 Hello, welcome everyone. I am Dimitri Perourlis, the head of Electrical and Computer Engineering and it's my pleasure to welcome you to this very special event. This is the Purdue Engineering Distinguished Lecture Series and we started the series back in 2018 where we're inviting world-renowned faculty and professionals to Purdue Engineering to encourage thought-provoking conversations and ideas with our faculty and students regarding the grand challenges that we are facing. Our distinguished speaker typically presents a lecture to a wide audience of Purdue faculty, graduate students and undergraduate students, followed by an interactive panel with Purdue experts as well. So to introduce our distinguished speaker, we will have our Dean of Engineering, Dr. Mang Chiang, and after the introduction I will cover a couple of the logistics of how we will conduct today's lecture. Dr. Chiang. Thank you so much, Dimitri. Hello everyone, this is Mang and as Purdue Engineering aspires to the pinnacle of excellence at scale, we welcome the visit, perhaps virtual these days, to Purdue Engineering from truly outstanding distinguished speakers such as today's speaker at the Purdue Distinguished Engineering Lecture Series. Now, I do have to brag as part of my job description. First, the Purdue Engineering research and grad program is ranked this year as number four in the United States. And our enrollment size undergrad master and PhD equals the summation roughly of the enrollment sizes of those three ranked number one, two and three in the country. I believe that no other engineering school with as many as 14,000 in residence students have ever attained to the top five in the United States. So we are immensely proud of our faculty student staffs accomplishments. And we are deeply appreciative of the outstanding speakers who are willing to discuss their research and their vision to a broad audience with us. And today's speaker is someone that I have admired for a long time and have had the pleasure to work with over the past couple of years on various projects. And I'm glad to also say that he's starting this month. Also the new Armstrong distinguished visiting professor to produce College of Engineering Dr. Marcus Weldon. And I'll have to compress the introduction of Marcus resume to just the two minutes, otherwise it will eat into his presentation about our collective future. Well, Marcus Weldon became the CTO for Alcatel Lucent in 2009. And then in 2013, he succeeded John Kim as the 13th president of Bell Labs, the iconic innovation engine for this country and much of the world. And then in 2016, following Nokia's acquisition of Alcatel Lucent, he became also the corporate chief technology officer for Nokia. And over the past a decade and more in these critical leadership positions of global technology innovation. Marcus has launched many important projects, including those in communications, but also broadly speaking such as the Bell Labs prize and the future X projects, which were all launched during his tenure. And he also is well known to be an advocate for the interactions between arts and technology and between the virtual and the physical size of engineering, which is precisely something that where the boiler maker engineers are very good at. And until last month, as the president of Bell Labs, Marcus has had a particularly remarkable impact to the evolution of solutions such as advanced communications 5G, such as artificial intelligence, such as quantum computing and other types of applied science and he of course he himself was trained at Harvard as a physical chemist, but the impact that he has had has been broad and deep. We are truly delighted to welcome Marcus to share with us his views, coming from a decade long leadership at Bell Labs, and coming from his insight, leading the technology of two of the most celebrated companies in the world of technology. So with that I first turn it back to Dimitri to talk about the logistics and then very warmly welcoming my friend and great colleague now Marcus Welder. That's wonderful and as we welcome Marcus I just wanted to say that Marcus has been very kind to make this an interactive discussion. I think we would very much enjoy having your questions that Marcus can address as he speaks. So I'd like to encourage everybody to think and put your questions in the chat box and periodically with Marcus permission, I will interrupt him and pose those questions. So thank you so much, Marcus. Thank you both. It really is an honor to be here and my initial enthusiasm for the Neil Armstrong Visiting Professorship that was offered to me has only been multiplied by spending today with faculty virtually of course. I did just get my second vaccine so if I, if I fade today, I will blame it on the vaccine, but it does mean I can be there possibly in person in the not too distant future. It is a great pleasure speaking to Purdue and I didn't know but congratulations among on the fourth place. The fourth is is is probably a the most prized place and that one, two and three almost certainly have bought their way to those those places by endowment and alumni. But I can understand why you're in fourth place it's a fantastic institution. And I'm actually going to give a talk that ends up at Purdue in some ways, meaning my talk is going to propose the need for a new way of innovating. I could argue that Purdue is set up ideally to be that innovation environment. Okay, so I'm going to get started. It's going to start on with a talk on the nature of innovation. One of the things that people often ask me about because of my roles is how do I see innovation and successful innovation I'm going to talk about in particular. I'm going to talk about sustaining and disruptive innovation, as well which some of you will be familiar with from Clayton Christiansen's work but I'll give my own take on it. Because what I want to do is show that we're at the dawn of a critical new innovation era around industrial automation and the need to automate our physical world and physical processes. But the challenge is going to be innovating what is now a multi dimensional complex system, which I think requires a new innovation model. So that's what I'll end up is proposing that we need a new innovation model and open up. I have a view on on that but I actually think Purdue and the mungs leadership and across the university can play a very leading role there. Because Dimitri said I'll stop periodically but feel free to send in questions and interrupt me. I'm going to try and get this done in about 4045 minutes and then we have 10 minutes of Q&A at the end. Okay, so let's start on the journey to create the future and we're going to start by thinking about the terms in innovation and invention. Alan Kay, many of you may know is a famous computer scientist, and he had this very popular quote, he's almost more famous for this quote than his work. The best way to predict the future is to invent it it's very pithy and seemingly wise, but it does not tell you what to invent, but it simply says you should invent the future. You need to combine that with an even older guy, and here is a relatively old guy, equally smart, maybe even smarter, Plato. And he was the one who came up with the phrase we commonly use is necessity is the mother of invention you see the way he said it was a little bit more historical, but necessity is the mother of invention. So if I combine those two comments about invention, you could argue that the what it asserts is that you should invent what is necessary in the future and that seems somewhat trivial as a statement. But I think it's it's relatively powerful it talks about necessary things for the future and that's where you should start focusing your intention. But there's more to innovation than invention, and that's where I'm going now. Another very well known individual, Thomas Edison said vision without execution is hallucination, a particularly favorite quote of mine. So what it says is, if I translate it into invention invention without a implementation is a hallucination or is a maybe a piece of marvelous academic work, but it's not something that can change humanity until it gets implemented. And in fact, that was also conceptually framed by an economist who actually ended up at Harvard, a guy called Joseph Schumpeter. And he said innovation is the market introduction of a technological or organizational novelty, not just its invention. What this is absolutely critical is that true innovation and this is why someone summarized this first slide is, you have to implement your invention in the marketplace for it to be an innovation. Otherwise, it's just an invention. And that's critically important because it tells us quite a bit about what we need to do it says we need to understand the markets. We need to look forward into the future, but also understand present reality, and then invent and implement solutions that get us from today to that necessary future. So let's build on this a little bit more. So that's the basic definition innovation is invention and implementation it's not actually strictly additive, but we'll leave it as that simple formalism, but I want to take you to step further. So what I would argue is, that doesn't tell you whether it's successful it says you it's necessary but not sufficient to invent and implement that you actually need it to hit the market at the time that there's a human need or market needs so I've recast this as successful innovation is those two things invention and implementation to the power of market factors and again the power is somewhat arbitrary but what I do want to suggest by that is. That's the more important term in many ways which, which often shocks people because logically or even morally it seems like invention and implementation should dominate, but these following market factors often end up being deterministic. And they are and I broadly describe them here, economic advantage you have to have a personal performance advantage relative to an existing solution I call that a market factor because it's stupid cost. It's not the invention per se it's not the implementation per se it is actually the economics of that which are somewhat independent market timing is their current marketing is, can I create the mark, meaning if there isn't a current demand can I create or increase that demand. And then there's an incumbency factor. Do you have the ability to displace the existing solution. Even if it's deemed not optimal, is it displacable is the incumbency too strong, and then is it sustainable. And the existing solution catch up and overtake, or kind of competitor come along that sort of the selfish view is, can I win marketplace with my invention or someone else when because they can offer an equivalent solution with other factors that favor it. So this is critical because if we want to change the world if we want to create the future. We need to invent implement and then also manage these market factors, and it's very hard to manage those talk a little bit more about that, but it's important, particularly as we look forward to the future if we want to win. Win meaning having a leading position in innovation globally, we need to manage the market factors to. I'm going to put these concepts together and again now I'm introducing the idea of two types of innovation. There's disruptive innovation and sustaining innovation. These are well known concepts to anyone who has read Clayton Christensen's work. I like this little graphic here that I found, but fundamentally it says disruptive innovation often starts and this is the key message with lower performance in the key features. The market has more defects and less speed, less power, but it's, it's the beginning of the future. It's simpler. It's more economical. It's, it's, it's more agile. So even though the thing is not complete in its conceptual implementation, it has the potential to overtake the more complete solutions that are typically arrived at by sustaining innovation. And frankly, the way to think sustaining sustaining innovation is satisfying customers current needs and disruptive technologies focus on customers future needs needs they don't even know they have. And that's the critical thing real innovation disruptive innovation is about solving for a problem that the regular company the regular marketplace doesn't even see exists. The innovators dilemma as cast by Christensen is how do I as a company do things that are sustaining because I keep my business running forward, but I actually invent something radically different. And when do I adapt from the sustaining to the disruptive and of course he points out that most companies never adapt what happens is one company is replaced by another. The analysis of this in terms of my factors, my factors if you remember invention implementation as the innovation factors and the market factors are the five factors hit, and I gave them a nominal score again nothing rigorous about this. The key is in this conception that you could argue that the disruptive innovation is actually better in invention. But less good in implementation meaning it's a it's a more basic implementation. If you if you think about it that way. And then on the economic factor the market factors. It does almost as well in some categories market timing it can control a little bit. It can do well in marketing is a lot of hype associated with new stuff so it can, it can compete favorably with much bigger marketing budgets from from larger companies. It has sustainable differentiation sometimes sometimes the sustainable differentiation is because you have a complex solution, or you have more R&D that allows you as a sustaining innovator to win. But the big factor that changes here is initially the disruptive innovation has no incumbents, and the sustaining innovation has massive in company. And that's the factor that switches the balance and I'm going to show this in a plot now is all other factors being. The disruptive innovation fails until it can overcome the incumbency, because it's innovation factor its economic advantage becomes preeminent and required by the market. And then that X mark I have the disruptive innovation becomes a massive check, because the incumbency becomes an impediment so I want to show this graphically. And this will be the end of my intro section and I'll pause for questions. If you think about what I'm saying. I'm saying that my successful innovation innovation metric which is the vertical axis here starts by in the, in the sustaining domain with expansion of an existing technologies well known S curve of adoption and off it goes. And over time that seems to be doing well and it's highly profitable for the companies that make it. But really what then happens is this. We have a disruptive innovation coming with a lower feature set, lower successful innovation metric, because it doesn't have that incumbents, but it moves with agility and speed. And with economic advantage to gradually initially but then very rapidly overtake the incumbent technology causing the incumbent technology to a very short order be eliminated. And that's the dotted line down to the bottom as the new technology gets adopted and this is a sort of classic take on the Clayton Christensen model. But you see that for a long time the sustaining innovation looks like it's, it's doing just fine. In fact, maximum profitability as I mentioned is at that point of the plateau. So it's, it's what's known in the industry as a cash cow is printing money when it reaches that plateau. So it's very tempting to say that world will last forever if you look at it looks inexorable. It goes on forever until the dotted line kicks in. So this is the dilemma zone, as I would call it and it in fact is the innovators dilemma is how do you navigate the sustaining innovation versus disruptive innovation space as any type of company, because typically you are if you're a successful company on the black curve, you see the red curve as in the rear view mirror or you know almost invisible perhaps you have to look at it to view it through a high power magnification because it's almost invisible in the early phase of the exponent. And then of course it starts taking off but by now, it's too late for you to adapt to that because it already has as insurmountable momentum. So this dilemma zone is actually the critical period of time for the companies, marketplaces to to adapt. And I would argue for the rest of my talk and as I said of course questions is we are in a new dilemma zone with a massive scope and scale that the topic of which is the next industrial revolution. And it's a complex landscape of many technologies coming together that have to be integrated to allow the automation of our physical world. We are equipped in the current setup, I would argue, to navigate this dilemma zone, and other nations, particularly, perhaps China is better equipped because it requires a more coherent approach. So I'll pause here just for the intro. See if there are any questions and I'll move on to what I think are the ingredients in the dilemma zone and then talk about the innovation model going forward. Any questions yet to meet you. Yes, we do have one question. So would you like to comment on the work of Amory Lovings and Rocky Mountain Institute and his concept of integrated design for energy efficiency and of reinventing fire. Not yet I think I'll save that to the to the end, I'll address that a little bit more. All right. Okay, I'll save that to the end for the Q&A session. Maybe just a couple more that they have just come in. There's a colleague that says I think of Alta Vista and Google from this diagram. Do you have any successful examples where companies were able to overcome the dilemma zone and become also the disruptive innovator? Google, I would say, was the disruptive innovator, admittedly in a bit of an open space, which was internet search, but there are others. There was Alta Vista, there was Yahoo, there was Jeeves early on, lots of these search engines and they obviously had a better algorithm with better scale. So there was a space was open, I would argue, which was sort of internet search and Google won in an open space by having a better algorithm. So I don't think the early Google matches this diagram. I think what it does is how the web scale players started owning large parts of cloud and networking infrastructure relative to the telecom players, for example. I think what has happened in the networking or ICT space is the web scale players use that initial entry point and success to actually disrupt, I would argue, the ICT space in large scale and have frankly overtaken computing services from the traditional offers of mainframe and cloud-based computing services. But I think that it does apply to computing infrastructures and companies. I think web search was an open space. And it's great when open spaces exist. There aren't that many of them. Any more, Dimitri? Wonderful. We have a few more. Maybe I will ask only one more right now because it follows up directly on what you mentioned and I'll save the others for later. There's a question actually regarding the obscurity between the user and the infrastructure control, particularly as the internet is becoming cloud-based. And so the question is, does the loss of physical control of infrastructure hinders our ability to innovate in the online marketplace or does it strengthen by outsourcing management of others? Oh, that's an interesting question. I think I'll defer it until I've discussed the industrial new reality because I think the current web is not so interesting to me. It's the new industrial internet that I think is interesting. And that's where there's a new coupling that has to emerge because the system level optimization is critical. I think with web services which are designed to be best effort and work relatively well with adaptive rate video streaming and no reservation of bandwidth in the access domain but just building out more and more capacity. That's relatively simple paradigm has worked for consumer grade. I think with industrial grade services, there has to be a tighter coupling between the ones owning the infrastructure and the ones owning the software systems, which is what makes it so interesting for an innovation epoch. Okay, so let's get on to where we're going. So that was innovation and that sets the backdrop for what we have to solve for and this is where we get to the industry 4.0. So let's look over the history of industrial revolutions, particularly the technological ones. You can read how I've characterized the industry 1.0 somewhere around the 1800s was mechanization of local physics tasks and steam engine was only operating statically in a in a locational in a relatively small scale. That's obviously we're printing presses and weaving looms and automation of those things happened industry 2.0 actually then was the first sort of manufacturing so larger scale mass production so automation was a big focus in industry 2.0 rather than sort of mechanization was 1.0 automation 2.0 and and yes, also increasing the scope. The idea is that was producing vehicles and infrastructure that could travel larger distances. There was an efficiency in that 3.0 is loosely the internet age, where what we did was stop focusing on sort of physical tasks, but 3.0 was about physical media and digitizing access to physical media, and then along came mobile as part of that because when we digitize physical media, we can consume it anywhere and that naturally makes sense to tether ourselves to consuming digital media on only fixed locations really didn't fit with the logic the logic is digitize something so that you don't have to be tethered to a location of record player CD player that you were using to consume physical media so so that that was 3.0 4.0 is actually then going to extend that these things generally a progression, instead of just being mobile access to physical media it's actually kind of completely wirelessly controlled physical systems, allowing global access and I think everyone COVID age understands global access meaning access from anywhere, but with high order local optimization so you're remote into something, and then locally control. Parodyne shifts from local human limited tasks at the early stage of these industrial revolutions to global local combined globally accessible locally optimized and human augmented. And down the bottom I've conjectured that the innovation leader was individuals then industrial complexes that were build on industrial power houses, then Bell Labs probably pioneered the industry treat it was among said, then a bunch of startups took those innovations and became web skills, and the question is, who's going to lead the next industrial era, and that's what I'm going to talk about for the next few slides. So if I look at this third industrial revolution, it's always good to look at the present to understand the future. So what happened in the networking space is not much this is the market capitalization of these companies and I could have picked any company in the ICT space, and basically for the last 30 years the internet age, they bumped along in response to market fluctuations. Nothing much has changed, they haven't added value, which is shocking really because they these are the companies that pioneered the ICT or internet space. The users and can have that space with these companies. And they've seen a massive up to Amazon for obvious reasons, with its.com business and its AWS business, Apple with its device and device ecosystem, and now Microsoft. And what I like to focus here on is not the Amazon's as well, although it's interesting or the apples, but actually Microsoft, because Microsoft is a traditional company that took its enterprise offering the office suite, made it available on the cloud on a next gen infrastructure available on any device so globally accessible on any local device. And as a result has turned around its trajectory to look like one of the pioneering companies of the internet age when in fact it's a company that, prior to that look like it was on the same trajectory as the classical IT companies. So, the point I'm making is, there's a possibility when we reimagine the future of digitizing the physical world, focused on enterprise and industrial systems, making them cloud native making them available on any device but globally accessible that we generate tremendous new value. And so Microsoft is some ways on the slide is the bellwether of that phenomenon. It's not necessarily going to be the apples and the apples, but again I'll come back to that question in the Q&A. So what is it we're trying to do it's always good to frame the problem correctly and here's how I like to frame it. If you look at this data from Robert Ford, this is his from his book, The Rise and Fall of American growth in the first and then second industrial revolution is the blue bars and productivity growth, this is productivity growth so all of these numbers are growth on the previous number was going along with a very impressive growth rate until the 1950s. The internet age which kicked in on the back of that has led to a decline in productivity growth or growth still an asymptotically going to zero. And this is a big deal because if you look at this and Paul Crockman the economist, his quote is, as you see, productivity is almost everything. And if you think about human endeavor, most of what humans try to do is increase the productivity of our lives and systems and processes and productivity simply put is the rate at which you or number of goods produced per unit time per unit worker. And so the reason we live in cities the reason why we have, you know, larger scale agriculture and transportation systems is all about moving goods from point A to point B more efficiently to increase productivity so it's hard to think of things in human life that are to do with commerce that aren't about productivity. You could even argue aesthetic things. What you do with the time you create by optimizing productivity and I'll make that point in a little bit. So even aesthetic things a couple to productivity as the companion function. So let's look at what has changed or is going to change because this is going to be my thesis for the future and it's very important slide because I'm going to show you in the 2020s and beyond what matters is not what has been mattering in the last decade. So if we look over the decades of the third industrial revolution in the 1980s was about supercomputing and the way to read this is the blue text means it had preeminent value or premium value and the commodity functions or assets were the great text. So 1980s supercomputing hardware mattered. They were big complex systems and OS is like Linux mattered. There really wasn't a sophisticated network. There weren't any applications and there wasn't a human machine interface, anything other than a punch tape or a clunky terminal. And that's what HMI stands for by the way 90s. Of course, Moore's law changes that we move to a personal computing paradigm. And we move computing onto everyone's desktop. And I emphasize desk. It was an enterprise driven transition not a consumer one where we went so that we could actually process data on office data or enterprise data on everyone's desktop. And that's when the app started pairing office suite starts appearing and human machine interface matters because it has to be much more naturally appealing or intuitive for anyone to use in parallel with that revolution that the 2000s saw the advent of mobile. And initially that was just mobile comps. And that's when the bubble of the telecom era happened, because this idea that you could mobile communicate anywhere was remarkable up to that point everyone had had fixed lines and your communications were defined by a location. And suddenly if you imagine, conjecturing to someone that within a decade, you could have global mobile communications and the infrastructure that would take is massive but it happened. Now, if you think about the 2010s. What really happened there was that people like Steve Jobs recognized and they that you can see the early origins of that in in the Apple Newton, which was a personal digital assistant or PDA. They realized that mobile computing was actually what the answer was to take the PC of the 1990s, the mobile telephone of 2000s and he merged them to create a mobile computer that consumed media and content, but actually became the basis of the current age way more and more than tablets and smartphones that are very closely related. There's been a little bit of regrowth of, of, of computing on, on, on harder systems or, but even those are all wireless now, mainly wifi, and their laptops so all computing has gone mobile, driven by the realization by Steve Jobs that people ultimately wanted to compute wherever they are, and not at a fixed location. Good, that explains the current world. But we're about to shift to solve a different problem. We've solved the problem of media entertainment on the go and maybe some financial systems we can access an e-commerce systems. But this is what we're going to do. We're going to move to solve the productivity problem. And for that we need a new set of, of value. So we're going to talk about this more. The new value set is nothing to do with apps running on smartphones and OS is running on smartphones and a mobile device hardware that is, you know, prized by consumers. The inversion, you see the blue text and the gray text is going to happen where we still need human machine interfaces and I'll come back to that, but they will end up being more native to us. There will be simpler devices. There won't be things we carry around, like this device. So the hardware won't matter as much. They may not even be a sophisticated OS in that device. Think of in the simplest way it's a, it's a sensory device. It's a sensory device with very simple hardware, no OS and no apps, because apps in some ways are a complexity that we don't need. What you really want in a future industrial system is the data coming from these new sensory systems running over a high performing network to an edge cloud. I'll explain the edge cloud in a minute to an AI based system or I call it augmented intelligence because humans are in the loop that actually computes the perfect outcome and then signals that outcome back to another piece of hardware that actually executes it. Again, it could be a human wearing that piece of hardware or robotic system. So, but it's a very different value system from the 2010s. I'm going to try and flesh it out for you a bit more is we are about to invert the value system like we have every decade. But, and there'll be a new set of winners in that space but one of the key things is the AI system the edge cloud the network and the human machine interface system have to work seamlessly together because the performance criteria are much higher. So I just give you a sort of a context set what coven is taught us is what I have just told you isn't just wishful thinking. It's a necessity if you think what we really learned in the coven era is that we want to remotely control interact treats assemble manipulate manage diagnose everything because that clearly allows us to a degree of freedom that is important. And so we're moating into everything but then having the ability to control all those systems so it's that global access to local control. I think we've all now seen the lights, whereas before it was an argument you have to make that that's what the world should be like and no one really agreed, because we had this sense that you have to be physically present to, to operate and manage your business or a process and now I think everyone has agreed. The one good lesson from coven is, there's a better way, and the better ways to be able to remotely access and optimize and perceive everything from wherever you are. But the problem of that is this, in order to access everything, you have an explosion of data and this is a curve called the, the Buckminster fuller knowledge doubling curve. And you can see he conceived of it in the 1970s, but his conjecture was that every half life knowledge doubled, and this has just continued. So you can see this place I call the IOT tipping point. This is where we're at roughly 2019 20 somewhere in that place, we had this IOT paradigm appearing where devices would start signaling and dominating human signaling. They may be not quite there yet but we're on the path to something that is unconscionable. And why is it unconscionable because of this. This is human ability to learn. It's a ebbing house forgetting curve. And it's based on a set of studies that say humans forget 90% of what they've learned after 30 days that's what the curve roughly says, if they don't refresh knowledge. So what we really have to do is constantly refresh knowledge but we can't do that. If of course the knowledge is constantly changing it's just impossible to be popular. This is where the paradigm of augmented or artificial intelligence, helping humans to actually pass this massive amount of data, interpretive, present optionality, and then allow humans working with machines to determine new outcomes. That's a critical part of this new story, because we otherwise we simply wouldn't be able to operate on the massive amount of data we're generating in real time we just can't consume it. The way to think about this is something called Morović paradox that some of you may be aware of he's a professor at Carnegie Mellon. He pointed out that there's a conundrum or paradox that humans and machines are actually optimized for different tasks. Machines can't do human tasks particularly well, and they are movement manipulation perception, because human machines have no conception of the physical world, yet we have been created and evolved to be experts in the physical world. Conversely, humans are not very good at mathematics logic data processing repetitive in tasks, because we're not optimized for those things those are relatively recent changes or advances we've made as humans, but we're not optimized for processing those tasks. So in fact, the reality is humans and machines are always symbiotic. And in fact, I like this other quote by a student banker that the main lessons of 35 years of AI research is the hard problems are easy and the easy problems are hard, and it's a recasting of Morović paradox. The mental abilities of a four year old that we take for granted navigating the physical world. In fact solve some of the hardest engineering engineering problems ever conceived. So this is a fundamental state of affairs until machines learn about the physical world which is a complex thing to teach a machine, given its limited capabilities and it's limited sensory perception. So, for the foreseeable future what I see is humans and machines coexisting and this is a way to frame this. If you think about it what AI will be useful and that means pure machine intelligence is high data replication and scale, the very little of the physical world that's the vertical axis. Human intelligence is good at intuitive stuff if you want to call it that, compared to repetitive low data, meaning scale of data and replication of the task, but high knowledge of the physical world. And really what we want to do is blend those two things and that's the, the great space, creating what you, I like to call augmented intelligence, where we move from a best effort realm of humans, applying a best effort in human intelligence to machines applying best of artificial intelligence to a symbiotic world of human machine augmented intelligence. And in some ways you can think of the two realms in that triangle of human checking a machine at one at the upper bound and machine checking a human in the lower bound to achieve what you could call life and death or mission critical performance. So this is what I call co-assistance or augmented intelligence and it's going to be foundational to that structure we're building for the industrials. So I'm going to go a little bit more into the industrial part and I'll keep going now because I want to finish up the talk and then I'll leave enough time for Q&A. One point I like to make is that we have to think about the human part of the equation here. And there's this famous hierarchy, Maslow's hierarchy about human needs and you can read about it. The basic premise is human needs, human aspirations to achieve the top of the pyramid. So you start with basic survival needs and then you have esteem and cognition and aesthetics and then learning and then teaching others. That's the way to think about, and the joke I make is this. In fact, we put free Wi-Fi or free wireless now as one of the preeminent human needs and maybe COVID has taught us that it's double down on that. But there's a digital set of needs we have. And this, think of this as in the digitizing the physical world. We actually need to be able to sense and connect and access data from the physical world and analyze it and automate that. And then perceive it correctly in a perceptual device that helps us become, if you want, superhuman. So we've got these traditional human analog needs, we've got these new digital needs which are digitizing the physical world. The trick is we have to blend them. We're going to blend these two things, our human and our human digital needs to achieve new productivity. And if you think about it, what I said earlier is, productivity is about creating time. You do something more efficiently, you create time by automating and augmenting tasks, doing them more efficiently. But trust is also going to be critical. So privacy and security are critical in this new realm. Okay, now moving on to some of the infrastructure questions. This is just a summary, I will provide the slides for anyone who's interested, but essentially I'm identifying here, what I will now show to you on the next slide. The fourth industrial revolution in summary is going to be a set of the things shown in the green box here. And I contrast what was required in the previous industrial revolutions but I'll move on in the interest of time. This is how I see human augmentation. I fundamentally believe that these are the seven domains, we need to be able to think better. We need to have our identity and privacy managed better. We need physiological to mechanical systems. So let's think of those as cyborg, in other words, I can take a stimulus from my physiology and or neurology and control the machine. I need pervasive communications in and more than oral or video. I need actually to have intuitive communications with machine systems and other humans. I need to be able to perceive more completely beyond reading or a virtual reality I need some new perception system that allows me to be almost omniscient. And with AR, I really can be know exactly what I need to know at each moment of time and that's that's critical. I need to have enhanced senses. So maybe I'll be hyperspectral I'll sense the world, more completely not just be limited by my five senses. And then I need to be monitored myself to monitor my physiological state, not just my mechanical or sensory state, but my deep physiology. And that of course helps with health and well being. So in different domains, I see of human augmentation coupled to this, I need infrastructure augmentation, and I'm going to start or focus with the computing systems and the networking systems in the middle, because these are foundational. There are a lot of what will assist the humans when I'm doing intelligent thinking and, and, and passing of data for humans I'm going to be using augmented computing systems. I need a networking system to either network me physiologically, a body area network will connect me to an intelligent system running on the computing system so at the center of the infrastructure systems, a new networking systems. And I think of those as 5G, and then evolving to 6G and augmented computing systems include quantum and analog computing systems and graphics processor based computing systems and icing model computing systems. These are optimized for processing physical world analysis. And around that a set of other infrastructure systems and you can read them here I went through all of them, but this is the complete set of infrastructure systems, and the way I see the world. In future, is that the infrastructure systems coupled to the human systems. And in fact that's the new playground if you want, or innovation scope that we need to foster. So you see how what a challenge that is if you think they're building up mobile infrastructure was a massive challenge and building the internet was a massive challenge. There probably weren't as many simultaneous domains of innovation that are going to be connected as what I'm proposing here. So a little bit on money. It's always good to look at the market remember I said those market factors. So this is studied by McKinsey that looked at this concept of IOT based productivity. And that was originally the report was called beyond the height of IOT and they really did a very good job of looking at the tasks so the way to read this is there are six industries they picked. And each of those industries they broke into seven essential tasks managing applying expertise you can see that unpredictable work data processing predictable. They said what fraction of those tasks were automatable in a let's say AI enabled 5G enabled highly sensorized world, and that's the percentages in these bars. The conclusion was that in most industries 40 to 60% of tasks were automatable. And if you automated them it would result in 4 to 11 trillion of global productivity 2025 and 11% of the global economy will be impacted which means enhanced GDP growth of 11% enormous it's unprecedented. So this is a big big deal and you see it's across for the industrial segments. A little bit more again I'll leave this with you specific types of tasks, but fundamentally, I want to use this slide to say, well, I'm really saying is we have to sense the physical space. We have to predict a new optimum state of that space we have implement the new state, and then we have to monitor and manage the new state that's what the new human machine. And that's what the resistance regime is all about and we have to be able to do that from anywhere, which means local remote mobile and fixed and for all contexts and environments which means live real time. What a challenge but what an opportunity. And I'll skip this but it basically just reiterates that this is a large part of the physical world has not been digitized at the top physical industries represent 30% of G of ICT spend. They represent 70% of GDP to physical industries 70% of GDP and we have not digitize them in the internet age and you see their productivity growth of the of them of the physical industries only 0.7% if we look at the digital industries I mentioned, media, communications, e-commerce systems, much smaller percentage of GDP but much higher growth rate. So the basic point is, we have this to realize digitize the physical world and industrial systems, and we will realize massive productivity growth. The timing imperative. If you look at what happened in the COVID era. Again, what you saw is e-commerce growth was 10 years growth was achieved in three months. The reason the infrastructure was in place. What just required was more software systems or platforms or scale to be deployed on top of the cloud infrastructure. So, we were ready. And, and those companies did phenomenally well that you only have to look at Amazon's market capitalization, Microsoft as well, and companies like that to see the leadership that is possible once you're ready once you're prepared for the new reality. The trick with the physical industry digitization is we have to build that infrastructure in the near term to be ready for leadership in the future. And again, I think that geopolitically this is going to be a critical question. Digitizing of physical industries and physical infrastructure will define the leading societies of the future and the leading industries of the future and the laggards will likely lose out to competition or other nations. So, I want to just do a couple of things on tech, and then I'm going to talk a little bit about the innovation model and a wrap up. The reason why we have to change everything because I've alluded to this that machines are different than humans so therefore we have to change everything but I want to really drive this point because humans have set up mainly to process data on 100 millisecond timescale because that's how we, that's about the time constant of our seeing hearing tasting and smell it. However, machine systems or industrial control systems, as shown in this sort of generic diagram, if you look at the specs of those systems which are in the table on the right, their cycle times are very short on the order of one or a few milliseconds and it's all to do with precision movement and precision control. Precision in time is precision in distance of movement. So, to achieve very precise movement, you need very short cycle times from the control system. So this sets a 100 fold change in the specification of these industrial systems compared to human systems. The rather nice point is there are two things that humans are faster at processing their touch perception and what's called the vestibular ocular reflex or VOR, which is how you perceive things in a headset. So, by solving the problem for industrial systems and machines, we will actually at the same time solve a problem that allows us to interact with those machines in a more natural, intuitive way through touch perception and visual or ocular perception. So that's where you see the remote local control coming to the fore with humans and machines operating in perfect harmony. I want to highlight one more parameter, the availability of wireless systems, because all these systems have to be wireless to allow them to adapt and reconfigure in real time, which is how we're going to achieve a lot of the productivity growth. The availability of wireless systems of six nines is unprecedented, which is why we're going to need a new wireless network. So let's start with the latency imperative that's at one millisecond latency. Fundamentally, we know the answer cloud has to become distributed because the speed of light is such that you can only travel approximately 100 kilometers there and back. You can only travel around trip in a millisecond. If I need a millisecond control, I need less than a millisecond propagation, because I've got all the application logic and the D2A conversion logic and the queuing logic. So I need cloud to move to within 100 kilometers, maybe within 30 kilometers of those industrial systems. And you see already the predictions of massive growth edge cloud compared to centralized cloud by many. It'll be almost as large as centralized cloud in the not too distant future. That's just by speed of light. Here's the networking problem. And this is why 5G is catalytic in this. Previously, we've built wireless networks that were really focused on capacity, consumer capacity, but they had almost no focus on reliability or latency. They were sort of two to three nines of reliability and there were latencies on the order of 100 milliseconds. So the shift from LT to 5G massively changes both the capacity dimension, which is critical because I've all these systems coming online, but also the reliability dimension from three nines to six nines, and the latency dimension from 30 milliseconds to one millisecond. So that's why you hear so much buzz about 5G. It's not about consumer broadband. It's about this being the enabling infrastructure for the industrial future. Combine that with edge cloud infrastructure and then all those those human augmentation and infrastructure systems I talked about, and you have the future. So I just want to end on the innovation model. I promised I would get there and I just about got there in time. Fundamentally, I came from Bell Labs, but Bell Labs have this interesting innovation model. It gets back to where I started with why it was so effective is that the research labs and the product development and the market were all in contact. The market was owned by AT&T. The product development was owned by the product division, which is called Western Electric originally in AT&T and then Bell Labs did the foundational research. So it was a tightly coupled system. The market was known and the innovation was directly delivered to the market. So it was the right innovation model because it was hugely successful, but for old market dynamics because the market dynamics now look like this. The telecom market is actually still there. But it's it's much now what it's comparable to the web infrastructure market. But most of the innovative thinking is now happening in the web infrastructure market. And so what you have is a model where Bell Labs is sort of over in the telecom segment you see in the gray segment and not enough in the web infrastructure market segment. So although the innovation model remains good, it's got the wrong market coupling and that's those market factors I mentioned at the beginning to remember you can invent and implement as much as you want. But if you're out of touch with the market and don't control the market factors, you can't be successful. And so here's the problem going forward and this will be really my last slide. The problem we have in this industrial innovation that I've talked about is it requires coupling of all the systems. The first breakthrough, the initial innovation, the product development, the solution up with the architecture, the integration and deployment, the operation of that and then a feedback loop of that to optimize. And yet we are not set up. No one is set up to innovate in a coherent way across all those domains and this is what I come back to my comment about Purdue. I want to think that somewhere like Purdue a leading engineering school number four in the nation with all its myriad different capabilities and both sort of an industrial segment stuff on agriculture and farming and outdoor systems and transportation systems, and it's leading engineering computer science mathematics is an unique place that could actually act as a virtual innovation that could innovate for the future, coupled to the critical industrial markets. That surround Indiana and the Purdue University as a whole. So this is the opportunity in front of us and I think Purdue is uniquely positioned to win in this new world order. So just to summarize, the new driver is not consumer broadband and media and it may feel like it because it's what we've spent a lot of our time doing during lockdown, but the new drivers industrial productivity. The new architecture I've mentioned, and I don't need to reiterate, and a new era, which is human machine co assistance, not coexistence co assistance. And underlying that the question I raise is do we need a new innovation model, and I would answer. Yes. So I've just about done in time with a few minutes left for questions. So I will stop there and hope it was informative to everyone. Oh Marcus thank you so much this was way more than just informative is this really sparks our imagination and thank you so very much. We have a couple of minutes for questions. Some actually the questions that have already come in. You have already responded to one question was related to the role of academia in the industrial innovation that you describe. And I was just wondering if you could elaborate a little bit more on that. Yeah, I think the problem of academia is not knowing not having market context. I had lots of people in the Bell Labs from academia and generally they would all agree what they lacked was the right context so they were solving problems that they thought were the right problems, but by not being coupled to the reality and daily reality you know in the market, it changes daily or monthly, they couldn't know they were solving the right problem. And I think that is the challenge for academia is to be able to know you're solving the right problem. The other is then producing a product of course academia isn't generally set up for producing products. So there needs to be an outlet. I think startups are one way but startups generally go actually quite small and focus. And we're talking about a big innovation space here, which grand industrial problems not small problems. And even if it were to be reduced to smaller problems that a startup as well. Those, those things need to be integrated. So the challenge of integrating a number of innovations that have a rich array of market context because I've got all these is a big one so I think what's unknown in my view is how to do it I think I have an inkling. It needs to be a new type of industrial academic partnership and an innovation hub where everything gets sort of pre integrated. I think long Chang has some very good ideas in that realm, but it's not so for yet but it certainly isn't individual innovators working in isolation. It's not recreating a Bell Labs at some level that's not possible across all industrials. So maybe there's an opportunity for a multi dimensional engineering university to act as the anchor for such a new innovation reality. And I will just ask just one more question. Because I know many of the other questions are coming you'll be able to discuss in the panel falling right after. What do you think when you're hiring innovators. I know we have a lot of students attending this talk. So what how do you do that. Yeah, it's a good question. I look first of all for super smart. Because I think smart people who are open minded and adaptive so I say my definition super smart is not just in one domain. It's really truly intelligent people think broadly. And so breadth is key because most innovation in my view happens at the intersection of different domains. So if you're expert in one domain you should not be unable to consume knowledge in another domain because that's where innovation happens on boundaries off. So I like smart broad deep. And I like driven to solve human slash industrial problem so I like the application to a problem and solving it all the way to the end. And I like broad deep and and human endeavor interested. So it's a bit of a and then collaborative I think those goes with that if you're sort of humanistic you tend to be collaborative, but collaboration is key because of the multi disciplinary nature and the multi component nature of true innovation. Marcus, on behalf of everybody here thank you so very much for sharing your innovation, your innovative thoughts your ideas about future in about industrial revolution here so thank you so very much again. I'd like to invite everybody who is here to join us in just about 15 minutes to the panel where we'll continue this discussion. There is a link right in the chat box for those of you who haven't registered yet. And I would like to give our speakers just a few minutes to prepare for for the panel so thank you everybody for attending and once again thank you so much, Marcus. That was wonderful. Thank you. Thank you all for listening in and it was helpful to is a tremendous opportunity for us all ahead.