Nearly two years have passed since Tesla’s first Autonomy Day event, with CEO Elon Musk making numerous lofty predictions about the future of autonomous vehicles, including infamous allegation that the company would have “a million robots on the road” by the end of 2020. And now is the time Osa Deux.
This time the event name is AI Day and According to Muskthe “only goal” is to get robot and artificial intelligence experts to come to work at Tesla. The company is known for its high turnover, the latest is Jerome Guillen, a key executive who worked at Tesla for 10 years before resigning recently. Attracting and retaining talent, especially top names, has proven to be a challenge for the company.
The event is scheduled to begin on August 19 at 5:00 PM PT / 8:00 PM at ET Tesla’s headquarters in Palo Alto, California. By invitation obtained Electrek, it includes “Elon’s keynote speech, Tesla engineers’ hardware and software demos, Model S Plaid test runs, and more.” Just like battery day, the event will be broadcast live on Tesla ‘s website, allowing investors and the media, as well as the company’ s many fans, to see what is being developed up close.
Musk and other top executives at the company are expected to deliver updates on the introduction of Tesla’s “Full Self-Driving” (FSD) beta 9, which began reaching more customers this summer. We may also receive information about the Teslan Dojo supercomputer, its neural network training, and the production of FSD computer circuits. The invitation also says that “internally, let’s see what Tesla’s artificial intelligence is next outside our car fleet.”
Let’s start with what we know and try to speculate on the future.
Introduction of FSD
The big news of Tesla’s first day of autonomy was the introduction the company’s first Computer Circuit, A 260 square millimeter barrel that Musk described as “the best chip in the world.” Originally, Musk had argued that Tesla cars don’t need hardware upgrades, just software, on the road to full independence. It turned out that was not quite the case; they would need this new chip – actually two – to be able to drive themselves.
A lot has happened between the 2019 event and the present. Last month, Tesla began delivering wireless software updates for FSD beta v9, its long-awaited, definitely not independent, but certainly an advanced driver assistance system. This means that Tesla owners who have purchased the FSD option (which now costs $ 10,000) could finally use many of Autopilot’s advanced driver assistance features on local, non-highway streets, such as Navigate on autopilot, Automatic lane change, AutoPark, Assemble, and traffic lights and stop.
The upgrade will not make Tesla’s cars completely independent, nor will it launch “a million self-driving cars” on the road, as Musk predicted. Tesla owners with full self-driving still need to pay attention to the road and keep their hands on the steering wheel. Some don’t, which can be tragic consequences.
FSD software, loved by fans and disliked by safety advocates, has received a lot of hot water from Tesla recently. In recently published emails Between Tesla and the California Automotive Plant, the company’s director of autopilot software made it clear that Musk’s comments (including his Tweets) don’t reflect the reality of what Tesla’s vehicles can actually do. And now Federal authorities are investigating Autopilot who want to know why Teslat with Autopilot constantly collide with emergency vehicles.
In addition to the introduction of FSD beta v9, Teslan has had to adapt to a global chip shortage. In a recent call for results, Musk said the company’s engineers were forced rewrite some of its software to receive alternative computer circuits. He also said that Tesla’s future growth depends on a rapid solution to the global semiconductor shortage.
Tesla uses chips to use everything in the airbag and vehicle seat belt control modules. It is unclear whether the shortage will affect FSD chips manufactured by Samsung. Musk and his cohort may give some idea of it during this week’s event.
Outside the car, Tesla uses a powerful supercomputer to train artificial intelligence software, which is then fed to its customers through wireless software updates. In 2019, Musk teased this “very powerful training computer,” which he called “Dojo.”
“Tesla is developing a [neural net] a training computer called Dojo to handle really large amounts of video data, he tweeted later. “It’s a beast!”
He also alluded to Dodo’s computing power and claimed that it was able to perform an exaFLOP operation or one quintillion (1018) floating point operation per second. It’s an incredible amount of power. “To answer what one exaFLOP computer system can do in just a second,” NetworkWorld wrote Last year, “you had to do one calculation every second for 31,688,765,000 years.”
By comparison, chip maker AMD and computer maker Cray are currently working with the U.S. Department of Energy to design the world’s fastest supercomputer with 1.5 exaFLOP processing power. Dubbed Frontier, AMD says the supercomputer has as much processing power as the next 160 fastest supercomputers combined.
When completed, the Dojo is expected to be one of the most powerful supercomputers on the planet. Instead of performing advanced calculations in areas such as nuclear and climate research, Tesla’s supercomputer uses a neural network to train its artificial intelligence software in self-driving cars. Eventually, Musk has said that Tesla will make the Dojo available to other companies that want to use it to train their neural networks.
Earlier this year, Andrej Karpathy, Tesla’s Director of Artificial Intelligence, gave a presentation at the 2021 Machine View on Seeing Computers and Identifying Patterns, during which he introduced more information about the Dojo and its neural network.
“Computer vision is for us the bread and power of what we do and what enables Autopilot,” Karpathy said. by Electrek. “And for this to work really well, we need to manage fleet data, train massive neural networks, and experiment a lot. So we put a lot of effort into computing.”
Earlier this month, Dennis Hong, founder of UCLA’s Robotics and Mechanics Laboratory, tweeted a picture of a computer circuit that many speculate is the hardware used by the Teslan Dojo.
But Hong is an interesting character for other reasons as well. He specializes in humanoid robots and participated in the DARPA Urban Challenge, which launched a self-propelled car race. (His team finished third.)
When asked on Twitter if his lab works with Tesla, Hong posted playful emoticons but otherwise declined to comment. We might learn more about how Hong’s work and Tesla’s hobbies intersect during AI Day.
Musk has said he wants Tesla to become more than just a car company. “I think in the long run, people think of Tesla as much as an AI robotics company like we do a car or energy company,” he said earlier this year.
A caveat to anyone tuning in to the AI Day live stream: take Musk’s predictions of upcoming achievements with huge salt. Issues addressed during the transaction are unlikely to have a measurable impact on the company’s business in the coming months.
Self-propelled cars are an incredibly difficult challenge. Even companies like Waymon are considered to have the best independent vehicle technology still struggling to get it right. Tesla is no different.
“The key question for investors is what is the latest timetable for achieving full independence,” Loene Funds CEO Gene Munster said in a note. “Despite Elon’s ambitious goal by the end of this year, our best guess is that 2025 will be the first level 4 public access to autonomy.”
The end of 2021 is already full of Tesla. The company must open factories in Texas and Germany. And it needs to increase production efficiency for the highly anticipated Cybertruck, which has been postponed until 2022. Full independence, as it stands, can be expected.