makes it sound like Tesla using safety drivers isn’t a huge issue. However, Tesla launched its self-driving program over a decade ago as a driver-assist system in its consumer cars (in other words, self-driving with a safety driver). As a result, they have significantly more driving data than Waymo, and if they progressed at the same rate as Waymo, they shouldn’t need a safety driver at all by now!
Nonetheless, I am glad that Tesla has opted for a safety driver, as the statistics regarding Tesla’s current FSD system are damning. Elon Musk has claimed FSD has significantly improved and only needs intervention every 10,000 miles. However, third-party testing of the most current version revealed that it required intervention every 31 miles. In other words, Tesla’s FSD is 99.87% less safe than was previously claimed. This is what happens when you use a vision-only approach to self-driving and don’t incorporate a redundant sensor suite. By comparison, Waymo, which utilises an extensive sensor suite with multiple levels of redundancy, requires intervention only every 17,000 miles, and it achieved this milestone back in 2023. Oh, and this also means it can function in the rain.
It’s no wonder the public is incredibly reluctant to trust FSD. The Electric Vehicle Intelligence Report found that 60% considered FSD “unsafe”, 77% were unwilling to use FSD, and 48% believed FSD should be illegal. With this context, Tesla might need to supply an in-person safety driver just to get people, even Tesla employees, to use the service.
And it isn’t just the public or sceptics like me that are worried. Piper Sandler analyst Alexander Potter has openly stated that he doesn’t think FSD can support truly autonomous vehicles without human control. Piper Sandler is a Tesla bull and still claims Tesla is undervalued, which is about as far from being a sceptic as you can get, yet the company are still concerned about FSD. This adds additional context to how dodgy it is that Tesla isn’t using a significantly overhauled version of FSD for the robotaxi service and is instead just using bog-standard Model Ys.
I can’t exaggerate how hollow and pathetic an attempt this is to live up to the hype and how little it will actually help Tesla.
Waymo currently has around 1,500 fully autonomous vehicles and conducts around 250,000 rides per week. Due to their surprisingly high safety rate, they are now able to dramatically expand into new cities, so expect these numbers to soar. Yet, despite Waymo rides costing more than an Uber, they are still far from profitable and rely heavily on their parent company, Google, for funding.
By comparison, Tesla, which has been in the self-driving robotaxi game for roughly the same amount of time as Waymo, only has a handful of robotaxis, and the system they are running is so flawed that they still require safety drivers. Even worse with Musk pigheadedly persevering with the current flawed FSD system (which is understandable, as overhauling the system would cost billions of dollars), Tesla is unlikely to pass the safety levels required for robotaxis to expand past Austin. And on top of that, with Tesla being saddled with the cost of a safety driver and the higher costs of their AI-heavy approach, their operational costs will be higher than those of the unprofitable Waymo.
All this launch has done so far is prove that Tesla is lightyears behind in the robotaxi race, not just in scale but in technology, and has backed itself into a corner. It demonstrated that the vision-only, AI-heavy route Musk forced Tesla to take (read more here) is inefficient and incapable of achieving safe, fully autonomous driving and that they have no solution to address this issue. Yes, it is still technically “early days”, but also, it isn’t. Tesla has been gearing up and driving hype for this launch for literally a decade. This doesn’t feel like a well-planned, well-thought-out execution. It feels like after years of bluffing, Musk’s hand has been forced by a failing stock price and investor pressure, only for Tesla to unveil that they have half-arsed their homework.
Thanks for reading! Don’t forget to check out my YouTube channel for more from me, or Subscribe. Oh, and don’t forget to hit the share button below to get the word out!
Sources: Forbes, The Independent, Reuters, The Street, The Dawn Project, Will Lockett, Will Lockett
The NAZI muskrat is incompetent and a liar as well! Gee, just like a former Orange friend!!!
Reports of traffic violations on Day 1 won't help.