The Intersection of Humans and Technology at the Edge | Intel Technology
- [Announcer] You're watching InTechnology, a video cast where you can get smarter about cybersecurity, sustainability, and technology. - Hi, and welcome to today's podcast. I'm your host, Camille Morhardt on InTechnology.
And today we're gonna talk with Joannie Fu. Joannie Fu is VP of Network & Edge Group at Intel Corporation. And we are gonna talk about the intersection of humans and technology at the edge. Welcome, Joannie. - Thank you, thank you for having me.
- Joannie, I think I met you when I hopped in the cab that you had pulled over, or maybe it was probably like one of the rideshare vehicles, I guess at the time. And just said, "oh, you're with Intel too? Well, like, can I'm going the same place you are, can I hop in?" And it was at a Lesbians Who Tech conference and you're like, "sure, get in." And then I found out you were, you know, kind of an important person and I was talking to you about it and you were just so friendly and genuine and nice.
And I've liked you ever since. So I'm really happy to have you on the podcast. - Likewise, I really enjoy your passion and your vision in regards to how diversity, technology, innovation all come together. I just remember that was a great experience for us. - You were also just fairly recently recognized as among the top 50 on the Queer 50 List in Fast Company.
So congrats on that. - Super proud and super humble. I think we need more of our representation within the community out there.
- We're going to meander back and forth between technology topics and a sort of human interest story here and importance of diversity and inclusion. So I want to start off by asking you how you got involved in technology. - Yeah, no, that's a great question. Actually, when I was thinking about the idea of intersectionality, humanity, of technology and everything that comes together, it really started with my childhood. I've been fascinated by technology and humanity and solving complex problems since then.
My father was the very first telecommunication engineer in Taiwan Telecom, which owns all the phone lines in Taiwan. So if you go back in history, 1945, 1950, that was when phone line was just started to be installed in Taiwan. So he was one of the groups that was in there. He took me and show how human voices as sound waves was translated into electrons and then magically the electrons becomes the human voices again at the end with the copper line and then later with digital transmission. It was just so amazing. And I just thought, wow, that was so cool to see how technology and humanity come together where you are connecting people to people even at that stage.
So for college though, I took a little way and I went to structural engineering and architecture with my undergrad. And it's a cross interdisciplinary learning where design and engineering intersects with psychology, society, and art form. So I was super interested in the way people interact with man-made environment and how man-made environment is actually influencing the people's behavior. So then how did I get to Intel? Well, after I got my graduate degree, I joined Intel first as a construction program manager to build the world's most complicated factories, which it's called fabrication plants or fabs as we call it at Intel that manufactured the most advanced technology by literally lining up electrons to move, process, store and communicate data on circuits that's thinner than human hair. So then when I finished that first project of installing the ultra pure water system in the New Mexico factory Fab 11X, I became very curious to learn more about semiconductor manufacturing.
And I went back to school and took courses in electrical engineering, industrial engineering, manufacturing engineering. And I coupled that with structural engineering and material science background for my architecture studies. And that became a foundational intersectionality of technology and humanity to drive emerging technologies of the future to advance human interest.
So it's just been an amazing journey and it is about intersectionality. - So how long have you been at Intel, did you say? - 23 years. - Now you're in charge of Network and Edge.
Can you take a minute and just define what network and what edge are? - That's a great question, Camille. There is no standard taxonomy or industry-wide definition about the edge. Depending on who you talk to, to be super honest and their motivation and their perspective, the edge can be defined by products, business, whatever orientation or whatever segment or whatever motivation the speaker and the writer is looking for.
If you talk to a sports person, they will tell you that the stadium is the edge. If you go talk to someone that's in the theme park industry or casinos, they will tell you that their casino is the edge. If you go and talk to a grocery store, they'll tell you that their actual cashier is the edge. So it depends on the person and depends on the industry.
Without question, if I were to define this, I would say that the edge is a new old frontier of different infrastructure domains. It is predicted that by 2027, more than 70% of data will be generated outside the core, all generated by edge, whether or not it's going to be a device, whether it's an endpoint, whether it's a camera, whether it's whatever the usage is used on. And it is also predicted that the overall total available market value of these items are going to be about 700 to 800 billion by 2027. To define edge, you need to look at it as in, it is a data gathering, processing, storing, networking and communication data that's generated between the core, that's the data center, and the endpoint.
And that includes everything I just talked about, all those sensors I talked about. But sometimes you'll hear the word near edge or far edge. So near edge is looked upon as it is closer to the data center, such as cell phone towers and network gateway, whereas far edge is compute at the point locations further from the data center, such as shopping mall, cars, and factory. Those examples I just gave, your smartphone is an edge device that we look at, and it is a far edge item in there. Then we further define and qualify when you talk about intelligent edge or an edge compute.
It's the placement of the resources to actually process and analyze the data at the source or the point of service delivery that enables actionable insight. A great example of that is facial recognition. You're putting in sensors and processing power within that camera for the camera to recognize you or recognize whoever it is that's the camera coming through and actually come up with executable insights. So when you go to a shopping mall and you're trying to try and close virtually, that is actually an edge application and edge compute that you're seeing.
So it is a very long answer to a very short question, Camille, but it is, to sum up, the edge is everything outside of the data center core. - Talk a little bit more about that when we have such distributed means of data collection and talk about how we, whoever we is, are deciding where various pieces of data are processed rather than shipping them all, I suppose, into the data center. They're being processed in different places. How is that orchestration occurring and is that changing over time? - So right now there is no standard ecosystem and everyone is going after this market and trying to find that ecosystem and that standard in there, but you would see it in terms of industry.
So as an example, autonomous driving, there's already a type of system development and standard and development kit and languages that actually goes in and define how autonomous driving could work by placing all the compute into the car and using LiDAR or very quick data transfers in order for the actual driver to be driving safely along with the AI that's within the item. So that is determined by the car industry. If you look at, let's say, and I was using the shopping mall example, shopping mall example is another great example where we're seeing retailers coming together and say, can we build out, for example, a standardized palette of color, skin color, or eyes color or avatar that enables our customer to actually be able to react quickly and check if they like the dress or shirts that they're trying on? And then the other great example is the fast food restaurant example where we're seeing using the edge, the McDonald's or Burger King, and you guys are seeing that right now is you can go into a fast food restaurant and order very quickly. That's an edge and it is being standardized.
So what we're seeing is the segments are being standardized overall, but whoever is able to connect all of that and drive that one platform where it's seamless within the industry will win this game. - What about how it's evolving alongside AI? Is that really changing? I mean, it seems like we have to have AI in every conversation. So I'm wondering, though, are various insights being generated closer to the Edge that previously could only occur in a data center or there's of course federated learning, which is distributed sort of AI or machine learning, and it really hasn't even been around all that long. So how are those kinds of use cases changing things in AI? - We're seeing smart classrooms being set up.
Now, smart classroom may or may not be a new idea for people when we talk about it, but let me throw this out. A smart classroom that has a camera that actually tracks the eye movement of the student when the teacher's speaking, when the teacher's displaying PowerPoints or texts or things that is actually teaching the students in the classroom or whether the virtual classroom or in-person classroom. The camera is actually capturing by eye movement and facial expression, whether or not the student is engaged. Is the student learning? Is the student confused? Are there actually understanding of that and give real-time feedback to the teacher, whether or not the teacher needs to slow down, the teacher is not as effective or the student is just absolutely not paying attention and it has a better classroom management item. If we use that as an example, you can see how AI actually enables the Edge compute and the Edge compute enables the AI use case in that way. And the end result goes back to the intersection of human and technology.
Why I'm so excited about this topic is that you would actually have a better classroom and a better learning experience and you're giving real-time feedback to the teacher to actually become better teachers. And that is so important, Camille, because as AI is coming together, you have to think about new skillsets and new ways that the human being have to interact with AI. But then let me tell you about the challenge. Challenges, security, ethics, trust. How do you know if the AI is actually being trustworthy? How is that data being used? How are those insets being used? Is it really equitable? Across the group, would it? And I'll just use the diversity angle a little bit. What if it's a person who is non-binary? What if it's a person who has a darker skin color? Would the AI pick up on that and treat it differently? We all know that standardized testing actually doesn't work well with inner city kids.
We know that. That's been study after study of that. But if we look at smart classrooms and the example I just gave, how does that change the educational value and educational insight that we're doing? And is it really gonna help or create a greater chasm between the haves and have nots? - Yeah, for sure. Thinking of if a parent does not want to partake in a camera evaluating facial recognition or facial gesture tracking or deriving insights from that. Even if you're up-leveling that or anonymizing it and sort of packaging it as the entire classroom as opposed to individuals within the classroom, what happens then for the people who decide to bow out of that kind of technology? Are they falling behind? Or do we have good recognition that some people may be interested in that and some people may be running out of the room when they hear about that? - My daughter's school, when COVID happened, it was really interesting as well, is that the kids all had virtual classroom, right? And there's a camera and the kids has background in it.
Two of the kids was actually told to remove their background and then they actually did the school then installed blur functions so that all the kids have the same background in order for the kids to be in there. But I wonder about that. Why is it a difference between how a kid has something in the background versus not, right? So there is a lot of ramification when you think about using this technology, whether it's intelligent edge, whether it's AI in this new world. And it's one of the reasons why I'm also super interested in AI ethics and responsible AI things, right? How do we do this so it makes sense and then that it doesn't overpower the humanity in all of us. - So what are people arguing about other than, you gave us a good kind of description of, people are arguing over what is the Edge to start with. And probably like you said, a bunch of privacy and ethics considerations.
Are people arguing over how to architect the edge? - I'm gonna pick a little joke, but it's almost like "The Lion King." It's literally a circle of life. If you think about the edge as I just defined it, it's everything outside of the cloud, just as simple as that.
It's everything outside of cloud, whether it's you're storing the data you're processing, you're computing data, you're analyzing data or you're moving the data. If it's all that, then you can actually think about the 1980s to the 1990s where the PC was the edge, right? PCs outside of the data center and it's actually connecting to connecting different things like that. Whereas if you look at 2000, 2010, the smartphones and the machines where you're seeing the explosion of the edge, where all connected device, going from Facebook, going social media, all the way through the computer, all that. And you look at those numbers, it's actually scary. In 10 years from 2001 to 2010, the US mobile users for the smart edge grew from 31 million to 856 million.
10 years. That is incredible number of that. So at that time, all the data goes to cloud, right? And so there is an easier way, if you think about it, there's a way to standardize, which we have, and there's a way to update and there's a way to oversee everything because the ecosystem and the dev cloud and all that has been set for that. It took over 30 years. Now we're in 2020, in this decade, we already seen the whole shift in what needs to happen.
So when you think about that oversight may not be there anymore and how do we as humans pull together and that's where the challenge will be. - We've seen a tremendous consolidation of information in the cloud. And there's only a couple handfuls of sort of giant cloud entities that have so much of this data. But I'm kind of wonder as we move, as the Edge becomes more and more powerful and networking more and more capable, that it's kind of like, I don't know which side of the coin to look at. On the one hand, possibly personal information could remain at the edge, kind of in a controlled or private environment where insights could be generated and then perhaps then shared with the cloud as opposed to transmitting raw personal information.
On the other hand, do I want whatever camera happens to be installed on the side of whatever building to be able to process my facial recognition? We have so many different vendors and so many different manufacturers putting out kind of edge equipment. Like what's your take on that trade-off? - This is happening right now, by the way, Camille. EU just passed an AI policy. China's looking at one. US government is looking at one. It is changing daily.
I was speaking with our lawyer overall and there are policies set in place to do that. So it is being watched by the government. And the question then is how do you set these regulatory policies without killing the innovation? That's always that challenge, right? How do we do that as well? But my point of view is very, very simple in a way.
I think that if we're thinking about the edge compute, we must not think about these data as permanent data. We must think of this data as in the presence of when my usage needs to happen. And it's a harder one.
So as an example, if you've ever been to Disneyland or Disney World or Universal Studio, you have an app, right? App tells you where you are, what you can do a fast pass to jump the line. You can make an appointment where the line is at. It tells you everything.
It's on-prem. That is actually an edge application. It's on-premise. But as soon as I walk out of that premise, all that data should be expected to be deleted. That would be the marriage of a perfect, I have the usage at the tip of my hand with this screen. And at the same time, I know that whatever I bought, whatever eyes I looked at, whatever camera, whatever image that is of me is gonna be deleted so that I actually get my privacy back.
- Well, what if you return to said park multiple times a year? So are people arguing over that? You know, when you cross over the line, now it's okay to sort of regenerate and remember what you liked, or is that kind of a consumer opt-in? - The memory should be on your phone or the memory should be on your computer. - What would you personally kind of worry about as the edge is being sort of orchestrated and networked? Also should note or ask you about, you know, all the variety of different software vendors who are putting together this kind of orchestration to decide where to process various kinds of information. Like what do you personally have concern over as we go forward? And I'm also gonna ask you what you're personally excited about.
- Privacy, security, and ethics are three, obviously area that everybody's concerned about. But going back to the theme of humanity, I am concerned about our next generation. I am concerned about our future, our kids. My daughter, I talk to her all the time. There are jobs right now that doesn't exist five years ago.
You're talking about data scientist, AI, big data, large language modelers, those jobs did not exist in that. So with technology advancement, what kind of skillset do human being need to have so that we can continue to grow as a human being and actually help our item? I know that my daughter doesn't need to memorize any more of the US history. I know my daughter doesn't have to go into a library and look up an encyclopedia.
She can literally type all that. She doesn't need to memorize the Pythagorean theorem. She doesn't have to memorize any of that.
But then what is she learning? Is she learning just using the edge compute to actually help her to get to where she is? I don't know. If you look at where we have come as a humanity for so many years, we're solving complex problems. With generative AI, with all these AI tools, we may get lazy. Then what's gonna happen with humanity? And so I think the key is how do we educate our next generation? How do we ensure that with all these advancement that there's a way for them to do and actually to learn and actually have new skillsets? - And now I have a question I just wanna ask. Because the better the edge gets at sort of customizing for its environment or for its human that it's helping, the more I guess fraught again with potential issues because if you're researching something, but it's coming to you from a well-orchestrated Edge, you may be seeing just more and more and more of the kinds of answers that you've already been interested in instead of tapping into like a broader data set that hasn't been customized to you.
- Yeah, that big bubble and you're living in this bubble and you're not thinking or the critical thinking skillset, which is what I always tell my daughter as well, is you need to question the data and the resultant summary and conclusion as just a first step in what you're learning. You don't know. You don't know. That can be an opinion. That could be an input.
Everything you and I are talking about right now is not industry standard coming out. Things are shifting as we know. And so to create that critical thinking skillset and the versatility to lead because leadership doesn't come from edge AI. Leadership is still a human skillset. And at the very end of the day, I think decision-making and quality database decision-making with critical thinking is the skillset that human being must develop and must advance. I just wanna say one very important thing to your audience as well is when we think about intersectionality, when we think about edge, when we think about technology, it can apply to your personal career as well.
If you heard my career that I talked about where I leveraged structure engineering, material science, industrial engineering, electrical engineering and everything, the continuous learning is the key in this new era of edge compute and intelligent edge. That is a very important thing. Being able to learn and connect things together is better than any AI can do. And that's really a very important point about career development. - So, and I cut you off because you were gonna talk about what you were excited about. - Oh yeah, no, I'm just excited about data democratization, right? If you think about it with edge compute, the data and the executable insight can happen with everybody, right? And if the knowledge and the information and the insight and foresight represents power in human society.
In the past, when it's cloud or internet, whatever it is back in the 80s and the 90s, where it's only accessible to a few and to the rich as well. People that have the dollars to pay for a $4,000 phone and able to pay the high speed internet, now it's everywhere. So that knowledge and that insight where it's everywhere, it distributes to all of humanity and that will make the society better and truly improve human lives. So that I'm excited about. - Well, Joannie Fu, VP in Intel's Network & Edge Group, really interesting conversation.
You're so thoughtful about humanity and technology and what it means for the two to come together. So thank you very much for your time. - [Narrator] Never miss an episode of InTechnology by following us here on YouTube, or wherever you get your audio podcasts. - [Announcer] The views and opinions expressed are those of the guests and author, and do not necessarily reflect the official policy or position of Intel Corporation. (gentle music)