News
Tesla is patenting a clever way to train Autopilot with augmented camera images
Tesla is currently tackling what could only be described as its biggest challenge to date. In his Master Plan, Part Deux, CEO Elon Musk envisioned a fleet of zero-emissions vehicles that are capable of driving on their own. Tesla has made steps towards this goal with improvements and refinements to its Autopilot and Full Self-Driving suites, but a lot of work remains to be done.
As noted by Tesla during its Autonomy Day presentation last year, attaining Full Self-Driving is largely a matter of training the neural networks used by the company. Tesla adopts what could be described as a somewhat organic approach for autonomy, with the company using a system that is centered on cameras and artificial intelligence — the equivalent of a human primarily using the eyes and brain to drive.
Tesla’s camera-centric approach may be quite controversial due to Elon Musk’s strong stance against LiDAR, but it is gaining ground, with other autonomous vehicle companies such as MobilEye developing FSD systems that rely primarily on visual data and a trained neural network. This approach does come with its challenges, as training neural networks requires tons of data. Tesla emphasized this point as much during its Autonomy Day presentation.
With this in mind, it is pertinent for the electric car maker to train its neural networks in a way that is as efficient as possible with zero compromises. To help accomplish this, Tesla seems to be looking into the utilization of augmented data, as described in a recently published patent titled “Systems and Methods for Training Machine Models with Augmented Data.”
Teslas are equipped with a suite of cameras that provide 360-degree visual coverage for the vehicle. In the patent’s description, Tesla noted that images used for neural network training are usually captured by various sensors, which, at times, have different characteristics. An example of this may lie in a Tesla’s three forward-facing cameras, each of which has a different field of view and range as the other two.
Tesla’s recent patent describes a system that allows the company to process these images in an optimized manner. Part of how this is done is through augmentation, which opens the doors to flexible and widespread neural network training, even when it involves vehicles equipped with differently-specced cameras. The electric car maker describes this process as such:
“Augmentation may provide generalization and greater robustness to the model prediction, particularly when images are clouded, occluded, or otherwise do not provide clear views of the detectable objects. These approaches may be particularly useful for object detection and in autonomous vehicles. This approach may also be beneficial for other situations in which the same camera configurations may be deployed to many devices. Since these devices may have a consistent set of sensors in a consistent orientation, the training data may be collected with a given configuration, a model may be trained with augmented data from the collected training data, and the trained model may be deployed to devices having the same configuration.”
Among the most notable aspects of Tesla’s recent patent is the use of “cutouts,” which allow Tesla’s neural networks to be trained using an optimized set of images. This was something that was discussed by former Tesla Autopilot engineer Eshak Mir in a Third Row Podcast interview, where he hinted at a system adopted in the electric car maker’s ongoing Autopilot rewrite that helped lay out “all the camera images” from a vehicle “into one view.” Such a process has the potential to help Tesla with 3D labeling, especially since the images used for neural network training are stitched together. Tesla’s patent seems to reference a system that is very similar to that described by the former Autopilot engineer.
“As a further example, the images may be augmented with a“cutout” function that removes a portion of the original image. The removed portion of the image may then be replaced with other image content, such as a specified color, blur, noise, or from another image. The number, size, region, and replacement content for cutouts may be varied and may be based on the label of the image (e.g., the region of interest in the image, or a bounding box for an object).”
Tesla is aiming to release a feature-complete version of its Full Self-Driving suite as soon as possible. Elon Musk remains optimistic about this, despite the company missing its initial timeline that was set at the end of 2019. That being said, Elon Musk did mention previously that Tesla is working on a foundational rewrite of Autopilot. In a tweet early last month, Musk stated that an essential part of the rewrite involves work on Autopilot’s core foundation code and 3D labeling. Once done, the CEO indicated that additional functionalities could be rolled out quickly. This recent patent, if any, seems to give a glimpse at how these improvements are being done.
Elon Musk
Tesla Roadster owners to get “Flying Stunt Mode” and looping track access
Musk says Roadster owners can command the unveiling stunts autonomously, with a Texas looping track.
Elon Musk says owners of the new Tesla Roadster will be able to ask the car to perform the stunts Tesla showed at its unveiling, with the car doing the driving itself. Musk made the comment Saturday in a reply on X to tech blogger Robert Scoble, and he added a promise of a looping track at Tesla’s Texas headquarters.
Scoble had argued that the person in the Roadster is not really driving it. In his description, the driver operates a computer that tells another computer what to do, with steering, braking and acceleration all handled by wire. Musk answered, “True. You will be able to ask the Roadster to perform any of the tricks shown autonomously, as it will be sentient. Also, we will have a wild looping track (crazier even than Hot Wheels) at our Texas HQ where you can test out your Roadster.”
@Scobleizer True.
You will be able to ask the Roadster to perform any of the tricks shown autonomously, as it will be sentient.
Also, we will have a wild looping track (crazier even than Hot Wheels) at our Texas HQ where you can test out your Roadster.
— Elon Musk (@elonmusk) October 17, 2026
At the October 15 event outside Waco, the car drove up a curved ramp until it was upside down, flew off on cold gas thrusters, flipped and landed. It also hovered over a Cybercab, Cybertruck, Semi, Model Y and Model 3 while moving. As we noted in our breakdown of the three Roadster configurations, Musk said at the event that the car “literally is just going to puff itself up” to perform the maneuvers whenever an owner wants. Saturday’s reply adds that the car would run them on its own.
Tesla has described some of the hardware that would make that possible. The Roadster uses steer by wire with a yoke, and its cold gas thrusters rely on electric actuators for small adjustments in flight. A fan system under the car creates suction, which Tesla says is strong enough to hold it upside down, a setup we covered in our look at the Roadster’s fan system and thrusters.
Musk did not say when an autonomous stunt feature might arrive, how it would work, or what he means by sentient. He also gave no size, location or opening date for the Texas track. Production is planned for Giga Texas in 2027 or 2028, and Tesla says the car is not approved for public road use. The Rocket version, the one built to fly, is invite only.
The comment follows a week of Musk teasing the event, including his claim that the demo would be a banger, which we covered before the stream. Tesla reports third quarter earnings on October 21, and shareholders can submit and vote on questions through the Say portal until 4 p.m. EDT on October 20, according to our earnings Q&A preview. A question about autonomous Roadster stunts and the Texas track would be a quick way to get specifics from management.
Elon Musk
How to use Grok Bot to block spam calls and texts
Musk told users to ask Grok Bot to scrub their phone numbers from spam lists.
Elon Musk has a new use case for Grok Bot, and it is the spam call. On Thursday he replied to an X user named Arash with a short instruction: “Ask your @Grok @Bot to remove your phone number from spam call & text lists!” Arash had said he got tired of the calls, had an AI bot find his number on listing sites and file removal requests, and packaged the process as a free skill for others.
The skill is called Spam-Call Scrubber, which is described to find a phone number on people search sites and data brokers, submits the opt outs, completes the confirmation emails those sites send back, and rechecks on a schedule the user picks. Using it requires the Grok Bot app. The pitch is time saved, since most broker opt outs sit behind separate forms on separate sites.
Ask your @Grok @Bot to remove your phone number from spam call & text lists!
— Elon Musk (@elonmusk) October 15, 2026
Musk said Grok Bot now gives each bot its own email address, and that it can detect scam emails. On Thursday he also said SpaceX would make users whole if the bot botches a purchase or leaks a card, then told users to shop away from Amazon after the retailer blocked the agent, as Teslarati reported. The terms of service we reviewed at the time put responsibility for what an agent does on the user.
For Tesla owners, Grok Bot is not just a phone app. Tesla brought Grok Bot into its vehicles on September 22, and we argued earlier this week that the Cybercab’s real advantage may be the errands it runs during a ride. Every Cybercab and the rest of the Model Y Robotaxi fleet has it enabled. The more routine tasks like this one become, the more a rider has reason to hand over something personal in the back seat.
The new Spam-Call Scrubber skill can be downloaded for free and works on sites that honor removal requests, while noting that listings can reappear, and illegal robocallers do not check any registry. Whether Spam-Call Scrubber cuts anyone’s call volume is something the first wave of users will find out, and Musk’s endorsement guarantees that wave is coming.
Investor's Corner
Tesla’s flying Roadster just revealed something bigger than the car
Analysts say the Roadster demo puts a Tesla and SpaceX merger back on the table.
The Roadster unveiling near Waco ended Thursday night, and by Friday morning much of the investor reaction had less to do with the car than with who owns the rocket technology inside it.
Wedbush analyst Dan Ives said the event “puts the Tesla/SpaceX relationship on the table in a tangible way,” pointing to the propulsion, pressure vessel and control hardware sitting on a Tesla product. Gene Munster of Deepwater Asset Management called the demo “a ridiculous display of engineering” with “no commercial value” beyond embarrassing traditional automakers, though he said it showed Musk wants the two companies together.
The more specific forecast came a day earlier from Barclays. Analyst Dan Levy told clients a deal could be announced before Tesla’s annual meeting, which must happen by December 6, and that an announcement in late October or early November could make shareholder approval harder. Barclays expects SpaceX to be the buyer in an all stock deal, with Tesla holders receiving SpaceX shares at a premium, based on market values of roughly $2.1 trillion for SpaceX and $1.5 trillion for Tesla. If nothing lands around the meeting, the bank still sees an announcement within 6 to 12 months.
The Roadster demo just added more weight to merger speculations. Tesla’s demo used cold gas thrusters and pressure tanks of the kind SpaceX uses on Falcon 9, and Musk described the car as a collaboration between Tesla and SpaceX engineers, as we mentioned in our breakdown of the three Roadster configurations.
Musk has not publicly addressed a deal since the event, as far as we have seen. His most direct answer remains the one he gave at the All In Summit, when he said that with so much collaboration, “who can imagine what action one might take” without ruling anything out. Ives has held to a 2027 timeline since his March note, and Musk pushed back on the China piece of the speculation when ARK floated a year end announcement in August.
The next test is Tesla’s third quarter earnings call on October 21. However, we recall in July that Musk told analysts, “we can’t talk about combining companies and that kind of thing on an earnings call,” so a clear answer wouldn’t be expected.
Tesla Roadster owners to get “Flying Stunt Mode” and looping track access
How to use Grok Bot to block spam calls and texts
Tesla’s flying Roadster just revealed something bigger than the car
SpaceX reveals how its 1 Million AI satellite network will work and prevent space collisions
Elon Musk’s Boring Company lands a new Middle East deal, and Nashville is about to get faster
