From the rapid rise of autonomous driving in the summer and autumn of 2016, to the cold winter of 2019 for related companies, there is about 2-3 years of room for operation.
According to different levels of intelligence, autonomous driving can be mainly divided into six levels, from completely unintelligent to fully intelligent:
L0 level autonomous driving: There is no automatic throttle, brake and steering wheel, and the entire process is controlled by the driver.
L1 level autonomous driving. The system will intervene under certain circumstances, such as assisting in functions such as accelerator and brake.
L2 level autonomous driving. For example, Tesla's autonomous driving system can control the speed and steering of the vehicle under certain conditions.
L3 level autonomous driving. In most road conditions, the driver's attention does not need to be focused on the road. But both hands must be on the steering wheel. This level of autonomous driving is basically the limit of major companies by 2023.
Moving forward to L4, humans basically do not need to intervene in driving at all, and only need to intervene in rare cases.
As for L5 level autonomous driving, humans become completely passengers. There is no need to even install a steering wheel, accelerator and brake on the vehicle, and the vehicle is completely treated as a subway.
In the eyes of the capital circle, AI technology has continued to make breakthroughs in 2016, and has surpassed the peak human brain in such a complex game as Go.
So in the direction of autonomous driving, let alone L5 level of complete automation, at least L4 level, it should be very easy for the driver to not have to pay attention to the road conditions at all times, right?
With this kind of blind optimism, companies related to autonomous driving have received overly high evaluations in those years.
Many related companies are unable to make ends meet, and after burning money for several years, they simply cannot produce products that meet expectations.
For example, argo.ai, which received more than $5 billion in investments from Volkswagen and Ford, claims to be building an advanced L4 system but has not even seen a single cent of it.
After all the money was burned, he just spread his hands and declared bankruptcy.
The L4-L5 route has come to nothing, and the industry market has become calm and rational. Car companies have focused on L2-L3 levels, the popularity of autonomous driving has also dropped a lot, and capital bubbles have burst.
"Currently, driverless driving is okay as a dream, but not as a business or enterprise. Technology cannot determine everything, and advanced technology may not necessarily solve market problems."
After ten more years of observing the development of autonomous driving, Meng Fanqi concluded that it is far from enough for autonomous driving to rely solely on advances in vision algorithms and bicycle technology.
It requires the transformation and reshaping of the entire urban transportation system, and overall terminal control and planning to be truly realized.
Without changes at the transportation system level, the moral and ethical issue of autonomous driving will never be avoided, that is, whether the algorithm should be responsible for traffic accidents.
If vehicles and pedestrians are not separated, this problem is estimated to be completely unsolvable for a long time.
Doing autonomous driving still requires a mature business plan, including the cost of recalls, etc., rather than a simple sales behavior. Meng Fanqi will only invest in this field in the last six months to a year, and will not pay too much attention.
However, Meng Fanqi knew very well that Baidu had already launched research and plans related to autonomous driving in 2013. Robin Li would not like or listen to any opinions he expressed rashly.
There is no need for me, and I have no position to remind the other party. In fact, it is always a good thing for entrepreneurs like Li Yanhong to be willing to invest tens of billions in this kind of technology.
It’s much better than doing takeout all day, using apps related to shopping for food, and exploiting the money of small restaurants and farmers.
There wasn't much video footage to be shot with Robin Li. It took about ten minutes and several shots from multiple angles of the two of them talking and discussing formulas on a whiteboard. What Meng Fanqi needed to do here was completely over. .
Li Yanhong didn't have many follow-up arrangements today, and the two of them hadn't seen each other since they signed the contract last time, so they chatted a little more.
"My plan here is that after shooting these materials today, the promotional video will be released the day after tomorrow at the latest." Robin Li's work style has always been very decisive, "We have tested extensively on different platforms and hardware in the past few days. Let’s look at the performance and adaptability of the algorithm. The algorithm unit you use is quite hardware-friendly, so there is generally not much work we need to do here.”
"Did you design it like this intentionally? Or did you do it unintentionally?"
"I did consider this limitation during the design." The version Meng Fanqi memorized was naturally the result of many choices. He shamelessly took the credit without blushing. "If the promotional video will be released tomorrow and the day after tomorrow, what is the specific time of the technical conference?"
"It will probably be delayed for another three or four days to give the market and the public a few days to buffer. Our publicity and marketing department will also try to stir up the excitement and try to mobilize everyone's attention." Li Yanhong is determined to invest heavily in this aspect. resources, because privately he has contacted some large leading companies with testing materials and specific parameters.
Without exception, the other party was shocked by this result. Whether they are familiar with artificial intelligence or not at all, they are amazed by the performance and speed of this result.
Most importantly, in their view, this kind of intelligence can greatly improve efficiency and save costs.
The feedback from these giant companies was beyond Robin Li's expectations and exceeded his expectations.
Therefore, regarding related publicity and releases, Robin Li has frequently urged and decisively stepped up efforts in recent days.
When I saw this young man again today, I found him pleasing to the eye.
"Let me tell you in private, the people above are also paying close attention to this technology and think highly of you." Li Yanhong put a hand on Meng Fanqi's shoulder, pulled him to the corner of the room, and whispered.
"Above? You mean... the official?" Meng Fanqi probably understood Robin Li's hint. He knew that Robin Li had been cooperating closely with the government, but he did not expect that it would attract official attention so soon.
"I won't say much now. I value the potential of this technology in security and military aspects." Robin Li smiled mysteriously, "You continue to do your research well, don't think too much, don't be nervous, this is a good thing. Related If they need it, they will contact Baidu. If they really need to know something from you, they will definitely contact you directly. If they don't contact you, nothing will happen."
"What should I say? When the press conference comes, would you like to come up and talk about the algorithm, principles or the story behind it?" Li Yanhong finished talking about the plan and sent out the invitation.
In fact, he told Secretary Liu yesterday that Meng Fanqi would be attending the meeting and would not have time to come to the press conference, but noble people often forget things, and today they forgot everything.
