Connect with us

Investor's Corner

Why Tesla Autopilot will ultimately prove the self-driving industry leader

Source: Tesla

Published

on

Tesla took an early lead in the race to develop vehicle autonomy, and its Autopilot system remains the state of the art. However, the technology is advancing more slowly than the company predicted – Elon Musk promised a coast-to-coast driverless demo run for 2018, and we’re still waiting. Meanwhile, competitors are hard at work on their own autonomy tech – GM’s Super Cruise, is now available on the CT6 luxury sedan.

Is Tesla in danger of falling behind in the self-driving race? Trent Eady, writing in Medium, takes a detailed look at the company’s Autopilot technology, and argues that the California automaker will continue to set the pace.

Every Tesla vehicle produced since October 2016 is equipped with a hardware suite designed for Full Self-Driving, including cameras, radar, ultrasonic sensors and an upgradable onboard computer. Around 150,000 of these “Hardware 2” Teslas are currently on the road, and could theoretically be upgraded to self-driving vehicles via an over-the-air software update.

Above: In its current state, Tesla’s Autopilot requires a hands-on approach (Youtube: Tesla)

Tesla disagrees with most of the other players in the self-driving game on the subject of Lidar, a technology that calculates distances using pulses of infrared laser light. Waymo, Uber and others seem to regard lidar as a necessary component of any self-driving system. However, Tesla’s Hardware 2 sensor suite doesn’t include it, instead relying on radar and optical cameras.

Advertisement
-

Lidar’s strength is its high spatial precision – it can measure distances much more precisely than current camera technology can (Eady believes that better software could enable cameras to close the gap). Lidar’s weakness is that it functions poorly in bad weather. Heavy rain, snow or fog causes lidar’s laser pulses to refract and scatter. Radar works much better in challenging weather conditions.

According to Eady, the reason that Tesla eschews lidar may be the cost: “Autonomy-grade lidar is prohibitively expensive, so it’s not possible for Tesla to include it in its production cars. As far as I’m aware, no affordable autonomy-grade lidar product has yet been announced. It looks like that is still years away.”

If Elon Musk and his autonomy team are convinced that lidar isn’t necessary, why does everyone else seem so sure that it is? “Lidar has accrued an aura of magic in the popular imagination,” opines Mr. Eady. “It is easier to swallow the new and hard-to-believe idea of self-driving cars if you tell the story that they are largely enabled by a cool, futuristic laser technology…It is harder to swallow the idea that if you plug some regular ol’ cameras into a bunch of deep neural networks, somehow that makes a car capable of driving itself through complicated city streets.”

Those deep neural networks are the real reason that Eady believes Tesla will stay ahead of its competitors in the autonomy field. The flood of data that Tesla is gathering through the sensors of the 150,000 or so existing Hardware 2 vehicles “offers a scale of real-world testing and training that is new in the history of computer science.”

Competitor Waymo has a computer simulation that contains 25,000 virtual cars, and generates data from 8 million miles of simulated driving per day. Tesla’s real-world data is of course vastly more valuable than any simulation data could ever be, and the company uses it to feed deep neural networks, allowing it to continuously improve Autopilot’s capabilities.

A deep neural network is a type of computing system that’s loosely based on the way the human brain is organized (sounds like the kind of AI that Elon Musk is worried about, but we’ll have to trust that Tesla has this under control). Deep neural networks are good at modeling complex non-linear relationships. The more data that’s available to train the network, the better its performance will be.

Advertisement
-

“Deep neural networks started to gain popularity in 2012, after a deep neural network won the ImageNet Challenge, a computer vision contest focused on image classification,” Eady explains. “For the first time in 2015, a deep neural network slightly outperformed the human benchmark for the ImageNet Challenge…The fact that computers can outperform humans on even some visual tasks is exciting for anyone who wants computers to do things better than humans can. Things like driving.”

By the way, who was the human benchmark who was bested by a machine in the ImageNet Challenge? Andrej Karpathy, who is now Director of AI at Tesla.

===

Note: Article originally published on evannex.com by Charles Morris; Source: Medium

EVANNEX carries aftermarket accessories, parts, and gear for Tesla owners. Its blog is updated daily with Tesla news.

Advertisement
Comments

Elon Musk

X changed how everyone gets paid, and this lawsuit shows why

X sued a Bitcoin account network over fake payouts as its creator pay model shifts

Published

on

By

Elon Musk’s X has taken a Bitcoin-focused engagement ring to court, and the case doubles as a receipt for how differently the platform pays creators today. The company filed suit in the High Court of England and Wales against Vivek Kumar Sen and Zamyang Sherpa, alleging the pair ran six accounts, including @Vivek4real_, @Bitcoin_Teddy and @TrendingBitcoin, as one coordinated operation to fake the kind of engagement that used to translate directly into money.

According to the filing, first reported by Gizmodo, the accounts posted near identical “BREAKING” crypto headlines seconds apart, in one case 11 seconds, then had three more handles like, reply to and repost the material to manufacture what X called “a false appearance of genuine, human communication and interaction.” X says the scheme pulled in at least £207,384, about $278,000, and pegs its own investigation and remediation costs at another £75,000. The accounts were suspended August 18. X general counsel James Burnham announced the case on X last weekend, writing that the company “will act forcefully to protect our platform and the earnings of genuine creators.” Musk’s own reaction, posted shortly after, was three words: “Don’t mess with 𝕏.”

The timing lines up with a a recent update to how X pays its creators. The program these accounts allegedly gamed, Creator Revenue Sharing, launched in mid 2023 and paid out based on how much a post got engaged with. Originality was never part of the formula, which is exactly how the platform ended up flooded with recycled clips, copy pasted “BREAKING” posts and replies engineered purely to farm reactions from paying subscribers.

X tried patching the model more than once, including an April cut to aggregator payouts and a March regional weighting change that Musk personally paused hours after it was announced. X retired Creator Revenue Sharing for good on September 7 and opened its replacement, Original Content Rewards, the next day.

Advertisement
-

The new math is stricter. Payouts now come only from qualified impressions, meaning unique Home Timeline views from Premium subscribers where at least half the post is visible, and replies no longer count toward eligibility at all. Copied posts, reuploaded media and reposts without meaningful changes are explicitly excluded. Allegra Jacchia, senior product manager for Creators at SpaceXAI, which now runs X’s product and AI work following xAI’s acquisition of the platform, put it bluntly, saying the goal is to reward creators who bring original ideas and perspective, “not those who have become best at gaming the system.”

Read that way, the lawsuit isn’t really about six crypto accounts. It’s X putting a dollar figure on what the old incentive structure cost, then suing to collect it right as the new one goes live. For live updates on how the case and the new rewards program shake out, follow @Teslarati on X.

Advertisement
-
Continue Reading

Elon Musk

Why automakers keep turning down Elon Musk’s Tesla Full Self-Driving offer

Elon Musk confirms no automaker has ever accepted Tesla’s offer to license Full Self-Driving software.

Published

on

By

Elon Musk gave a brief answer on X Monday that confirmed that Tesla’s standing offer to license Full Self-Driving to other automakers still has zero takers. Sawyer Merritt wrote that “Tesla has for years openly invited other automakers to license FSD. None of them have accepted,” responding to a prediction from Boom Supersonic founder Blake Scholl that Tesla would eventually open FSD the way it opened its Supercharger network to rival brands. Musk’s reply to Merritt was one word: “Exactly.”

It is not the first time Musk has made this point. He said something similar in November, when he called legacy automakers reluctance to adopt FSD “crazy,” and Tesla has floated the offer publicly since at least 2021. Scholl’s prediction touches on something real. Once NACS became the de facto charging standard, adoption from Ford, GM, Rivian and others followed within about a year. FSD licensing was supposed to work the same way once Tesla built enough of a lead that switching made sense for everyone.

The case for licensing now is stronger than it was two years ago. Waymo and Zoox are logging hundreds of thousands of unsupervised autonomous miles, along with Tesla’s own Robotaxi fleet. Every automaker still selling driver assist systems that lag FSD has given the robotaxi conversation to Tesla, Waymo and Zoox by default. Licensing FSD would let a GM or a Ford compete on the same field without spending a decade and billions of dollars building a stack from scratch, the same argument Tesla made when it opened the Supercharger network to bring more EVs onto its chargers.

But FSD is not a connector standard. As one reply to Musk’s post pointed out, licensing FSD is not a software license the way NACS was a plug spec. It requires adopting Tesla’s eight camera layout and its onboard compute architecture, meaning a licensee’s cars would effectively become Tesla hardware wearing someone else’s badge. That is the visible obstacle. The less visible one is data. A licensed FSD stack would report back the same telemetry Tesla collects from its own fleet, giving Tesla a continuous read on how a competitor’s cars are actually driven, where they struggle, and how often drivers intervene. For an automaker trying to build its own autonomy program, or simply trying to keep its build quality and safety record private, handing Tesla that visibility could be a bigger cost than the hardware bill. It is the reason the Supercharger comparison only goes so far. Opening a charging plug cost Tesla very little. Opening FSD would cost a rival something it cannot get back.

Advertisement
-
Continue Reading

Elon Musk

Tesla eyes supply partners for Optimus mass production

Tesla certified three Chinese suppliers for Optimus mass production, signaling its robot timeline is accelerating.

Published

on

By

Concept rendering of Tesla Optimus in mass production

Tesla’s robotics team traveled to Ningbo, in China’s Zhejiang province, on September 16 and spent the following day auditing component suppliers for Optimus, according to a Bloomberg report cited by RobotAIGeek. The visit moved three manufacturers from provisional status to certified mass production partners: Tuopu Group, which handles actuators and chassis components, Ningbo Joyson Electronic, a sensor supplier, and Zhejiang Sanhua Intelligent Controls, which builds thermal management systems. All three already supply parts to Tesla’s electric vehicles, and the audit reportedly came with fresh orders that supply chain reports put at an initial batch of roughly 5,000 units.

Tuopu, Joyson, and Sanhua built their manufacturing base serving the automotive industry, where tolerances and volume requirements are already close to what a mass produced humanoid robot demands. Sanhua in particular has history here. Teslarati reported last October that the company had received a roughly $685 million order for linear actuators tied to Optimus, a volume industry watchers estimated could cover around 180,000 robots once production ramped.

Supply chain reports tied to this week’s audit put Tesla’s near term production goal at about 1,000 Optimus units a week by late September, rising to 2,000 to 2,500 units a week by the end of the year. That pace would put real weight behind the timeline Tesla has been building toward since May, when it wound down Model S and Model X production at Fremont to convert that floor space into a dedicated Optimus line targeting one million units annually. JPMorgan analysts who toured the factory in August confirmed the conversion took roughly four months, a pace Musk has called unprecedented for a facility that size.

New drone video shows Tesla’s Optimus Factory reaching a turning point

Fremont is only the first phase. A second, larger Optimus plant is rising at Gigafactory Texas, where drone footage shared by Joe Tegtmeyer last week showed the structural steel nearing completion on the north end of the building. Tesla has said that facility is meant to eventually support production of up to 10 million units a year, though volume output there is not expected before 2027.

Commercial sales of Optimus are still targeted for the second half of 2027, but production is expected to start well before then. JPMorgan analyst Rajat Gupta has said Tesla’s “Optimus Academy” program, which uses early units to collect real world training data inside Tesla’s own facilities, is expected to be running later this year. Bloomberg Intelligence analyst Ian Ma described the Ningbo audits as “a positive commercialization signal for China’s humanoid supply chain,” noting that sentiment could improve further if the visit leads to confirmed supplier nominations and larger orders. The Solactive China Humanoid Robotics Index rose about 1.4% on the news, though it remains down roughly 30% for the year.

Advertisement
-
Continue Reading