Legendary engineer and former Microsoft vice president Qi Lu returned to China in January this year to become CEO of Baidu. Wired magazine spoke to him about how the company plans to achieve AI leadership not only in China but around the world.
We have a platform we call Baidu Brain, which is powered by over 60 AI services. It is larger and wider than Google and Microsoft.
Does Baidu have a counterpart to Siri and Cortana?
The only difference is that in China we are the undisputed leaders. She has the most skills and partnerships. Duero’s has already been implemented by more than a hundred companies: it is found in refrigerators, televisions, air conditioners, speakers and other home devices. She speaks excellent Chinese. In absolute terms, Amazon is still ahead, but only due to a huge partner network in the United States.
Alexa, Echo and Cortana are adapted for American homes, which are unique: they have special acoustics designed for huge rooms. Therefore, they are not suitable for China, where everyone lives in small apartments. In Japan, India, Brazil and even Europe, people live differently than in North America. Therefore, Duero’s has more options in these countries. Google and Microsoft have the best technologies, but Amazon has bypassed them by creating a better ecosystem.
In the spring, Baidu announced work on the Apollo autonomous driving system. Why did the company focus on this area?
If you seriously want to excel in AI technology, you need to deal with autonomous systems. Today, a robotic vehicle is the first commercial application of such systems. The largest ecosystem is now built around the smartphone. In the future, the same will happen with the car: devices, sensors, chips, software will be produced for it. And then, on this basis, you can deal with robots, drones and other autonomous products.
When will autonomous vehicles hit the streets of Chinese cities?
Technology alone is not enough. While they are still far from complete autonomy, they must learn to read human handwriting and understand the language well. For full autonomy, new rules and laws are needed. But there are many opportunities for commercialization with partial autonomy. Apollo will allow the car to move in traffic jams, which make up most of the traffic in Shanghai, Beijing and other major cities.
Why did Amazon beat Microsoft in AI?
I personally worked on Cortana for four and a half years, but our mistake was focusing on the smartphone. At Baidu, we are experimenting at our headquarters, where almost all employee movement is based on facial recognition. We want to create a canteen where orders will be taken using AI, and people will not have to touch, drive in and count anything at all. Amazon figured it out before.
Have you managed to find a replacement for Andrew Woo in the field of AI?
There is much more talent in the US than in China, no doubt about it, but the gap is closing. Baidu will strive to create more and more technologies within China. But at the same time, we continue to invest in Silicon Valley and Seattle, where we opened a campus following the acquisition of Kitt.ai. The US is by far the best supplier of the top tier of talent, and we want to make the most of that.
The same goals as Baidu set themselves and the authorities of the Celestial Empire. China has unveiled a strategy for the development of artificial intelligence technologies in the country until 2030. The main goal is to become the world leader in this area and bypass the USA
AI will make oil and gas even cheaper
IT consultancy Tessellar is using artificial intelligence and big data to predict the corrosion of metal structures in oil wells, writes Adnet.
“Many of our clients have invested in AI after reading analysts and then they were disappointed,” says Ray Hall, director of energy at Tessellar. However, in his opinion, the industry is using outdated technologies, and new methods of analysis involve the introduction of machine learning to improve performance.
From a competitive perspective, it is important for oil and gas companies to overcome difficulties in adopting AI and other new technologies because the industry is facing challenges due to low oil prices. To cope with this problem, new technologies are needed. In particular, it is necessary to use robotics for drilling.
In addition, the industry is experiencing an aging workforce. “Up to 40% of the workforce in the oil and gas sector will retire in the next 5-10 years,” says Hall. “It will be very difficult to replace them, but machine learning, as well as AI, will help reduce reliance on employee experience when making decisions.”
Finally, as the global economy’s dependence on oil and gas decreases, companies must turn their businesses into full-fledged suppliers of renewable energy. Tessellar is helping oil companies implement AI, particularly for well corrosion calculations.
“The oil companies had a lot of accumulated historical data, but they couldn’t analyze it correctly,” Hall says. “With the help of AI, we tried to find connections in this data to find the cause of the corrosion.”
Self-driving dump trucks and robotic drills are already being tested in many quarries around the world. In the oil and gas industry, robots repair gas pipes deep underwater, and drones inspect offshore oil platforms.