"Nvidia is going straight for a fully autonomous vehicle"

Jeff Herbst photo: Amir Meiri

Nvidia VP business development Jeff Herbst sees AI taking over everywhere, and Nvidia's platform leading the charge.

No "Globes" interview with a senior manager from a global company goes without the question "Is this your first visit to Israel?" In the case of Jeff Herbst, vice president of business development at Nvidia for the past 17 years, the number of visits, about 20, exceeds the number of years he has been in his position.

By virtue of his position and long experience, Herbst has a rare perspective on the world of technology, twenty years back and twenty years ahead. Few senior managers are prepared to prophesy decades ahead, but Herbst, who came to Israel to promote the annual Nvidia GTC 2018 Israel developers' conference that the company will hold on October 17th and 18th, confidently produces long-term forecasts.

"Artificial intelligence is penetrating every industry and every enterprise. Before long, there won't be any such thing as an artificial intelligence company," he told "Globes." "It's like what we saw with the Internet boom 20 years ago. They talked then about Internet companies, but today there is no such thing, because the Internet is part of the fabric of every successful company. This is exactly what's going to happen with artificial intelligence."

He says that the reason is that "Artificial intelligence essentially enables software to write software. Instead of people writing code, you use the data to program. This technology is so strong and revolutionary that every company, developer, and cloud computing provider will have to assimilate it in some way.

"We're only at the beginning of the artificial intelligence revolution," Herbst says and immediately puts Nvidia front and center. "We believe that our technology can become the most important computing platform in the world. We no longer think of ourselves as a chip company. Our vision and goal is to become a very influential and important company in this sphere."

"Globes": More than Intel? "It's hard to answer that," he answers, and goes back to talking about his company. "We have a computing platform called CUDA in which we've invested $1 billion in research and development annually. Now we want to take this amazing technology and install it in every industry. We feel that the company is in a good position to do this. We have turned a company of graphic cards and semiconductors into a platform for visual computing, artificial intelligence, and autonomous cars."

Nvidia is also aiming high in the autonomous car field. "Our strategy is to use artificial intelligence and deep learning to develop the future autonomous car, but our aim is the use our technology to reach a fully autonomous car. We won't begin with stage 1 or 2; we'll go straight to stage 5 - the most difficult one," he declares confidently and immediately adds disparagingly, "To best of my understanding, that's a little different from what Mobileye is doing." In the context of the autonomous car business, stage 5 means that a car drives itself completely independently, in contrast with stages 1 and 2, in which the technology aids a human driver.

Meanwhile, Nvidia's hardware is popular in a different sector that the company never expected to specialize in: digital currency miners are using its graphic cards in their work; in the past year, they have quickly emptied the shelves. This situation will be frustrating for gamers, Nvidia's usual customers, so despite the revenue potential of the new market, the company has come to their aid by limiting sales of its cards. "Currency mining is not something we have focused on; it's a byproduct of what we built," Herbst says. "In the end, everyone uses GPU and the customer decides for what purpose. The same technology used for gaming is also used for an autonomous car and for data centers, so someone discovered that GPU and parallel computing are also suitable for mining. We have no way of identifying these users and it won't be the only application that GPU is used for. There are other applications that involve blockchain, but have nothing to do with currency mining. We want to cover a wider range of applications, with an emphasis on artificial intelligence."

Ambassadors in universities

Herbst, an ex-lawyer and a graduate of Stanford University, does not wait for questions; without prompting, he concisely and precisely relates the company's history and its metamorphosis. "This company, which began as a manufacturer of graphic and gaming cards, has become something completely different - a company that develops graphic cards in which artificial intelligence is used. The change took place at a time when the acceleration rates of ordinary central processing units (CPUs) were slowing. Moore's Law says that multiplying transistors will multiply computing power, but it eventually ran into the limitations of physics, and this type of processor no longer produced the exponential performance advances that it once did - instead of multiplying its computing capability, as in the past, this type of processor now produces increases of only 5-10% every two years."

"The world, however," Herbst continues, "is being filled up with large quantities of information, with people who have to conduct serious analyses, and with people who want to build artificial intelligence applications. The computing demands are super-high, so graphic processors have become the platform of choice for fulfilling them. Once upon a time it was odd to put a graphic processor in a data center, but today, all of the major corporations are installing graphic processors in their data centers, or at least seriously considering it. The same is true of all the cloud computing providers, who are joining the auto designers, cinema designers, and architects."

Herbst's visit to Israel is part of Nvidia's effort to become the global leader in artificial intelligence (AI). At his side is Liron Freind-Saadon, Nvidia Israel's director of developer relations. Freind-Saadon is responsible for fostering the ecosystem of AI developers in Israel and their connections with Nvidia, the company that is still called the graphic processors giant and is less identified with the growing AI field. "Nvidia has realized that someone has to go and foster an AI ecosystem here - talk to the developers, understand their needs, work with them, and connect them directly to engineers or between them and a large corporation in need of a specific solution that they can provide," he says.

Nvidia is searching Israeli universities for what it calls "ambassadors" - potential AI lecturers. Herbst explains that usually, these are doctoral and masters students with the right knowledge who can communicate it to people outside the computer science and mathematics faculties "so that they will educate students who would not have received such knowledge in other circumstances. We want to educate the entire ecosystem so that people will implement deep learning solutions in every industry."

At the same time, Herbst and Freind-Saadon are aiming to increase the proportion of Israeli startups with which they cooperate from 70 out of 2,000 startups worldwide in 2017 to 150 Israeli startups out of 3,000.

Published by Globes [online], Israel business news - www.globes-online.com - on August 6, 2018

© Copyright of Globes Publisher Itonut (1983) Ltd. 2018

5 Comments
View comments in rows
Update by email about comments talkback
POST
Comments
Jeff Herbst photo: Amir Meiri
Jeff Herbst photo: Amir Meiri
Twitter Facebook Linkedin RSS Newsletters גלובס Israel Business Conference 2018