Whether in the past or in the future, the answers to many questions are hidden in time.
In 2012, the Tesla Model S was launched, in 2015, the Tesla Model X, in 2016, the Tesla Model 3, and in 2019, the Tesla Model Y.
Aside from the Cybertruck, which has encountered technical difficulties in body materials, Tesla's "entry-level family car" that has been brewing for many years should come out. However, with a tweet sent by Musk on the X platform, Tesla's new car lineup began to become confusing.
On April 6, Beijing time, Musk announced a brief news: Tesla will release a driverless taxi (Robotaxi) product on August 8. This news is not surprising, because according to Musk's original plan, the product was originally scheduled to be released in 2020, and this debut has been four years late.
But two days later, Reuters reported that Tesla had canceled plans to develop low-cost cars and instead continued to develop self-driving taxis on the same platform. In other words, the Robotaxi and the "entry-level family car" will be integrated into one product, and it is the mutual accommodation of the two that has led to the debut of both products later than expected.
Although Musk was quick to respond, bluntly saying that "Reuters is lying again", combined with Tesla's sales in the first quarter, it is difficult to say that this news is groundless.
In the first quarter of 2024, Tesla delivered 386,800 new vehicles, down 8.5% from the same period last year and well below Wall Street expectations. As a result, Tesla's stock price fell 4.9% on April 2. Three days later, Musk was also robbed of the title of the world's third-richest person by Meta CEO Mark Zuckerberg.
The combination of various headwinds has shown that Tesla needs a "thirst-quenching" product to stimulate the market, and Robotaxi seems to be becoming the answer.
In fact, Musk, as the CEO of a top technology company, is obviously not only ambitious in the field of automotive products, such as Robotaxi, recyclable rockets, Starlink, Mars colonization... These visions and products are significantly more compelling. In Musk's eyes, the future of human travel is a whole, and the current bottleneck period is inevitable and must be broken.
Interestingly, some of the technologies that have made Musk fanatical are strongly related to the automotive industry, such as self-driving, AI, and robotics that we are familiar with.
●The unspeakable nature of FSD
From the perspective of R&D investment, there is probably no CEO of an automaker in the world who is more obsessed with autonomous driving than Musk.
According to Tesla's 2023 financial report, the company's total revenue last year was about 692.2 billion yuan, of which R&D expenditure was about 28.4 billion yuan, which not only set a record in the company's history, but also exceeded the total R&D of many new domestic car companies.
We don't know how much of this contributed to autonomous driving, but Musk's attitude can tell us something.
Not long ago, Musk publicly stated that in the future, driverless taxis will be more common than human-driven cars. He even agreed with a netizen on platform X that driverless taxis would "kill" public transportation.
With a goal, Musk will naturally implement this set of ideas to Tesla. He believes that if the ability to fully drive itself cannot be realized, Tesla's commercial value will be "basically zero". It can be seen that no matter how well the Model series sells, Musk has never regarded it as a car in the traditional sense, but a platform for autonomous driving.
In Jiang Wen's words: For this bit of vinegar (Robotaxi), I made this dumpling (Model series).
At the 2019 Autonomous Driving Day, Musk publicly said that the first phase of Robotaxi will launch a business model similar to online car-hailing and car sharing, and Tesla will build its own fully autonomous Robotaxi fleet, which will be operated by itself in the initial stage, and its taxi cost will be lower than the current bus fare.
Even in Musk's vision, Tesla will not only operate its own fleet, but also allow Tesla owners' cars to join the car-sharing team, and car owners can make a profit of about $30,000 a year.
However, this is all built on a future where autonomous driving has become standard. The current reality is that Musk's ambitions in the field of autonomous driving have not been realized.
Open Tesla's official website, there is a line similar to the "exemption clause" for Autopilot and FSD: the currently enabled functions still require the driver's active supervision, and the vehicle has not yet achieved full self-driving. This means that both Autopilot and FSD are still far from the technology that Musk envisions.
Many people said after experiencing FSD that it was like handing over control of the car to a novice teenage driver, which once caused a lot of doubts in the industry.
Philip Koopman, a professor at Carnegie Mellon University who studies the safety of autonomous vehicles, said: "Other companies have found that they originally thought that autonomous driving was a project that would take 2~3 years, but in fact it takes 10 or 20 years to develop. In fact, Tesla has also discovered this. ”
In December, Tesla agreed to recall more than 2 million Autopilot-equipped vehicles under pressure from U.S. safety regulators, despite raising objections to the U.S. government's safety concerns, and made remote software updates to the system to add more owner warnings.
Clearly, the global development of autonomous driving is not going well, and the latest example is that Ghost Autonomy, a self-driving software company, announced in early April that it would cease its global operations after only five months of working with OpenAI, citing many uncertainties about the road to profitability.
The cost is large, the revenue is slow, this is the microcosm of the entire autonomous driving industry. Without the support of autonomous driving, the future of Robotaxi will become more and more uncertain.
The California Department of Motor Vehicles (DMV) suspended the operating license of Cruise's driverless taxis last October after a Cruise's driverless taxi crashed into a pedestrian towing several meters, and in February, Waymo announced a recall of software previously used for its driverless car after two robotaxis crashed into the same pickup truck within minutes.
Musk's dream of autonomous driving may have to accompany Robotaxi for a long, long time.
●比Makeup更「Open」的I
Lose in hake, but gain in herring. Don't look at the uncertain future of Musk's Robotaxi, he has received recognition and applause from a large number of users in the field of artificial intelligence.
On March 17, local time, Musk's artificial intelligence startup xAI officially announced the open-source large model Grok-1, which follows the Apache 2.0 protocol and opens up the model weights and architecture. According to the official website, xAI has open-sourced the weight and architecture of Grok-1 on GitHub, a software hosting platform.
The matter of open source sounds like "the predecessors planted trees, and the later generations enjoyed the shade", but in fact, Musk has always scoffed at technology monopoly. On March 11, Musk said on X that xAI will open the source code of the chatbot Grok, which means that the public will be free to use the code behind the company's large model technology, and then Meta CEO Zuckerberg also commented on X to express support.
某种意义上,它是比Makeup更「Open」的I.
In fact, Musk was an early investor and co-founder of OpenAI, but because he and Sam Altman did not agree on AI security issues, Musk left OpenAI in 2018, and later he became one of OpenAI's most high-profile critics.
For CEOs of technology companies, an open mind to breaking down patent barriers can not only promote the development of the industry, but also establish core standards. That's true for Grok-1, it's true for Android, it's true for Tesla.
In 2014, Tesla disclosed about 270 patents, and the top patents classified by patent number were H02J7/00 battery charging circuit design, H01M10/50 thermal management temperature control system, H02J7/04 charging current and voltage control, H01M2/12 appearance packaging, H02J7/16 charging current and voltage field strength change and other patents.
Combined with the previous data, it can be found that most of Tesla's patents published at that time were battery management systems (BMS). This proves that Tesla hopes that through the disclosure of patents, new energy companies can uniformly adopt standard battery management systems, promote charging piles and charging technology, improve supporting charging stations to quickly promote the market share of electric vehicles, unite against traditional fuel vehicles, and grasp the right to speak in the industry.
This is the same idea as when Google open-sourced Android and seized the smartphone highland to improve the ecosystem. You know, the new energy vehicle market in 2014 is far less hot than today, in order to leverage the fuel vehicle market, Tesla can be said to have done its best.
On June 12 of that year, Musk made it clear in his open letter "AllOurPatentAreBelongToYou": "Our real competition is not the small number of non-Tesla electric vehicles being produced, but the large number of fuel vehicles that gush out of the world's factories every day, and we believe that applying the concept of open source to our patents will strengthen rather than weaken Tesla's position in this regard." ”
In recent years, the concept of technological equality has been mentioned repeatedly. This open ecosystem not only accelerates technology iteration and optimization, but also brings potential economic benefits to the company. If the development of autonomous driving did not take advantage of the time and place, then at least the open source of AI has allowed Musk to usher in people.
● Robots need love, but they also need "death"
Autonomous driving, AI, robots, they can be different things or they can be one thing.
In Interstellar, Nolan depicts the feasibility of intelligent robots in the future, but in fact, no matter how the image of TARS changes, it will still be a mobility terminal in the end, which is very close to the concept of today's car.
Not long ago, Jiyue 01 was officially launched, although the product named "car robot" is suspected of excessive marketing, but the functions it covers do have the prototype of the future mobility terminal, such as fully localized voice AI algorithm, Wenxin Yiyan's large model, intelligent driving of pure visual solutions, etc.
And what about the humanoid robots that we are familiar with? This is a gem in the automotive industry.
Musk recognized the future of humanoid robots, so they created Optimus. Coincidentally, companies such as Xpeng, Xiaomi, and NIO have set their sights on humanoid robots, and they seem to agree that robots will shine in the automotive field of the future. It can be both the creator of the product and the product itself.
Not long ago, the 2024 Chinese Humanoid Robot Ecological Conference opened in Shanghai, and Yi Peng, head of NIO's intelligent manufacturing engineering humanoid robot, said that humanoid robot is an integrated tool with a human body and a high degree of intelligence, and has relatively large application potential in the replacement and automation of dangerous operations in automobile manufacturing.
Yes, the ability to cope with dangerous operations is a major core of humanoid robots entering the industrial chain. Although it sounds a little cruel, such products are used as tools, and "generosity to death" is the best response to "love". We don't have to feel guilty about this, just imagine, those cars that do crash tests before leaving the factory, don't they also end up unrecognizable, but in the end they serve humanity?
Especially for the current problem of rising labor costs in global car manufacturing, humanoid robot car manufacturing also provides a new way to solve the problem, which not only brings lower production costs, but also promotes the global low-carbon transformation.
NIO has made two main attempts to build humanoid robots, one is to fix the humanoid robotic arm in a small space and use 5G and vision algorithms to perform multi-task operations, and the other is to conduct quality inspection of some items of the humanoid robot.
The advantage of humanoid robots made by car companies is that various algorithms and hardware sensors of vehicles can be perfectly adapted to the robot products themselves, such as Optimus, which uses the chip of Tesla's Dojo project, and its underlying algorithm is the same as FSD, and even the hardware of the camera is the same.
It is not only a product of the transformation of mobile terminals, but also a new singularity of autonomous driving, which realizes a closed loop with technologies such as FSD and AI.
Moreover, from the perspective of industrial development, with the participation of humanoid robots, it is possible to realize a complete closed-loop of industrial automation in the future, which inevitably requires human participation.
In January this year, Figure AI, a hot overseas humanoid robot unicorn company, signed a cooperation agreement with BMW Motor, the former to deploy humanoid robots in phases at BMW's South Carolina plant to integrate into BMW's manufacturing process. It can be seen that there are not only new forces that are eyeing the cake of robots, but also traditional luxury brands and other technology companies.
According to Bloomberg, Apple has plans to enter the field of home robots after abandoning its electric car plans, and it may become Apple's next major project along with the Vision Pro.
Specifically, Apple's engineers are developing a mobile robot that can follow the user's movement around the home, but not in the same form as the Samsung Ballie, but closer to the Optimus. The project is still in the early stages of development, with Apple's hardware engineering department, artificial intelligence department, and machine learning team overseeing research and development of domestic robots.
The application scenario of this robot is to provide users with assistance, entertainment or other customized services, Apple internally calls this project "Skunk Works", and the technology transfer has also learned from Tesla's approach, and the autonomous driving, sensor technology and artificial intelligence algorithms accumulated in the previous car project have been reapplied to these domestic robots.
It remains to be seen how much of a repercussion they will have in the automotive industry in the future, but it is a fact that costs have fallen sharply. Last month, Musk said on X, "Although the humanoid robot is more complex per unit mass, I think the final manufacturing cost will still be less than half that of Tesla cars." ”
Similar to NIO's vision, Optimus' original goal was to make it perform dangerous, repetitive, or tedious tasks.
Adam Jonas, an analyst at Morgan Stanley, previously estimated that the market potential for humanoid robots is huge, and they can replace the workforce around the world. He wrote in a report at the end of December that the global labor force market is about $30 trillion, or about 30% of global GDP.
If humanoid robots can penetrate our lives in a variety of ways, not only the automotive industry, but the whole world will be turned upside down.
● Write at the end
A product based on the world must have its irreplaceable role. For cars, this is the case for mobility.
However, with the continuous development of technology, the concept of mobility terminals is slowly blurring the attributes of automobiles, autonomous driving, robotaxis, artificial intelligence, and robots, all of which ultimately aim to serve the same group of people in different ways.
It's like the blood-pumping mechs in Pacific Rim – designed as a Gundam to defeat monsters, and a cannon to defeat monsters. For us ordinary people, in addition to some beautiful imaginations, we should also give them some time. The answer may lie within.