Home > News content

Baidu, which can't keep talented people, auto-driving at crossroads

via:博客园     time:2019/3/20 12:32:59     readed:161



After Lu Qi left, Zhang Yaqin took over Apollo. In a twinkling of an eye, Zhang Yaqin also retired from Baidu in October. Who will take over the bowl this time? On the road of not retaining talents, Baidu is becoming more and more anxious. Where to go?

Spring River Water Warm Duck Prophet


If Google's autopilot project comes from the technology presentation of mature challenges, geeks like Page are keen to tamper with these novel technologies. Baidu's choice of automatic driving is more from the strategic layout and prospects of its business.

Baidu began to lay out AI in 2010, and devoted its research and development to machine learning and natural language processing. In fact, it was not until 2013 that Baidu put AI on the stage and officially announced the establishment of an in-depth learning institute to launch the unmanned vehicle project. In the following year, Robin Li invited AI star scientist Wu Enda.

It was also in 2014 that Google announced the first prototype of a driverless car, Firefly, designed and manufactured from scratch. After that, the mountains and rains are about to blow all over the building. Uber's CEO, Karennik, experienced a Google drone, quickly drew 50 scientists from Carnegie Mellon University to form Uber's Autopilot team, ATG; on the other hand, Madman Musk launched Tesla Autopilot.

During this period, Baidu's business transformation has been a bit overwhelming, and the full-time bet on the O2 O encountered a cold winter bankruptcy tide. Later, the evil of search engines plunged them into the whirlpool of public opinion. Robin Li is a little nervous, and AI is the new antidote. Autopilot is the main ingredient of this antidote. Baidu then set up an autopilot business department in late 2015, under the responsibility of Wang Jin, and formally stepped into the autopilot industry. Little did it know that this action also buried the hidden danger of the first case of domestic autopilot disputes.

The direction of the wind has changed, and only the key people who know where the wind comes from are at these key points.

Baidu Runaway and Rising Domestic Autopilot

Baidu, a toddler, drives its own car, but it's a bit out of line in its growing period. Robin Li was horrified to find that Baidu had fallen into a trap. From 2015 to 2016, Baidu's automatic driving team ran away with a lot of technical personnel.


The first one to leave was Yu Kai, the executive vice president of Baidu Research Institute. He failed to catch up with the establishment of Baidu Autopilot Business Department. He started Horizon Robot with two scientists of Baidu Deep Learning Research Institute: Yu Tienan and Huang Chang. The goal is to provide high-performance, low-power, low-cost autopilot software and hardware solutions.

Most of these people leave from 2015 to 2017, which is also the climax of domestic self-driving entrepreneurship. Among them, 2016 was a turning point. The battles of Alpha Go and Li Shishi in March of that year made AI World War I popular. GM's $1 billion acquisition of Cruise Automation, which had only a few prototypes and more than 40 people at that time, made investors squeamish.

Data show that between 2015 and November 2017, there were 193 investment and financing related events in the field of autopilot, amounting to $143 billion 800 million, accounting for 87% and 97% of the total number of events and sums since 2009, exceeding the total market value of General Motors and Tesla. And Baidu's self-driving start-up company is also the fragrance of the capital market.

These people have achieved the upsurge of self-driving entrepreneurship in China, and indirectly promoted the development of self-driving. If Google is an enlightener to Baidu and Uber, Baidu is also an enlightener to them.

From high-precision maps, lidars, autopilot chips to a whole set of solutions, the technicians who leave Baidu have also found a clear direction of subdivision. They are not persistent in doing Android system in autopilot like Baidu.

On the other hand, Luqi readjusted Baidu's internal organizational structure and merged the original Auto Driving Division (L4), Intelligent Automobile Division (L3) and Car Life etc. into Baidu Intelligent Driving Group (IDG). However, even Baidu's self-driving heroes, Lu Qi still can not escape the magic spell of Baidu.

When Baidu puts auto-driving into its core strategy, those Baidu people who run away are also living in their respective battlefields. Baidu's self-driving entrepreneurs have not since been associated with Baidu's old death. For example, Zhang Tianlei's main line technology appeared at Baidu's developers'conference last year and cooperated with Apollo. However, Baidu is still on the opposite side of the competition, Baidu's large and complete platform will inevitably have some details overlap with their business.

However, there are extreme cases of tearing faces. At the beginning and end of 2017, two lawsuits attracted great attention. Waymo sued Uber/Otto at the beginning of the year, and Baidu sued Jingchi and Wang Jin at the end of the year. Both cases ended when the founder of the defendant company was forced to leave.


Changes in Google's Driverless Personnel

Where should Baidu go if it repeats the same mistakes?

Faced with the departure of the veteran of autopilot technology, Google chose to break the boat and make the project independent. In December 2016, Alphabet, the parent company of Google, split the autopilot project from Lab X and set up Waymo, an autopilot company.

The same problem lies in front of Baidu. Baidu chose to bear it silently.

The question arises. Why does Baidu suffer the same brain drain as Google? And those who leave large companies are choosing to return to self-reliance in the field of self-driving?

An interview with Wang Jin made the answer come out. In the article "Wang Jin: I'm different from Baidu", Wang Jin takes Google's split Waymo as an example, advocating Baidu to follow the same path, dividing the autopilot business into separate companies.

Various factors make Wang Jin choose to split up with Baidu.

It is undeniable that large companies will inevitably encounter problems of corporate management and decision-making in the process of trying out some new businesses and transforming. This contradiction is particularly prominent in the talent-centered AI industry. Google has the courage to choose a strong man, Baidu can only make up for the new talent blood more urgently. At this point, to see which model is better failed to reach a final decision.

However, back to the industrial level, Baidu's automatic driving route is open source, with a big basket to put everything in the industry. This route is quite different from Google's, and naturally there are differences with some people's understanding of automatic driving. Those who leave Baidu are more willing to choose a vertical professional direction and strive for one place.

Of course, for the emerging technology of self-driving, the environment of start-up companies is easier for aspiring people to let go of their hands and feet. However, looking at it from a different angle, technology entrepreneurship is open to everyone who has the ability, and everyone who passes through the outlet has the chance to succeed.

This is the dividend brought by technological iteration. It is also why Robin Li said that the Internet demographic dividend has disappeared and the future growth momentum is technology.

But in turn, the dominant factor behind the technology dividend is still people. In February this year, Baidu made a new attempt to announce that three vice-presidents, Shen Jie, Wu Haifeng and Zheng Zibin, should be adjusted in cadre rotation to train and reserve compound managerial cadres. Zhang Yaqin is retiring after announcing the talent echelon construction and senior executive retirement plan last week. Another key autopilot has left. I wonder if Baidu can do a good job in talent construction during the window period.

China IT News APP

Download China IT News APP

Please rate this news

The average score will be displayed after you score.

Post comment

Do not see clearly? Click for a new code.

User comments

Related news