Communication group| Enter "Sensor Group/Skateboard Chassis Group/Car Basic Software Group/Domain Controller Group", please scan the QR code at the end of the article , add Jiuzhang Assistant , be sure to note the name of the exchange group + real name + company + position (no remarks Unable to pass friend verification)
Editor | Su Qingtao
When they need to be "arty" and improve their style, they like to call "software-defined cars" like others, although their understanding of the concept of "software-defined cars" may not even be "smattered"; When talking about money, they decisively believe that software is either only worthy of being "whoring for nothing", or it can only be sent "under the fence" of hardware.
The root cause behind this absurd phenomenon is that most of the bosses of OEMs lack "software thinking". Furthermore, the way of thinking of many people in power has not "kept up" with the era of "software-defined cars".
Strictly speaking, this cannot be regarded as a serious "article", but a "patchwork" of a series of fragmented observations and thoughts.
The main topics covered in this article are:
1. Only know benchmarking, pay attention to competing products rather than consumers
2. OEMs move from "separate designation" to "designated Tier 1"
3. The autonomous driving suppliers who advise car companies to "don't self-develop" are basically doing "full-stack self-development"
4. The biggest obstacle for L4 companies to obtain orders: the "mother-in-law thinking" of OEMs
5. The OEMs that have contributed the most to the rise of China's intelligent supply chain
6. The excessive publicity of "L3" back then was a pit dug for today
1. Only know benchmarking, pay attention to competing products rather than consumers
Many people, when picking up a girl, their attention is not on the girl, but on the rival. Many people who are responsible for product definition in OEMs usually do this—they pay less attention to consumer needs and are less sensitive to user experience, but they pay too much attention to competing products.
Some time ago, the marketing director of an autonomous driving company said:
Many decision-makers in OEMs do not understand cutting-edge technologies, but they are very serious when benchmarking against competing products. In particular, once something they didn't believe in before is produced by a competing product, their attitude will immediately change drastically.
For example, before the mass production of DJI Vehicle's cooperation project with Wuling, many bosses of OEMs came to ride the demo car to see if it was necessary to cooperate with DJI Vehicle. However, there was nothing to do after returning; After Wuling's cars were mass-produced, they all bought them for research.
"It's all caused by the bidding." The marketing director concluded.
2. OEMs move from "separate designation" to "designated Tier 1"
Previously, OEMs tended to "separately determine" autonomous driving suppliers. The main reasons are: to strengthen the ability to compete with suppliers and reduce the dependence on a few specific suppliers (Tier 1); we do not want suppliers to master a complete product definition.
However, after several years of practice, some OEMs have begun to abandon the "separate designation" strategy and return to the old path of "designated Tier 1". The reason is: After separating the fixed points, I found that supplier management is too difficult, especially, many technologies are not understood by myself, and I don’t know who to ask for responsibility after problems occur. It's easier to come to the bottom.
A technician from a supplier company said: OEMs still need a Tier 1 to complete many distributed tests and verifications in engineering mass production. They are looking for a blame man, and Tier1 is a strong blame man.
It seems that after suffering some losses, OEMs began to gradually recognize the boundaries of their capabilities, so they had to give up some "looking good" wishes.
3. The autonomous driving suppliers who advise car companies to "don't self-develop" are basically doing "full-stack self-development"
In the past half a year, the author often joked with friends in the supply chain: "All suppliers tell OEMs that they don't have to do full-stack self-research, or at least "I do this, you don't have to do it yourself." On the other hand, most vendors are 'full-stack' within their capabilities."
In the article " Software and Hardware Decoupling: Full of Ideals, Skinny in Reality " published in mid-April , we mentioned a phenomenon: In addition to OEMs, chip manufacturers, algorithm companies, and domain controller companies are also making middleware , while middleware companies have to do algorithms and domain controllers because their living space is squeezed.
In an exchange before the auto show, the author asked the person in charge of the software of an autonomous driving domain controller company: "Why does your company that makes domain control hardware also engage in middleware?"
His answer is: "It's not that we' insist on doing it, but if we don't do middleware ourselves, it's hard for us to convince customers that our domain controllers can help them solve any problems. Simply put, domain controllers Vendors don’t need to sell middleware as a product, but they must have (middleware), and middleware must be made in order to make demos; if they don’t make middleware, domain controllers may not be sold.”
At the same time, some domain controller manufacturers are also working on perception algorithms.
During the auto show, the CEO of a domain controller company explained to the author the reason for his company to develop a perception algorithm: If the perception algorithm relies too much on external partners, our ability to expand the market will be affected. For example, I want to expand into the Southeast Asian market, but if the visual algorithm partner has not paid attention to driving scenes in Southeast Asia, and their algorithm cannot recognize road signs in Indonesia, then we will be in trouble.
It seems that the "stretched front" is indeed out of helplessness.
Companies that originally focused on perception algorithms, in addition to doing middleware, are also expanding into regulation and control algorithms. The reason is, “If perception and regulation are done by different companies, once an accident occurs, it will be difficult to pursue accountability. Is the problem in the perception link or the regulation link? It is more difficult to pursue accountability within the same company. The efficiency of coordination between the two companies is much higher.”
What's more worth mentioning is that almost all algorithm companies are developing their own hardware. The most typical is the self-developed domain controller, and Zongmu has extended its tentacles to 4D millimeter-wave radar, ultrasonic radar, 8-megapixel camera and other sensor fields.
It can be said that most algorithm companies have done "full-stack self-research" within their own capabilities. Judging from the situation learned by Jiuzhang Zhijia, the main reasons are as follows:
1.
Software and hardware synergy can make the algorithm work better
Many low-frequency problems encountered in mass production of autonomous driving are at the bottom of the system. To solve these problems, engineers often need to have a certain understanding of each module of the system. In this regard, full-stack self-developed The advantages of the company will be more obvious.
People from many autonomous driving companies believe that the strongest suppliers in the pre-installation mass production market are Huawei and DJI, two companies integrating software and hardware. "To make a good system solution, you must have rich hardware experience."
In addition, the CEO and CTO of Zongmu Technology, the company that has won the most pre-installed mass production projects in China, have a long experience in software development in chip companies, and "software and hardware are integrated" in the knowledge structure.
2.
In order to avoid being dragged down by "pig teammates"
The CEO of an autonomous driving company that started out with algorithms said: When building a complex system, the higher the degree of external dependence, the greater the probability of problems. If you have 5 partners, each partner may have a 10% probability of losing the chain, which means that the success rate is only 90%. If you multiply these 5 90%, the success rate is less than 60%.
What makes Algorithm particularly wronged is that many problems are obviously caused by the underlying software made by the domain controller manufacturer, but when something goes wrong, the host factory is the first to hold the algorithm accountable.
The relationship between an algorithm company and a domain controller manufacturer is different from that between an OEM and a domain controller manufacturer. The host factory and the domain controller manufacturer are "the relationship between Party A and Party B". When A makes a request to Party B, Party B has to cooperate; but the algorithm company has a "left-right relationship" with the domain controller manufacturer. They found that the domain controller manufacturer provided If there is a problem with the plan, you can only make some suggestions to the other party, but you have no right to ask the other party to cooperate, let alone ask the OEM to replace this "pig teammate".
Since the command doesn't move "left and right", and you don't want to be dragged down by the opponent, you can only go into battle in person.
3.
OEMs have less trust in pure algorithm companies than hardware companies
After trying to "separate hardware and software" for a period of time, many OEMs are still more inclined to find a Tier 1 to "cover the bottom line". Of course, when choosing Tier 1, OEMs have relatively high trust in companies with hardware backgrounds, but relatively low trust in pure algorithm companies.
Although this cruel truth makes pure algorithm companies extremely unhappy, it is easy to understand.
First of all, for decades, OEMs have been purchasing from Tier 1 with a hardware background, and software only exists as an accessory to the hardware. Therefore, OEMs basically have no experience in dealing with software suppliers. Since we have never dealt with each other before, where does the sense of trust come from?
Secondly, when hardware companies invite customers to visit their own companies, they will show them their laboratories and production lines, which are all "real" investments; Apart from the demo car, there is only ppt - of course they will say that they also have a large number of high-end talents, but can they really expect those customers with mechanical background to know how to price their algorithmic talents?
In addition, being a Tier 1 OEM requires strong supplier management capabilities, while the upstream supply chain of pure algorithm companies is relatively simple, so the accumulation of supply chain management capabilities is indeed insufficient.
In addition, even if there is only a small problem with the hardware, it may need to be recalled, and the supplier has to pay a huge cost for this, which will also make the hardware company more in awe of the quality standards of the automotive industry; In contrast, software problems can basically be solved through OTA, which makes it easy for pure algorithm companies to think of some complex problems as "simple and light", so their acting style is "too bold", but the problem is that you If you are "bold", the customer may be frightened into "timid".
4.
OEMs are unwilling to pay for software, software needs to be bundled with hardware to sell money
When communicating with Jiuzhang Zhijia, many people in the industry mentioned that "pure algorithm companies" are easily replaced, so they have no right to speak in front of OEMs.
An autopilot product manager of an OEM said: "To be honest, pure algorithm companies do projects for traditional OEMs, and the contract amount is 3 million. Later, they may really get back a little more than 1 million, and the remaining 2 million Wan, the main engine factory will not give it to you. Because he thinks that what you make is 'only worth 1 million', and you can't tell them anything."
A self-driving algorithm company found that it can make money by outsourcing labor services to OEMs, with gross profits as high as 200%-300%, but its main business (selling algorithms) has a gross profit of only 10%. "Pure software can't be sold for money. Unless you bundle the software with the hardware, rely on the hardware to collect money, and bring the software in."
For a long time, in the automotive industry, software was just an accessory to hardware. Although it is now known as the era of "software-defined cars", you have to count on those big and small bosses of OEMs who have been working in machinery all their lives and lack software thinking from the bottom of their souls to truly understand the value of software at the ideological level. It seems to be a bit embarrassing for them.
Many friends from autonomous driving companies mentioned when they communicated with the author, "People in OEMs think that software has no cost, and they should not pay for software."
Of course, there are a large number of elites in OEMs who are aware of the value of software and think that they should pay for the software, but how to do it? Procurement can be confusing: software doesn't have a BOM like hardware, so how do I price software? How to explain clearly to the accountant, what is the logic of my paying 100 yuan for this software? Finances are even more of a headache, and he doesn't know how to explain it to his boss.
(Pricing according to the BOM structure is nonsense—should the price be determined by cost instead of value? Pricing according to the BOM is purely based on the value of "suppliers should not make money".)
Compared with algorithm companies that are almost flocking to make hardware, Jinmai, TZTEK and other companies that make domain controller hardware do not intend to get involved in algorithms. In fact, there are far more algorithm companies forced to do hardware than hardware companies involved in algorithms.
In a recent salon, when talking about "how to make more money", the technical VP of a self-driving company that transformed from software to a domain controller said: "If you are looking for a job, you can earn more money as a software engineer." money; if you’re starting a business, it’s easier to make money doing hardware.” (To be precise, it’s not “hardware is easier to make money” but “software is harder to make money.”) How ironic.
Software companies believe that software is the soul and hardware is the body, but it is extremely sad to find that the soul is inseparable from the body far more than the body is inseparable from the soul.
The reason is that if hardware companies don’t make algorithms, they just “make a little less money”, and if algorithm companies don’t make hardware, they simply “can’t live”.
The root of this problem may lie in the fact that the way of thinking of many downstream customers is generally not "worthy" of the era of "software-defined cars".
4. The biggest obstacle for L4 companies to obtain orders: the "mother-in-law thinking" of OEMs
Regardless of whether the author was questioned "How much advertising fee did you charge L4 company before you dare to speak nonsense here" because of the previous article " Engineering capability is not an insurmountable obstacle for L4 autonomous driving companies | Nine Chapters of Autonomous Driving Essay II ", It is really not easy for those companies that transition late to win mass production orders from OEMs. At least, it will be difficult to gain the trust of OEMs for a while.
But the crux of the problem is not that these self-driving companies are not capable, but that the decision makers of OEMs are not at a good level of cognition.
As we said in the previous paragraph, even the people in L2 Company believe that "engineering capabilities are not enough to constitute a barrier to competition. As long as there are projects, L4 Company can take some time to make up for them" (after all, L2 Company also started from "engineering It has come from the stage of "poor capability", but the main engine factory is still holding on to "engineering capability".
No matter how awesome the company is, the first generation of products will always be unsatisfactory. For example, is Apple's iPhone 1 awesome? For another example, in 2010, when the author came into contact with Huawei mobile phones for the first time, the impression was still "low-end phones, free with phone charges". Who would have thought that it would become the high-end of the high-end in the future?
Just like people, the company is constantly iterating. If you insist on using his level of making the first product to conclude that his level of making the second and third products is also the same, isn't that "pure hardware thinking" up?
It is difficult for L4 companies to get orders now. On the one hand, it is really difficult for these companies’ current engineering capabilities to meet the requirements of OEMs. On the other hand, the thinking of those responsible for screening suppliers in OEMs is static and cognitive Is fixed - they only look at the past and present of the supplier, have no interest or ability to assess the potential of the supplier. This is essentially a "mother-in-law thinking".
What is "mother-in-law thinking"? A poor boy, his girlfriend believes that he has a bright future, poetry and distant places with him, and is willing to accept that he has nothing now, but the short-sighted mother-in-law only cares about your current monthly income, whether you have bought a house, and what kind of car you drive?
Chinese mothers-in-law are incapable of understanding the prospects of their prospective son-in-law, and many mediocre people in OEMs are also incapable of evaluating the future of suppliers. Therefore, apart from "what projects have you done in the past", they don't know what other indicators to use to evaluate the supply business ability.
Once, the author saw a job advertisement of McDonald's, which said: Others evaluate your experience, we tap your potential. One sentence points out the essence of a great company compared with ordinary companies.
As early as more than ten years ago, the author discovered that those powerful bosses pay more attention to the potential of candidates (mainly the bottom-level ability) when recruiting, so as to stock up potential talented people first; while some second-rate bosses only look at How rich the candidate's past experience is, as for how high the gold content of the candidate's "three-year experience" is, is it real three-year experience, or just "simple repetition of an experience for three years", the second force the boss Absolutely no judgment whatsoever.
Assessing potential is very difficult, but looking at experience is very simple. As a result, many second-hand bosses have given up the "difficult and correct" method and chose the "simple and mediocre" method. As a result, it is conceivable whether the recruits are excellent or not.
The bosses of the OEMs who have sentenced their future to death because of the "lack of project experience" of some autonomous driving companies are very similar to the latter kind of "secondary bosses".
Decision makers who think too much about hardware have no sense of the concept of "iteration". In the era of "software-defined cars," this is a sad thing. I don't know how these people have the nerve to talk about "software-defined cars"?
At present, USD funds cannot invest in Chinese technology companies, but most RMB funds are short-termists and do not believe in poetry and distant places (this is why ChatGPT cannot be the first to appear in China), so they will not be moved by the story of Robotaxi. However, in order for autonomous driving to move from assisted driving to unmanned driving, it still depends on these L4 companies that "look up at the stars" and the talents they have cultivated.
Visionary OEMs should give these companies the opportunity to grow with them for a period of time.
Entering 2023, an obvious trend in the autonomous driving industry is that the direction of "full-stack self-research" by OEMs has changed drastically. "More details are covered in the fourth article of the series of articles), practice has proved that most OEMs are indeed incapable of self-development. Since I am incapable, it is better to get rid of the shackles of "mother-in-law thinking" as soon as possible and support several potential suppliers.
Perhaps, many people from OEMs can also see the long-term potential of these L4 autonomous driving companies, but they just don’t dare to be the “first to eat crabs”—they may plan to hitch a ride. However, in my opinion, choosing a free ride on such a major issue is a typical manifestation of "low emotional intelligence".
When you choose an L4 company that clearly has potential but cannot meet "goods knowers" when they need help the most, it is a timely help. And with the experience of sending charcoal in the snow, both parties will have a revolutionary relationship in the future, and the supplier will be "destroyed" when doing things for you. It can be imagined that at some point in the future, a certain technology of advanced automatic driving is very scarce, and only his family has it, but his service ability is limited, so he must give priority to serving you well, after all, you have had him before." The kindness of knowing you." Well.
When I was young, I read the biography of Du Yuesheng, the boss of the gangster in Shanghai in the 1920s and 1930s, and found a very interesting phenomenon:
Du Yuesheng was able to rise step by step from an ordinary boy from a poor family to the top class of society, and became a man of the day. One of the key secrets was that he especially liked to help those down-and-out potential stocks within his ability. After the success of the potential stocks, they are naturally willing to repay Du Yuesheng, even to the point of utter desperation.
5. The OEMs that have contributed the most to the rise of China's intelligent supply chain
Before going to press, the author chatted with a friend who has worked in the automotive supply chain for more than ten years and is now working in a leading company in a certain industry about the "mother-in-law" thinking of the OEM, and the other party particularly resonated:
Yes, especially in traditional OEMs, the decision makers themselves have mediocre judgment and are even more unwilling to take risks. They only want to find ready-made mature players, which leads to being passive everywhere. The significance of new car-making forces is that they dare to join forces with innovators, grow together, and eventually create a strong and leading industrial ecology.
"New car-making forces dare to work with entrepreneurs and grow together", which reminds me that at the end of 2018 or early 2019, Ideal's internal test results for Horizon J1 were "poor", but later, Ideal still became the first to follow An OEM with deep cooperation with Horizon.
In early November last year, the author stated in the circle of friends that "although Ideal is not the OEM with the highest sales volume, it is the OEM that has contributed the most to the rise of China's automotive intelligent supply chain." This circle of friends has resonated strongly with many friends who work in companies in the supply chain.
Fortunately, there are such forward-looking OEMs as Ideal, otherwise, if only rely on those OEMs with deep-rooted "mother-in-law thinking", it will be difficult for companies such as Horizon and Hesai to stand out.
Many suppliers who have cooperated with Ideal have a very high evaluation of Ideal. Even some suppliers who have not formally cooperated with Ideal, but have only been in contact with them, have a high evaluation of Ideal.
The marketing director of a sensor company said: "Ideal is a company that integrates knowledge and action. They pay great attention to transparency. book system."
The CTO of a self-driving company said: among so many of our customers, the ideal is to provide us with the most OTA demand - usually tell us the OTA demand 3 months in advance, and they are willing to pay the development fee (mostly ).
Mentioning that this CTO especially emphasized that ideals are willing to pay the development fee to the supplier for OTA, the author was surprised: "Aren't other customers willing to pay the development fee when they ask you to do OTA?" Some of them are prostitutes.”
Many people in the main engine factory are very strange-when they need to improve their standards, they shout "software-defined cars", but when they need to talk about money, they think that software is only worthy of being "free prostitutes" or only Can be sent under the "hedge" of the hardware.
The root cause behind this absurd phenomenon is that most of the bosses of OEMs lack "software thinking". Furthermore, the way of thinking of many people in power has not "kept up" with the era of "software-defined cars".
Judging from what the CTO of the self-driving company said, "Ideal is willing to pay for OTA", it can be seen that Ideal is indeed one of the few OEMs that can keep up with the era of "software-defined cars".
6. The excessive publicity of "L3" back then was a pit dug for today
At present, the most serious challenge facing the entire autonomous driving industry is: end users generally do not have high satisfaction with autonomous driving functions, so they are unwilling to pay for automatic functions; even if they pay, the utilization rate is not high.
The reasons here include: OEMs have insufficient research on user needs during the product definition stage, resulting in a gap between product functions and actual user needs; the ODD of the automatic driving system is not clear, and consumers do not know when it can be used. When not to use; the actual capabilities of the autonomous driving system do not match the user's expectations.
Regarding "user expectations", at a salon during the auto show, the product director of a self-driving company said: Although OEMs and self-driving companies only dare to say that their products are "L2+", consumers' expectations for it are not. It's "L3", maybe even L4. "Users hope that they don't have to do anything, and the system can be fully responsible." "If a high-level automatic driving system requires my concentration, then this is similar to the previous basic L2. I might as well not spend this money."
In other words, why do consumers have L3 or even L4 expectations for L2 products?
In 2020, when many OEMs can't even do the AEB of L1, they will start to brag about L3.
In the first half of 2020, the author said the following when arguing with others about L3:
Even if L3 is technically feasible and legally permitted, what value can it provide users?
Under normal circumstances, human beings rely on "muscle memory" to drive, and many actions are completely subconscious behaviors, which can ensure safety even without using their brains; but L3 only frees their hands, but does not free the driver's brain. Even after freeing their hands, The driver's muscle memory and body inertia can't play a role, so the nerves will be highly tense, and it will be much more tiring than driving by yourself (system reliability is 99.99% and 80%, for the user, it takes the same mind, if He wants to make sure nothing goes wrong).
If the above analysis is true, the user experience of L3 will be much worse than that of L2. To have a good user experience, users have to be allowed to "let their guard down," but that's where the danger lies. The best state is that the driver has turned on the automatic driving function, but he forgets that the system will "take over" when he encounters danger.
When car companies can't handle L4, they use L3 as a selling point, which will make the market feel that they are better than those companies that only provide L2. But to determine whether there is a market for L3, the research object may not be ordinary white people, but algorithm engineers in the autonomous driving industry, especially engineers or safety officers who have participated in "L4" road tests, to see if they will buy L3 cars , if you will buy it, is the purpose just to "experience it", or will you use this function often?
If these people who know about autonomous driving are Ye Gong’s favorite dragon or three-minute enthusiasm for L3, then L3 will become a tool for car companies to use information asymmetry to “cut leeks”.
Let's talk about the harvested users.
The inherent weakness of human nature is that between being a "reliable person" and an "unreliable person", there are always a large number of people who are more inclined to choose "being an unreliable person" (because of gullibility and relax their vigilance), because it is more comfortable; even if unreliable may lead to disastrous consequences, they will still choose "unreliable" under the control of luck.
The result is that no matter how many car companies make repeated orders that "attention cannot be liberated, you must always be prepared to take over the vehicle", but there are still many users who will take it lightly-from their standpoint, it is easy to understand, "I spend more Isn’t tens of thousands of dollars just to be more willful and indulgent when driving?”
You can think of these people as "stupid Xs", but you cannot ignore the existence of these "stupid Xs" when thinking about the feasibility of L3. After the L3 is on the road, as long as there are three or five "stupid Xs" among the car owners, after a few accidents, the public's trust in autonomous driving will suffer a severe setback.
Probably starting from the second half of 2020, domestic car companies have realized that it is easy to get burned if they brag about L3 again. Therefore, they began to use the English name (NOA/NOP/NGP/NOH) in PR instead of L3.
Although many people in the industry believe that TJA belongs to L2 and NOA belongs to L3, strictly speaking, there is no necessary correspondence between the two, because TJA and NOA emphasize "application scenarios" and say "where to run ", while L2 or L3 refers to the problem of "who will be held responsible if something goes wrong".
NOA is translated as "navigation assisted driving" in China, and this translation is quite "chicken thief". Because, on the one hand, car companies hope that consumers will think that "NOA is L3", that is, the auxiliary system, and lure car owners to "free their hands". L2 under standard.
In addition, in the PR of "navigation-assisted driving", "manual takeover" is a word that appears frequently, but this statement is nonsense, and it even has the suspicion of seriously insulting the readers' IQ.
As long as it is "assisted driving", no matter how advanced it is, the division of driving responsibilities is "system assisting people", not "human assisting systems". Specifically, when I accidentally doze off or take an emergency call and get distracted, the system can temporarily "save my life" instead of promoting "hands-free" and letting me hand over the responsibility of driving to the system. When it fails, I go to "save its life".
Therefore, in assisted driving, he should be driving more than 99.9% of the time; for a person who should be responsible for "more than 99.9%" of driving tasks, you actually say that he "takes over", he should not be "always in charge" ? The accurate statement is that the system "takes over the vehicle" when he fails to drive well.
The purpose of using the word "manual takeover" in the promotion is to "lure" consumers to use L2 as L3. The result is the so-called "L3 if there is no accident, and L2 if there is an accident".
In fact, Japan also showed such a mentality when formulating the regulations on L3: it hopes that domestic manufacturers can seize this opportunity, but does not want to put too much pressure on them; it also encourages consumers to "try boldly" ", and feared that he was "too bold".
Many people have not noticed that the L3 referred to in Europe, America and South Korea is also called "single-lane automatic driving", which is actually "lane keeping". Do you understand L1 better? But China does not seem to have many such roads.
In addition, L3 requires the automatic driving system to have redundancy in all aspects such as perception, positioning, decision-making, and execution. However, according to the product director of an automatic driving solution company, currently only ESP And EPS adopts a redundancy scheme, and there is no redundancy in other links.
Under the condition that the main engine factory dare not take responsibility and the system redundancy is not in place, according to the definition of SAE in the United States, L3.9=L2.9=L2. What's the use of blowing up those bells and whistles?
Now, everyone is talking about "consumers haven't experienced the value of autonomous driving, so don't want it anymore". The higher the expectations, the greater the disappointment.
If you manage user expectations well from the beginning, users will still think that "autopilot is valuable".
Some time ago, a friend asked me: "If you are a consumer, what is your demand for autonomous driving?" I said, "I need an L5 ADAS." Everyone laughed knowingly. Yes, if it is clearly stated that it is ADAS, consumers will recognize its value; if it is said that L3 "liberates hands", consumers will think that "automatic driving is useless".
If it exceeds expectations, it is also "good" if it is not good; if it does not meet expectations, it is also "bad" if it is good.
At the end of March, Wang Chuanfu talked about unmanned driving at BYD's 2022 annual report exchange meeting. In the end, it is an advanced assisted driving." At that time, a friend asked me what I thought, and the author replied: If it is only for passenger cars, I agree with Wang Chuanfu's point of view 100%.
I don't think Wang Chuanfu's words are "cracking down on the custom driving industry". On the contrary, I think it is a rational, responsible voice that is more conducive to the long-term healthy development of the industry.
END
The third video of Nine Chapters Observation is officially launched
Welcome everyone to pay attention , forward , like and comment! !
Communication group| Enter "Sensor Group/Skateboard Chassis Group/Automotive Basic Software Group/Domain Controller Group", please scan the QR code below, add Jiuzhang Assistant , be sure to note the name of the exchange group + real name + company + position (no remarks Unable to pass friend verification)
write at the end
About Contribution
If you are interested in contributing to "Nine Chapters Smart Driving" ("knowledge accumulation and sorting" type articles), please scan the QR code on the right and add staff WeChat.
Note: Be sure to note your real name, company, and current position when adding WeChat, thank you!
Quality requirements for "knowledge accumulation" manuscripts:
A: The information density is higher than most reports of most brokerages, and not lower than the average level of "Nine Chapters Smart Driving";
B: Information needs to be highly scarce, and more than 80% of the information needs to be invisible on other media. If it is based on public information, it needs to have a particularly powerful and exclusive point of view. Thank you for your understanding and support.
Recommended reading:
◆ Nine chapters - a collection of articles in 2022
◆ "Your budget is too much, so we can't cooperate" - Jiuzhang Zhijia 2nd Anniversary Review (Part 2)
◆ What is the comprehensive SOA-based electrical and electronic architecture?
◆ Application of deep learning algorithm in automatic driving regulation and control
◆ Challenges and dawn of wire control shifting to mass production and commercial use