Why is it so critical to Tesla’s long-term strategy?
Aug 5th 2024 at 8:37AM
·
Elon Musk has been talking about Dojo, the AI supercomputer that will serve as the foundation for Tesla's AI goals, for years. Musk believes it is so critical that the AI team at the business will "double down" on Dojo as they prepare to unveil the robotaxi in October.
However, what is Dojo exactly? And why is Tesla's long-term plan so dependent on it?
To put it briefly, Dojo is Tesla's specially constructed supercomputer, which it uses to train its "Full Self-Driving" neural networks. Developing Dojo further aligns with Tesla's objective of achieving complete self-driving and launching a robotaxi. About two million Tesla cars currently have FSD, which can handle some automated driving functions but still needs a person to pay close attention behind the wheel.
Tesla's Dojo backstory
Tesla should be more than just an automobile, or even a supplier of solar panels and energy storage devices, according to Musk. Rather, his goal is for Tesla to become an artificial intelligence business that has figured out how to make automobiles drive themselves by imitating human vision.
In order to sense the environment, the majority of other businesses developing autonomous vehicle technology use a combination of sensors, including as lidar, radar, and cameras, in addition to high-definition maps for vehicle localization. Tesla thinks it can accomplish fully autonomous driving by using just cameras to gather visual data, processing that data with cutting-edge neural networks to determine the best course of action for the vehicle.
At the first AI Day for the manufacturer, Andrej Karpathy, the former head of AI at Tesla, stated
What is a supercomputer?
Dojo is Tesla’s supercomputer system that’s designed to function as a
training ground for AI, specifically FSD. The name is a nod to the space where
martial arts are practiced.
A supercomputer is made up of thousands of smaller computers called
nodes. Each of those nodes has its own CPU (central processing unit) and GPU
(graphics processing unit). The former handles overall management of the node,
and the latter does the complex stuff, like splitting tasks into multiple parts
and working on them simultaneously. GPUs are essential for machine learning
operations like those that power FSD training in simulation. They also power large
language models, which is why the rise of generative AI has made Nvidia the
most valuable company on the planet.
Even Tesla buys Nvidia GPUs to train its AI (more on that later).
Why does Tesla need
a supercomputer?
Tesla’s vision-only approach is the main reason Tesla needs a
supercomputer. The neural networks behind FSD are trained on vast amounts of
driving data to recognize and classify objects around the vehicle and then make
driving decisions. That means that when FSD is engaged, the neural nets have to
collect and process visual data continuously at speeds that match the depth and
velocity recognition capabilities of a human.
In other words, Tesla means to create a digital duplicate of the human
visual cortex and brain function.
To get there, Tesla needs to store and process all the video data
collected from its cars around the world and run millions of simulations to
train its model on the data.
Dojo pics pic.twitter.com/Lu8YiZXo8c
— Elon Musk (@elonmusk) July 23, 2024
Tesla doesn't want to put all of its eggs in one basket, in part because Nvidia chips are pricey, even if it looks that the company depends on Nvidia to power its existing Dojo training computer. Additionally, Tesla aims to improve things by boosting bandwidth and lowering latency. The automaker's AI division made the decision to develop a proprietary hardware program with the goal of training AI models more effectively than those trained by conventional systems.
Tell me more about these chips
Tesla and Apple share the belief that software and hardware ought to be created with each other in mind. For this reason, Tesla is developing its own processors to power Dojo instead of relying on the GPU technology that is currently in use.
On AI Day in 2021, Tesla debuted the D1 chip, a silicon square the size of a palm. Production of the D1 chip started at least in May of this year. Seven nanometer semiconductor nodes are used in the chip production process by Taiwan Semiconductor production Company (TSMC). According to Tesla, the D1 has a massive die size of 645 millimeters squared and 50 billion transistors. All of this to say, the D1 is expected to be incredibly strong and effective, capable of handling complicated
What does Dojo mean for Tesla?
Tesla's humanoid robot Optimus Prime II at WAIC in Shanghai, China, on July 7, 2024. Image Credits: Costfoto/NurPhoto via Getty Images)
As it and TSMC ramp up chip production, Tesla may eventually be able to rapidly and affordably add massive quantities of processing power to AI training programs thanks to its takeover of chip manufacture.
This implies that Tesla might not need to depend on Nvidia's chips in the future, which are getting more and more expensive and difficult to obtain.
Musk claimed that it is frequently difficult to obtain Nvidia hardware due to the tremendous demand for the technology during Tesla's second-quarter earnings call. "I think this requires that we put a lot more effort on Dojo in order to ensure that we can actually get steady GPUs when we want them," he stated, expressing his anxiety.
How far along is Dojo?
Reuters reported last year that Tesla
began production on Dojo in July 2023, but a June 2023 post from Musk suggested that
Dojo had been “online and running useful tasks for a few months.”
Around the same time, Tesla said it expected Dojo to be one of the top
five most powerful supercomputers by February 2024 — a feat that has yet to be
publicly disclosed, leaving us doubtful that it has occurred.
The company also said it expects Dojo’s total compute to reach 100
exaflops in October 2024. (1 exaflop is equal to 1 quintillion computer
operations per second. To reach 100 exaflops and assuming that one D1 can
achieve 362 teraflops, Tesla would need more than 276,000 D1s, or around
320,500 Nvidia A100 GPUs.)
Tesla also pledged in January 2024 to spend $500 million to build a Dojo
supercomputer at its gigafactory in Buffalo, New York.
In May 2024, Musk noted that the rear portion of
Tesla’s Austin gigafactory will be reserved for a “super dense, water-cooled
supercomputer cluster.”
Just after Tesla’s second-quarter earnings call, Musk posted on X that the automaker’s AI team
is using Tesla HW4 AI computer (renamed AI4), which is the hardware that lives
on Tesla vehicles, in the training loop with Nvidia GPUs. He noted that the
breakdown is roughly 90,000 Nvidia H100s plus 40,000 AI4 computers.
“And Dojo 1 will have roughly 8k H100-equivalent of training online by
end of year,” he continued. “Not massive, but not trivial either.”
Elon
Musk's strategy to improve Tesla's artificial intelligence skills, with a
particular focus on improving the company's autonomous driving technologies,
heavily relies on the supercomputer Dojo. The following explains why Dojo is
essential to Tesla's long-term plan:
Improving the Capabilities of Autonomous Driving:
- Massive Data Processing: The sensors and cameras on Tesla cars gather
enormous volumes of data. Dojo is built to handle this data at previously
unheard-of speeds, enabling machine learning that is both quicker and more
precise.
Improving Models of AI:
Tesla's neural networks will be trained more effectively by the supercomputer, leading to improved autonomous driving algorithms. This will increase Tesla's likelihood of achieving full self-driving (FSD) capability.
2. Dependence on Hardware from Third Parties Reduced:
Combination Solution:For its computational requirements, Tesla currently
uses hardware from other parties. Creating Dojo internally
For more detail.
https://techcrunch.com/
.jpg)
إرسال تعليق