Author: Michelin
Recently, “the people” are especially busy.
Since Wuling launched the slogan “We produce what the people need” in the automobile industry, “the people” have become a brick to be moved wherever it is needed.
“Building cars for the people”, “defining intelligent cars for the people”, “the people need intelligent cars”…
These headlines make me wonder: Do the people really need intelligent cars?
What is an intelligent car?
Let’s take a look at what kind of intelligent cars are offered to the people.
Since the policy of clean energy replacing traditional fuel, new energy vehicles have become the mainstream development trend, and the intelligentization of automobiles has ushered in greater space. Electrification, autonomous driving and intelligent cabin have become the standard equipment of contemporary intelligent cars.
With the coordination of the “three-piece suit”, clean energy has gradually replaced traditional fuel; intelligent driving can assist human beings in driving, and even completely liberate their hands and feet; the diversified interaction in the cabin also brings users different services. Cars are not just transportation tools. Under everyone’s beautiful imagination of future intelligentization, cars are even endowed with the functions of “second bedroom” and “second office”.
From a God’s-eye view, we certainly need intelligent cars
In this era where even the broom in the home is gradually replaced by a sweeping robot, intelligent cars replacing traditional cars seems to be an inevitable fact of historical development.
Moreover, under the transformation of the energy and automobile industries, intelligent cars have created more new opportunities.
Looking at the recently released TOP5 of global electric vehicle sales in Q1 2021, among the car companies leading in traditional fuel cars, only Volkswagen made the top list, while three of the other four are Chinese brands: Wuling Hongguang, BYD, and Euler. This was unimaginable in the past era of traditional cars. And this overturning of the pattern of the automobile industry is exactly what electric vehicles or intelligent cars have brought.
Not only the automobile industry, but also the intelligentization and interconnection of all things have also provided other industries with new opportunities. Looking at recent automobile news, technology companies, internet companies, smartphone manufacturers, home appliance companies, and even the liquor industry have joined the “war” of intelligent cars.
Yes, the times require intelligent cars, the automobile industry needs intelligent cars, and Chinese companies need intelligent cars even more. However, do you really need an intelligent car?## The “Martyr” of Technological Progress Should Not Be the People
As electric car battery range approaches the target of 1000 kilometers, chip computing power increases, sensors become sensitive up to centimeters or even more, and 5G technology reduces communication delay to microseconds, these technological innovations have made smart cars full of limitless imagination, appearing to be within reach.
However, ideals are abundant, but reality is stark.
Any new technology that is put from laboratory, test bench to mass-produced cars requires a huge investment of time and money, hundreds or thousands of trials, numerous reliability verifications, and data iteration updates. Currently, automakers choose the route of “developing while popularizing”, by putting products into the market early, to balance the high R&D investment and time consumption, to drive up enthusiasm for smart cars and stock valuation, and further optimize functionality through user feedback provided by products already on the market.
This approach does indeed enable us to use smart cars more quickly, but the R&D costs, iteration time, and reliability verification cannot be completely eliminated solely by technological progress.
“All gifts of fate are secretly marked with a price tag”, and smart cars are no exception.
When immature technologies enter the market through “developing while popularizing”, accidents such as electric car self-ignition, autonomous driving out of control, and electric control system failure are bloody prices to be paid. Furthermore, there are also user privacy issues under in-car cameras, software data security, and early electric car sudden power depletion, causing real troubles for users despite being unrelated to life and death.
Ultimately, these costs are borne by users, who become the “guinea pigs” of smart car progress.
The Intelligence of Cars is Slower Than Imagination
The promotion of smart cars makes us believe that intelligent vehicles will arrive quickly and that the cost and effort are only temporary and limited. However, in fact, the pace of smart car evolution is much slower than we imagine, and true smart cars are still far away from us.
First, let’s take a look at the most sensitive issue for consumers, the price.
In front of the traditional hundred-year history of the automotive supply chain, the supply chain of smart cars is still in its infancy, and sales are only a fraction of those of gasoline-powered cars, resulting in higher costs. Moreover, the smart components such as chips and sensors used in smart cars, as well as the early R&D costs, make the current price of smart cars generally higher than that of traditional cars.Using the recently released BAIC Arcfox αS as an example, the Huawei HI version starts at a price of 388,900 yuan, which is 50% higher than the standard version. Even though the price is more than ten thousand yuan, it is likely only enough to cover the cost of this HI intelligent device.
With such a high cost and a price approaching 400,000 yuan, how many users can afford it?
In a 2017 car sales statistics, the price range of 80,000 to 160,000 yuan was the most popular among domestic users, accounting for 45.8%, almost half, while users willing to spend more than 400,000 yuan accounted for only 4%.
Although the cost and price of electric vehicles have been reduced in the past year, there are not many options in the 80,000 to 160,000 yuan range for those who want to buy an intelligent car with sufficient endurance, automatic driving, and complete in-cabin functionality.
In fact, the new sales model derived from intelligent cars has already lowered the threshold for car prices and avoided high cost prices that discourage users. For example, using the form of hardware embedding and software OTA upgrades separate the prices of software and hardware, such as Tesla’s FSD and XPeng’s NGP, which make automatic driving an optional feature. Some directly use different software, hardware, and intelligent functions as new versions, such as the aforementioned BAIC Arcfox αS Huawei HI version.
However, this means that users who think they bought a cheap “intelligent car” may only have just entered the threshold, and in order to truly experience the intelligent functions, they still need to make “in-app purchases”.
Of course, as the intelligent car industry chain gradually matures and production volumes increase year by year, costs will gradually reduce, making intelligent cars potentially as affordable as traditional cars. However, this still takes time.
If we don’t consider the money, can we just buy an intelligent car at will?
In fact, even if we don’t consider cost and price, the evolution of intelligent technology is not that fast.“`
Using intelligent driving functions as an example, the past year has seen a significant improvement in a car’s perception abilities thanks to the mass production of on-board lidar, the 4D imaging effect, and the advancements in the new generation of chip capability. The application of deep learning has made cars as intelligent as possible, almost like humans. However, real road conditions are far more complex than simulation systems and test benches. Even humans can feel helpless in unfamiliar environments or situations they have never encountered before. Similarly, intelligent cars, despite improvements in software and hardware, can still only mark corner cases with patches to solve the problem the next time they encounter a similar one. What about this time? It’s still up to the driver.
This means that autonomous driving, which truly frees the driver’s hands, requires massive testing and experience of all real-road scenarios to reduce the proportion of corner cases to 0.1%, 0.01%, or even smaller.
As of September 2020, Waymo, the autonomous driving giant, had driven 10 million kilometers autonomously, equivalent to ten round trips to the moon. However, even with such a long test distance, there were still 18 accidents and 29 human interventions to avoid collisions. Last month, Huawei’s stunning demonstration of autonomous driving was supported by Huawei’s hardcore technology and large-scale early testing of highly accurate maps of the same section.
The acquisition of massive data and the update frequency of dynamic data per minute or even per second, the time and cost required for repeated testing of actual roads are all pitfalls in the development of autonomous driving.
Therefore, despite the fact that some brands have announced the production of “L4 level autonomous driving cars” at the Shanghai Auto Show, the vast majority of companies in the industry are still focusing on L2+ and L2++ levels of autonomous driving, targeting autonomous driving functions in specific scenarios such as parking lots and highways. The seemingly simple goal of “freeing hands” is still a headache for all companies.
The so-called L4 or even L5 level of autonomous driving is still far away from us.
Are all the features in the cabin really useful?
Not just for autonomous driving, but for seemingly harmless intelligent cabins as well.
Compared to autonomous driving, the concept of smart cabins is more “youthful” and, because it is “youthful,” smart cabins are still in the exploration stage. In the cabin, you can see a variety of attempts, and even the shadow of smartphone applications.
“`The many features available in modern cars have made our lives more convenient. For example, voice-activated functions, eye-free controls, and adjustments for air conditioning and seats have made the driving experience more comfortable than ever before. However, some of these features, such as those requiring prolonged screen time, may not be practical for use while driving.
Until fully autonomous driving becomes a reality, humans will still need to keep their hands on the wheel and their eyes on the road. This means that activities like watching videos, playing games, shopping, and ordering food are unlikely to be able to be done while driving, and can only be done when the car is parked.
If car manufacturers focus solely on flashy features as the selling point for their smart cabins in the absence of fully autonomous driving, they may find themselves involved in a new cycle of one-upping each other with more visually stunning features rather than practical ones.
In addition to flashy features, many car companies are also promoting their ecosystems as selling points. These ecosystems include platforms like SOA, which connects the car to smartphones and smart home devices, and are aimed at users who want to make their cars more intelligent. However, just like the first generation iPhone software, which lacked 3G support and an app store, and was criticized by users for being difficult to use, current car ecosystems are still very much in the “every man for himself” phase. The options for customization are limited and the functions that connect the car to other devices are heavily reliant on smartphones and smart infrastructure. Overall, the intelligence of car cabins is not as rich and colorful as we might expect.
You may feel that this article has been overly critical of smart cars. As the publication GeekCar, you may wonder where the geek spirit has gone. In reality, we have been waiting for the transformation of cars and the advancement of smart cars, and we are eager to lend our own voices to support and promote technological development. However, as smart cars enter their growth phase, it is important to temper expectations and provide a dose of reality. Blindly promoting and exaggerating the benefits of smart cars may not only deceive consumers, but also hamper the development of smart cars.
Smart cars are still in their infancy, having only been around for about a decade. Just like a child cannot grow up overnight, smart cars will also experience countless trials and adjustments before becoming perfect. By addressing the problems that come with trends and ensuring that products meet users’ needs, we will be able to create a world where people can use smart cars that are safe, reliable, intelligent, and responsive to their every need.
This article is a translation by ChatGPT of a Chinese report from 42HOW. If you have any questions about it, please email bd@42how.com.