news

Today'sZeekr, is experiencing the most difficult period since the establishment of the brand.

As for the cause of the incident, I believe most readers are already very familiar with it, so I will not elaborate on it. In contrast, what I want to discuss more is, "Why did this new force in car manufacturing make such a big fuss and pay a high price to putZeekr 001Will the intelligent driving system switch to a self-developed solution? "

The answer is very simple, "It is the general trend and we have to do it."

In fact, after the controversial press conference, Zeekr conducted an exclusive media interview, in which An Conghui, the helmsman, emphasized: "Some decisions are really difficult, but we must pass the test of self-developed intelligent driving."

The implication is that as the brand's main sales unit, Zeekr 001 has always had major "shortcomings" in high-end intelligent driving, and has been criticized by many users.

Even though the delivery performance of the 2024 model has been good in the six months since its launch, it is in the most competitive 200,000-300,000 yuan market segment and must remain vigilant at all times and continue to make improvements in order to further consolidate its slight advantage.

So, the 2025 Zeekr 001 suddenly came.

Taking advantage of the situation, the operation of this new car-making force once again confirms a truth: "This year, high-end intelligent driving has undoubtedly ushered in a thorough blossoming."

For more and more potential consumers, the technical strength and user experience of the OEM in this dimension have a real impact on their car purchase decisions. To put it more bluntly, without increasingly mature high-end intelligent driving, what smart electric vehicles are there to sell?

But it was precisely during this process that a number of "empty fires" emerged.

In recent times, if we try to select the most valuable promotions related to "advanced intelligent driving", they will definitely be firmly occupied by "NIO, Xiaopeng, Li Auto and Huatai".

On July 5, Ideal took the lead and held a summer intelligent driving launch conference, revealing all of its core technologies and demonstrating its determination and ambition to leap from a recognized poor student to the first echelon.

On July 27, the annualNIOThe "Technology Innovation Day" was held as scheduled, and this new car-making force came to the fore with the "World Model NWM". At the same time, many unheard-of concepts were mentioned, including spatial cognition, time cognition, GOA, group intelligence...

July 30,XiaopengTaking the opportunity of the "AI Intelligent Driving Technology Conference", it announced that it will push the AI ​​Dimensity system XOS 5.2.0 version to all users around the world. Through this update, XNGP will be officially upgraded from "can be driven nationwide" to "easy to use nationwide", realizing "no restrictions on cities, routes, or road conditions" nationwide.

On August 6, the "big devil" Huawei made its final appearance, andEnjoy S9Also launched at the same time was HUAWEI ADS 3.0, which showed everyone what it means to "not make big promises but just implement them directly."

When it comes to high-level intelligent driving, the top priority that "Wei, Xiao, Li and Hua" tries to spread and demonstrate is, of course, end-to-end, which is rapidly changing the rules of the game like a spring breeze overnight.

This year, the high-end intelligent driving has become so popular, which is definitely related to its promotion. In order to embrace end-to-end, OEMs are willing to pay a lot of money, start to adjust the original organizational structure, and even hype up the marketing like a demon.

So, what exactly is end-to-end?

There are endless long articles and professional interpretations on this topic. To put it simply, the advanced intelligent driving that used to use perception, decision-making, and execution is like a student learning to drive in a driving school. How to drive, where to drive, and how to drive all require the instructor's instructions, which makes them seem very dull.

Switching to end-to-end advanced intelligent driving is like a novice driver who has obtained a driver's license and is already on the road. He will gradually be able to cope with most road conditions. As he drives every day and every month, his driving skills will continue to improve.

Breaking it down further, "end-to-end" actually gives high-level intelligent driving a so-called "brain", allowing the vehicles equipped with it to have the ability to think proactively and adapt to changes, rather than being as rigid as a machine as before.

As another dimension of evidence, in March this yearTeslaOfficially launched FSD V12 in North America.

It is worth sharing that this version of the model architecture has undergone a major iteration. It is a new end-to-end application that allows advanced intelligent driving to continuously learn and improve like humans. From Musk's promise, "Wide application will further reduce accidents caused by human driving errors and improve the active safety performance of vehicles."

In the end, the interim results proved that compared with the previous version of FSD, V12 increased the proportion of Tesla users' completely unattended trips from 47% to 72%, and the average takeover mileage increased from 116 miles to 333 miles.

The effect is immediate.

In short, whether it is the crazy internal competition of "Wei, Xiao, Li, and Hua", or Tesla's full bet, the value of end-to-end is still skyrocketing. Gradually, a trend has emerged in the entire industry: "If you don't talk about end-to-end, what's the point of talking about advanced intelligent driving?"

At the beginning, I mentioned Zeekr's expensive product launch. I think the fundamental purpose was to avoid being the first to fall behind in the technology competition related to the future. As more and more OEMs and suppliers joined in, a secret war without gunpowder slowly began.

The emergence of end-to-end has undoubtedly ignited the flames of high-end intelligent driving. Of course, the emergence of any new technology will inevitably bring a lot of pain and chaos.

It is true that some people are leading the times, but there are always people who take advantage of the situation.

“Publicity: We will soon mass produce advanced end-to-end.”

“Engineer: But we haven’t started yet.”

"Propaganda: It doesn't matter, and you can't dig into the code. No one will find out."

"Engineer: What should we do if we find out?"

"Propaganda: It's OK. Our modules are end-to-end, including perception, planning and control."

"Engineer: Do you want to hear what you are talking about?"

"Publicity: It doesn't matter. It doesn't matter where it is sold. What's more important is to occupy the publicity high ground. No matter how bad the sales are, everyone will be affected."

A few months ago, the "joke" at the beginning of this paragraph went viral. Everyone is welcome to identify who the target of this Weibo blogger's satire is.

What I want to say is that the sudden rise of end-to-end this year has indeed injected sweet spring water into high-end intelligent driving. However, many brands have exaggerated their marketing. Now, for some reason, an increasingly exaggerated atmosphere is rapidly enveloping this sector.

Extremely radical slogans, test videos flying everywhere, and the radical media's crazy pace have all misled many end consumers.

Without clarifying the safety boundaries of advanced intelligent driving, people began to blindly believe that it could drive like a human, and even thought it was infinitely close to autonomous driving. In the end, a variety of dangerous cases occurred, leading to irreversible tragic events.

Although, from the perspective of major OEMs, the urgency of winning this secret war is understandable, and no one does not want to be the first to grab the tempting "cake", but when it comes to high-end intelligent driving, which is a matter of life and death, shouldn't we sometimes "slow down" a little so that every step we take is more solid and with a clear conscience?

In fact, I am a deep and advanced intelligent driving user. As an automotive media, I can always get the latest versions of various companies first by taking advantage of my position.

But the strongest feeling after long-term use is, to borrow the words of a friend in the same industry: "This thing must be driven a lot. After driving 10 kilometers, it may give a rating of 95 points, but after driving 30 kilometers, it may only give 85 points, and after driving 50 kilometers, it may only be 80 points, and after driving 100 kilometers, it becomes 75 points."

In other words, mileage accumulates gradually, but points decrease rapidly.

Not long ago, I encountered an accident in Shanghai when I turned on the "pilot assist" of a popular car of a certain brand. The fundamental reason was my own distraction, but after experiencing such an incident, I realized that I should never trust any high-end intelligent driving system too much.

In addition, I would like to emphasize that: "At present, after an in-depth overall experience, the high-speed navigation assistance functions provided by several recognized leading echelon members can indeed reduce the burden and pressure on drivers while ensuring a certain cruising efficiency, allowing everyone to feel the inclusiveness brought by technological progress."

In sharp contrast, "once we get to the urban area, the road conditions become twice as complicated and emergencies occur more frequently. The papers submitted by these top echelon members basically fail in my opinion."

This is by no means a deliberate smear. Every time I turn on the high-level intelligent driving in the city, I always pay more attention than when driving by myself, for fear that I will make a big mistake due to improper operation of the vehicle.

Moreover, in terms of commuting efficiency, it is far inferior to driving by yourself, and it can even become a traffic obstruction during certain morning and evening rush hours.

In short, no matter how fantastic the promotional statements are, the major OEMs have only solved the problem of "being able to drive" so far, and there is still a very long way to go before they can truly be "easy to drive".

Next, if someone around you over-mythologizes advanced intelligent driving, then please ask: "Is it really useful?"

At this point, the article is gradually coming to an end. The last thing I want to clarify is that the essence of writing this article is to put out the "false fire" in this sector. I always feel that the high-end intelligent driving that has been hyped up should return to a more benign and healthy development model. But on second thought, according to the current competitive situation in the Chinese auto market, it is almost impossible.

So, I urge everyone to remember: "All high-level intelligent driving is assisted driving, keep the steering wheel firmly in your own hands!"

- END -