Intel’s Chief Technology Officer speaks to Balqis Lim on how data and artificial intelligence will drive the digital economy
DATA has become one of the most vital resources of the century. To help consumers prosper in this new digital economy, the world’s largest manufacturer of computer chips, Intel has shifted its focus from being a PC-centric company to a data-centric company.
Its Chief Technology Officer Dr Michael Mayberry says in today’s world, cameras and road sensors are collecting data, and this data has to go somewhere for things to be done computationally.
“How do you make sense of the data? How do you turn it into useful information, whether it’s a map of traffic congestion or of smog in some cities around the world?
“All these are examples of things where you start with raw data. You then turn the data into information, and that information is then used to make decisions. It’s important that we consume the information as quickly as possible to make sense of it. It’s no longer a person-centric thing, it is a data-centric thing,” he says.
Mayberry says when he started working at Intel 35 years ago, it was primarily a memory company, although the company had already started manufacturing central processing units (CPU) at some point.
In the 1980s, Intel made a transition, going from memory centre to a CPU centre. In the ‘90s, the transition was from CPU-centric to PC-centric. Today, the company is in the midst of a transition from PC-centric to data centre.
“What we mean by that last transition is that PCs are an example of something that’s optimised for humans to interact with. It’s a personal computing device. Data-centric means increasing the amount of data that we collect and we need to make sense of it. We then use the data to make decisions, sometimes in real-time,” he says.
Mayberry says the data collected comes in multiple forms, for example, from a video that was shared as a memo to the staff.
“It could be a video camera at a corner intersection. It could also be a sensor collecting information about smog, or the tides, or any of those things. What we want to be able to do is to make sense of all that by combining a lot of pieces of information together. So it’s not simply one piece of information, it’s making sense of all the data we collect all the time throughout the day.”
And as more data is collected, he says there is a need to store it or move it somewhere and a need to do computation around it.
“For example, when your phone’s storage becomes full, you have to offload those pictures to the cloud. That requires network, so that’s the movement of it. The network then may say, ‘hey, I’m going to tag the faces in those pictures’ and that’s compute. All three of those things work together to make sense of the world,” explains Mayberry.
Depending on which piece of product Intel is working on, the company also plays multiple roles. Having driven the PC ecosystem for a couple of decades, Intel is also into standardisation so that the machines can talk to one another.
“We work on that standardisation around how people use technology. We also work on standardisation around how to compute with technology. For example, on artificial intelligence, we have an initiative that’s called OpenVINO that allows people to write programs that make sense of data and do a control of the systems.
“It’s intended to make it more accessible for people to get into artificial intelligence without having to go to school and get a degree and train for years. We do things like that to make our technology easy to use.
“At the same time, we’re making technology more complicated. We’ll build a faster computer, a larger memory device, a faster network, and those, in turn, will require more choices and so we have to work on making those easier to use as well.”
Intel has also embarked on Project Athena, an innovation programme to aid laptop makers meet modern-day computing needs and build revolutionary laptops for the future.
“A decade ago laptops were typically between 1.3kg and 1.8kg, which was considered very light. We are trying to enhance the features to make it convenient to carry around, as well as preserve the things that make a PC useful to people.
“So we’re going to have longer battery life, ease of connectivity, and standards around the display, as well as around the responsiveness. Ultimately, we would like devices that are still as useful as the PCs were, but far more attractive where portability is concerned. So we work with the ecosystem to essentially certify those designs,” says Mayberry, adding that the first batch of these premium products have begun to enter the market.
FUTURE OF COMPUTING
Mayberry says one of the directions going forward is AI, to handle a large variety of different kinds of data, computing tasks, and sharing of information. “For example, artificial intelligence is one example of things where if we put a particular kind of compute capability in our chips, then it becomes much more useful for that kind of data.
“So as we have been modifying and refining the performance of our computer chips on traditional tasks, we’re also adding instructions. We’re adding compute units, we’re adding software layers to get ready for that future of computing.”
Another direction is increasingly rapid communications, or as Mayberry calls it, 5G. However, he says it is a collection of technologies. There are increasing ways for machines to become autonomous like automating the factory floor.
He says it can be automated to handle more variations such as the orientation of parts or doing adjustment processes in real-time because humans don’t necessarily want to do tedious, unsafe, repetitive tasks. “We think the robots will be able to do better going forward.”
Mayberry says Intel is also working on its 5G protocols to handle a combination of data types, with shorter latency and faster response.
Rather than using cell towers that are far apart, it is using micro cells to have higher bandwidth and ultimately a more rapid exchange of information that will become one of the enablers for machine to machine communication.
As technology becomes more sophisticated, we will see more generational changes where one generation was raised doing one thing, while the next generation will see it as obsolete.
Mayberry says 2020 will not be the era of automated driving, meaning it won’t be broad enough to happen soon. “But my children’s children may never learn to drive because the world will have advanced so much.
“My predictions for next year will be like this year, only better. We’ll have smaller laptops that are lighter and with longer battery life because the electronics are already being developed right now and being shipped.”
As for Intel, he says the company is bringing a lot of 10-nanometre products, promising better performance and lower power, to the market.
“Some of those go into laptops, some into servers, and some into networking. We will have a much wider range of products this time next year than we have today.”