Rewiring Lenovo with Art Hu, SVP & CTO

Today, we're talking to Art Hu, SVP and CTO at Lenovo, about all the latest happenings at Lenovo and beyond. Thank you to DigitalOcean for sponsoring this episode. For simple cloud and powerful AI that's built to scale, visit digitalocean.com or just click the link in the show notes. You're listening to Joel Beasley, Modern CTO.

My first question is for you. How have you been? It's been maybe a year or two since we've talked.

It's been a hot minute, right? But the world moves quickly, and so, no, things have been well. Business is going well, and I've got a few new additional titles since we last spoke, actually, Joel.

Oh, share those with me.

Yeah. So just as a quick—I've been the CIO at Lenovo for a number of years, but I'm also now the Chief Technology and the Chief Delivery Officer for our Solution and Services business. In a nutshell, that's the tip of the spear. We're really trying to move Lenovo in a services-led transformation direction that builds on our proud hardware heritage. Everyone knows what a ThinkPad is. Everyone knows what a ThinkServer is, and we've got great client and enterprise devices. But to really be even more relevant for the AI decade, the services-led and AI-led transformation are going to be key for that. So, same pay, three jobs.

Why delivery? Why Chief Delivery Officer?

Yeah. So for our Solution and Services Group, it's really about managed and professional IT services. And it's an interesting story. Our chairman and CEO had the vision to say, well, if you think about what IT does, it's really those managed and professional services, except you're doing that for Lenovo, but we want to do that for all of our customers. And so it's not too far of a leap to say, let's take your IT team and the best practices that you've been doing within the company, and let's use that as a platform and a springboard to go accelerate the work we want to do to serve and deliver those outcomes for all of our customers worldwide. So that forms the nucleus and some of the drive for why that's the case.

And how long have you been at Lenovo now?

So I've been at the company fifteen years. Even saying that actually is a little shocking for me. I've been the CIO the past ten years, and I've had the delivery and technology role in the services business in the last three years. So it's an exciting new part of our business and a work in progress.

You know, I've seen a lot of—we've been doing the podcast almost ten years now, and I've seen a lot of CTOs, CIOs become CEO and then grow through that. Like Kyle from Verizon, he was the CTO and then he became the CEO of Verizon Business. Why do you think the technology leaders end up making the transition to the CEO type role?

Well, I think a couple of threads there that you can think about. I think one of it is, of course, the increasing importance—and I think this is part of a longer term trend—on just how much technology is in the business and relies on or, in fact, you cannot do business without the technology. And so I think being very deep in understanding what the technology is, but also I think the best CTOs are very business and customer oriented. And it's not enough just to talk about the research papers, but it's really where research gets applied to specific enterprise scenarios and real world business problems. So I think CTOs who are comfortable living at that intersection really have an opportunity to thrive. I think the other aspect also is when you really have a customer hat on, it really forces you to think how you land that technology. And that's one of the things as I've added some of my additional roles, especially around being the Delivery Officer. If I even look at how I spend my time, more than half of it is now on the road and with customers. And so you can't help but really be deep in what's happening on the ground with the customers that you're trying to serve. So I think that also ladders up and is a really solid foundation for understanding how the business is running because you literally see what's happening at the ground level. So I think those are some factors that contribute to the transition of senior technology leaders into business leadership roles as well.

What are you seeing happening with the clients? What are they all trying to figure out and all trying to solve?

So I'll go in a different direction to start, because I think we'll get to AI, and I'm super confident that will definitely show up. But I tell my teams even, and even when I talk to our customers and for even internal meetings, it's okay to talk about things besides AI. AI doesn't have to be—and I would argue if all you ever do is talk about AI, you might be doing so at the risk of a balanced portfolio and looking at other trends. One of the things that is very front and center is this notion of global local. And you can decide what you want to call it. Some people call it globalization in reverse, deglobalization, the rise of global local, regionalization. But the idea is—and it's so fascinating because much of the world went in one direction on globalization for much of the past twenty-five years, which was globalization is single direction, and it is an unmitigated good. There's only good things. We will keep doing this forever, and things will get more and more global. And that's not really the case. And so that has very real implications in terms of how you wire your company, both as a business and how you run. And it has very real implications for what your underlying tech stack and technology architecture looks like. And so I think that's actually underplayed because there's so much excitement, there's so much attention on AI. But I think at the same time, we would be remiss to not think about how some of these other megatrends are going to impact how we work in the next decade as well.

Yeah. I would agree with you on that. Well articulated about globalization being a single direction that it was unmitigated good. And when I saw that and the conversation taking place, I couldn't help but think about distributed systems and monoliths. I was like, hold on a second. If we have all of our dependencies, we need a distributed system. We need distributed manufacturing and critical supply creation across everything.

No. Absolutely. I think, Joel, you hit it on the nail on the head. There's always tradeoffs. It's just that in the framing and kind of the popular thought and the mainstream thought, we tend to emphasize and pivot on certain aspects. But if you take—and this is not necessarily how companies actually operate, but globalization and centralization in an extreme, in the theoretical, would be single point of failure. And you can imagine if the company just had one site with all employees, all your manufacturing, all your functions. Well, that might be really great for standardization and moving quickly in that one site, but it's a little brittle. It's a little brittle. And then, in architecture terms, that's a single point of failure. And so I think we're reacquainting ourselves with how to adapt to some of these megatrends, which will have a bigger regional mix in things. And I say it's so important both on the business side and the technology side because it's not just the technology. You have to run your company differently when you think about your long-term capital and technology investments for the architecture. If you think about some of the large regulatory regimes, whether they're the European Union, whether they're the US or they're China or India, Japan, they're all over the world, countries are thinking very differently because I think they've learned a lot from the rise of cloud and all the things that happened in the past decade. And everyone is very eager to apply those lessons and what it means because we think there will be a further intensification on the use of data. Who knows what, what questions you can ask, what rules with the rise of the latest generation of the GPTs, the generative AI and also the agentic. And so it's a very real thing. And now we're starting to shade into AI. But I think there are other megatrends that are at work, which I think are extremely important to have in the picture as well.

So when we see these trends, they come and go. We went one way towards globalization. At almost the same time, you also saw companies in technology—I remember eight, nine years ago, microservices, everything's going to—and so they were all going to microservices. And so we tend to just dogpile on these trends. Is that just how humans work? Do you think we will ever get better at it, or do we just accept it, realize it, and try to do the right thing in our own context?

Embrace all of it, the good and the bad. And what I mean by that is there's no point in denying human nature. And I think that's where you want to harness the good when you have trends and when you are in the bubbly, frothy space. And I mean that more objectively without any value judgment. The goodness is people get excited. There is all this ideation. And I would argue without that excitement and that tremendous momentum, you wouldn't get as much of the flourishing of the ideas. If people are all ho-hum, every trend is the same and we're just so even keeled, you don't get that rush of excitement of let's really try new things. Maybe we're really opening up a new frontier. And I think that's where our work as technology leaders is very important, which is you want to channel the enthusiasm. In some sense, you want to ride the wave because that is the tip of the spear for how you get people excited. How do you generate change? Well, it has to start from excitement. It's like, well, why should I change? Well, Joel, it's the same thing as always. Just do what I say. It's not exciting. If it's, hey, be a part of the future. Create something different. You want to actually have a new business model, new capabilities, things you couldn't imagine before. Okay. Well, now people are excited. So I think that is absolutely the positive, and I think we should recognize it. Anyone who's like, well, let's not have bubbles. Let's not have new excitement. Let's not have new waves. A, it's against human nature, and B, it misses an opportunity, because I think there's a lot of good that comes out of all that excitement and the investment, which is what we're seeing in the last few years with AI. I think the role as a technologist is how do you channel that, in some sense, modulate that oscillation, guardrails around it, where you don't want to take leave of the business fundamentals and that you can ultimately draw some lines back from all that excitement and the rocket ship that feels like it's taking off every day, and then keep it tethered to some kind of business reality. So how do you both keep tapping into the business creativity while helping channel that excitement into things that are more productive? And, of course, you can never have a 100% match. It's just not possible where every dollar is spent, especially the ones that are more risky, in longer time frame. There's just not something where you shoot a hundred times and you score a hundred times. But you accept that, and I think the really good technologist can help manage the portfolios so that it's balanced. And then you can either dampen some of the extreme oscillations, so things that are likely to result in very high deviation outcomes, or accelerate the timeline so that more of the portfolio can land on things that are meaningful for the business at the end of the day. So I think you should absolutely embrace it. There's nothing wrong with the enthusiasm.

We're allowed to be humans.

No. Exactly. Perfectly put.

Real quick, before we continue, I wanted to give a special shout out to DigitalOcean for sponsoring this episode. DigitalOcean is cloud infrastructure that's simple to spin up, complete with integrated AI dev tools plus 99.99% uptime SLAs and industry leading pricing on bandwidth. We all know DigitalOcean. It's super reliable. It's been around forever. And now with DigitalOcean's Gradient platform, developers can train, fine tune, deploy, and scale AI workloads all in one place. DigitalOcean is reliable and affordable at any budget. Companies can save up to 30% on their cloud bill when they come to DigitalOcean. Use code DO25 and get $200 in free credits to get started. Simple cloud, powerful AI, built to scale. That's DigitalOcean.

Now back to the episode.

Okay. So what's the story of the moment you were deciding, hey, I need to change the title? The title needs to change. Was that a conversation amongst your peers? How did it come up and actually go from, I recognize the initial need, I decided on the title, and I made the change?

Yeah. So this one, actually, I think there's a lot of credit to our chairman and CEO. And I think it goes back to the company strategy of, hey, we've got great client and enterprise devices, but it's hardware. How do we—if we think into the future, where are our clients going? What are they asking of us? And so I think it was very much a top conversation that pulled together a lot of things that I think we were and I was experiencing, but really crystallized it. And so I think this is just an example of been doing really good work in terms of helping transform the company and rewire around digital transformation. Because as I mentioned, that's one of the advantages of being a tenured CIO, which is you get credit for things that work. You also get the blame for things that don't, but that goes with the territory. And on balance, hopefully, you put up more wins on the scoreboard than the other way around. But all those things kind of merged at the right time and crystallized in the discussion about, hey, what you do in IT looks really similar to what we want the future of the company to look like. And that was framed that way. And I think this is where the vision from the company and our CEO is so helpful is that just pulled it all together. It's like, oh, yeah, I do IT. We want to do IT for our customers. This is the logical building point. And so that was, I would say, the key insight. And while at large companies, there's always, of course, a lot of discussions and alignment. But I think that was the moment of, oh, right, you're doing in IT what we want to build a business out of in the future. Let's use IT as an accelerator and as an augmentation to the engine and a fast start to what we want to go do.

Okay. So they looked at it like an internal portfolio company, and they're like, alright, this is going really well. We want to apply these principles to other parts of the business.

Yeah. Exactly. This is going to be the future, and this is going to be the business that's going to help bring us through the next ten years.

Oh, that's awesome. And then almost immediately, your day to day changed?

Well, I think we layered it in. Because keep in mind, Lenovo is now a $70 billion company. And so in that sense, we're a little bit like the aircraft carrier. We're always going to have great devices—phones, tablets, PCs, new computing form factors. We're always going to have that, servers. But for the services business, we were really building in a more ground up way. And so I would say the activity started changing immediately, but it took us some time to get momentum and to figure out, well, what does it really mean to evolve into a services provider that has a strong OEM hardware heritage?

(Joel Beasley at 00:15:54) That's awesome. And you guys don't make semiconductors, do you?

(Art Hu at 00:15:57) No, we do not do fabrication or chip design at that level. We are primarily on the hardware side. We create the world's best hardware portfolio. But as you know, Joel, the ecosystem is so complex, and I think that's one of the things that's exciting. AI, and especially the generative AI and hybrid AI era, has opened up things from the silicon all the way up in the tech stack at each layer. So I think what computing looks like, whether it's for clients or enterprises at the edge, has a lot more possibilities than we would have been able to envision just five years ago.

(Joel Beasley at 00:16:33) What do you think computing will look like for individuals in the future? Do you think it'll be just like we all get this super AI that we just interact and talk with and it shows us stuff, or do you think it'll be very similar to what we have now?

(Art Hu at 00:16:49) Yeah, I think this one bears some segmentation. I think on balance, we are going to move more and more, especially as a voice. Not only—I think, of course, among enterprise use cases, it's sometimes said the next and best programming language will be English. But for consumers, I think what that means is as the capabilities become better, that voice will be a more natural way to interact, as the ability to have real conversations, as the ability to have persistent conversations. And I think even if you look at what we're trying to build, which is kind of a single AI agent that works across all your devices—whether that's your phone, whether that's your PC, whether that's your earbud, whether it's some other wearable—it fundamentally does a couple of things. More and more, it's going to be able to see and sense what you see. What's your environment? It's going to be able to process that and remember what you remember and augment you in that response and in that sense. And then finally, it'll be able to be your digital twin. It'll be your kind of personal AI twin, which is it can even, like, what would you think based on—for example, hey, what Joel has liked in vacations or how he likes to spend his time and how much, how densely packed should an itinerary be—to help be an agent or a proxy for you. So I think the notion of a personal AI twin is going to develop in interesting ways. And ultimately, then it becomes a—the PC is not really a form factor. Historically, it's been a personal computer. But I think in the future, it really will start becoming more about personal computing. As we get better about the capabilities and embedding those capabilities onto the array of devices, all of these things will work together on your behalf. So it really is personal computing, and it's going to be much more ambient. And that's super interesting because the past thirty years, we're so used to there's a box. You have a GUI that you kind of click on and icons that take certain meaning, but that might change in the future as and when the capabilities get better so that more of the computing becomes ambient because it just knows. It knows what you were thinking. It knows what you're looking at. It knows what a typical day looks like. And I think as the systems get better, it'll provide the capability for you to lean into that. And the part that I think is segmentation will be how much do you want to lean in. I think there is going to be a segment of people who will never get used to it. But I think more and more as users in demographics who've just grown up with it, who are more native to digital and more AI native, those are the people who will lead the way.

(Joel Beasley at 00:19:38) Have you seen the movie Dune?

(Art Hu at 00:19:41) Yeah. And read the book. It's wonderful.

(Joel Beasley at 00:19:43) Classic. I feel like the technology—it seems like very advanced technology, but it does feel ambient. It feels incredibly simple and organic. It's not the Star Trek-style dashboards and the ships and things like that. And when I watched the—I think they just came out with a new one a year or two ago—and when I watched it, I said that. I think that imagination of our future, not the Dune spice—

(Art Hu at 00:20:09) Or maybe not dystopian aspects.

(Joel Beasley at 00:20:11) Not the dystopian aspect. The way that the technology was interfaced with, I was like, that seems sufficiently advanced. That seems pretty good. And then as we've grown, you know, I had a conversation with Shri, and he was telling me about ambient payments. He was like, "When—" because, you know, he's a PayPal guy, and we were talking about payments. And he asked me, he said, "Joel, when's the last time you paid for an Uber?" And he got me because you don't—there's no moment where you, like, actually click pay. You select the ride. It knows that you got picked up. It knows that you got dropped off, and the charge happens without your interaction from there on. And so that's how he explained ambient payments to me. And as you're using this phrase, ambient technology, I can see it. I can see that that's the future.

(Art Hu at 00:21:00) Yeah. I mean, battery technology continues to improve. The ability of us to have sensors improves, and especially the computing per unit of energy consumed. I think the efficiency there to have good usage life—because you don't want to be charging your device every few hours. For these things to be ambient, it means it's exactly the point you just made. It's only going to be as ambient as the experience is frictionless. If you're like, "Oh, man, every two hours, my device ran out of juice. I've got to charge it," it's not going to become ambient because it's going to be a pain. If, for example, out of every 100 words you speak, you have to stop, or every minute, you have to stop twice to correct it. It's like, "No. That's not what I meant." It's not going to be ambient because you're going to have to think. You're going to say, "Well, that was annoying." And so I think what will be fascinating is where exactly is the tipping point? Because here, numbers don't necessarily do a good job intuitively. If you say, "Hey, the speech recognition accuracy is 95%," that sounds good at the top level. But if that translates into every five sentences, you have to make two corrections, that's actually a terrible experience. And so I think that's the part where as we move to ambience, how much friction can we remove? And I think the bar is actually quite high because voice is one of the primary ways that we interact. And when we do it naturally, like, over an hour in person, there's almost no latency. It's a very rich interaction. And so I think that's really where, as technology companies, we have a lot of work to do to make sure that it is seamless.

(Joel Beasley at 00:22:32) Have you been playing with the voice, like, the Grok or the GPT, different voice interactions?

(Art Hu at 00:22:38) Yeah. No, I think certainly you can see the improvement over time. I do regularly, you know, across a wide range of models. But I think it's really over the last year, it's really made significant bounds in its ability to both have low latency and much higher accuracy. And so I think, again, those are, let's say, leading indicators that we can and are making significant progress.

(Joel Beasley at 00:23:03) Yes. I have been following it, and I'll check in on it every couple months until about two months ago. I checked in on it, and I felt like it was ready. I was like, "This is ready. This is able to be used in my life on a semi-frequent basis." And the current application for me is if I'm driving and I'm by myself, I can just talk to Grok because I've got the Tesla, so it's all integrated. I just press the button, and I can just have a conversation about the future in AI, and it goes back and forth. And it's really fascinating from just a burning time social perspective.

(Art Hu at 00:23:43) And I think that's something important because I think there's a lot of parallels between—I spoke a little bit and we touched on it—it's really important. Every company is thinking about how will all this technology rewire my company? How will it rewire my teams? At the individual level, I think the same thing applies because what you described is this is actually rewiring your routine. It's rewiring how you think about getting educated, how you think about thrashing out ideas. What's the direction? What's the story line you want, your podcast, Modern CTO, to take over the next few months? Who are the right guests? In the past, maybe it's a brainstorming by yourself or with your team. Now you've got an additional team member. They happen to be silicon-based and digital. But that it's a very different thing because that's probably something five years ago you weren't kind of talking in your car to an agent brainstorming about the future. So I think, you know, it's not just companies that are going to be rewired. I think it rewires us, and therefore, I think society. Because if everyone is getting rewired and have different expectations about what this can do, that's going to reshape things fundamentally. Maybe not in ways that we can necessarily fully anticipate today.

(Joel Beasley at 00:24:55) You're exactly right. I had a conversation this morning with the CTO of The Telegraph, like, a large newspaper organization out of the UK, about the different ways we're now interacting. Like, I just ask Grok for an update on the news, the things I'm most interested in, and I've got pre-programmed guardrails in there that I don't want certain things that are unusually dark or things like that. I just want news that aligns with my interest.

(Art Hu at 00:25:27) Yeah. And, again, I think that's opportunity. Because now we start getting into very interesting questions about what are those words that you just said mean? How should the model interpret too dark? Because I'm pretty sure your idea of too dark is going to be very different than mine or Josh's or anyone in the audience. And how does it make that judgment? And so I think this is also one of the boundaries. This is kind of the push-pull of how does the boundary between what I want to hand off and trust the agents and the system with versus what I still want to retain control over? Where does human judgment come from? Because at some point, you still need to decide as Joel, like, "Hey. Does that match with my values? This is going to make me think differently about something." And so I think sometimes we get it wrong of the, "Oh, we're just outsourcing our brain and it's just going to rot and we're going to be just vegetables being fed AI slop." That's too extreme. But I do think it does mean we will have to have a different mode of discernment. Basically, critical thinking skills in the AI age will be different than what it has been in the past.

(Joel Beasley at 00:26:40) And the people that have those will accelerate. They absolutely will. Yeah. I've seen a couple of those. I haven't dove deep into them, but I have seen a couple of the studies that will say something along the lines of, "We watched brain activity for people using these AIs and people not using these AIs, and their brain is essentially turning to mush by using them." Well, I don't have the details, and I didn't look into it from a medical or scientific perspective. I just read the headlines and I was like, "I can see that, but I can also see the inverse. The fact that you have PhD-level people you can think and bounce ideas off of, it could create an even stronger mind."

(Art Hu at 00:27:23) Yeah. And I think this is where it's important because sometimes you'll see the equivalent of a moral panic. And that we have to be careful not to slide into the trap where we only see one side of the technology. It's exactly as you said. And this one, I always chuckle because I think it's important—I don't know how apocryphal this is, of course—but in ancient Greece, when writing first became a technology, there was the equivalent of a moral panic that said, "Well, writing is just going to destroy society because the oral tradition meant everyone—you were able to and you had to actually remember everything you wanted to talk about. You couldn't actually store information another way." And so when this new technology of writing and permanently committing to some store information and encoding it so you could decode it and have it across time, that was the technology, and there was moral panic that said this is going to destroy society. Now a couple of thousands of years later, you and I can decide, well, this probably isn't the destruction they had in mind. But I think what you said is so important that we again, as technologists, part of our mission is to bring that rationality. It's like, yes, if you misuse technology, for sure, some of the downsides can happen. You know, it would be the equivalent of Joel, you just kind of prompted the agent and said, "I'm going to have an agent just run the podcast, decide the questions, interview the guest, and have no human in the loop." It's probably not going to be very thoughtful. But if you use it to thoughtfully dig into who's up and coming, who's making a splash, then it can be an amplifier. And so I think that's a really important point to keep in mind. We have to make sure people understand in the context. There's pluses and minuses, which ones are we going to choose, and how do we deliberately exercise our muscles in the direction that we want to go.

(Joel Beasley at 00:29:20) Yeah. And that's one thing that's been incredibly helpful with the AI searches. Before, when I'm trying to find people for the show and topics I want to discuss, it largely is just network-based. Smart people tend to know smart people, good ideas, and so on. And while that's still the case and that still works well, there wasn't really a way for me before to Google search, "Show me the CTOs that are having new ideas and new thoughts about digital transformation from a perspective that is uncommon or whatever it may be," but I can ask more advanced questions and it can dig in a different way than Google was traditionally digging, or at least how I knew to use Google to dig, and it would be able to unearth interesting writings and thoughts and all of that and then connect me to those people. So it's been very useful for the show.

(Art Hu at 00:30:15) Well, and I think that's exactly one of the really big questions for AI, which is, will AI as a new technology really become that general purpose technology that doesn't just directly affect the things that it's operating on today, but really unlocks entirely new fields or new ways of working as you said. It's not just, "Hey, you're five or 10% better," but used properly, this makes you two X, three X better potentially. And the more cases like what you said can happen, and I think that's true at an individual level as well as for enterprises, the more likely we'd be on the bull case side of AI as a general purpose technology. Like, another kind of GPT. Not the generative pre-trained transformer, but from a macroeconomic perspective, a general purpose technology that will accrue benefits beyond just its immediate use cases in the here and now.

(Joel Beasley at 00:31:09) Have you used the GPTs, LLMs at all for parenting stuff?

(Art Hu at 00:31:16) Well, that's a good question. And so the answer here—because when you said parents, I immediately keyed on just, again, just sort of my frame of mind with some of conversations with my parents. But, actually, as I reflect on your question, it's—right. Because many of us are parents and we have parents, of course. But we are also, in many cases, parents to children. And I think it's absolutely interesting to help surface new ideas, different ways that you can communicate. And I think that's really relevant because it's often just like with AI, both especially in enterprises, but I think it applies to parenting as well. You have to be able to engage.

(Art Hu at 00:32:02) It's not simply because ultimately, there's a lot of things, especially with kids, that you would like to help influence them, right? A different way to think, a different way to do something, why to do something. And so I think where it's particularly helpful is, well, one, in helping identify information. Right? But I think, secondly, thinking in terms of alternate perspectives.

(Art Hu at 00:32:24) Right? How can you actually engage the audience? Right? And I think that's true if you are doing the parenting, right? If you are being parented, right? Or you're speaking to your parents. And I think it's also true in companies.

(Art Hu at 00:32:36) Right? Whatever the message is, the ability to really look at it critically and see, is it connecting with the audience? Right? A lot of my time as a technologist is actually spent, and that's arguably a lot of this big question of, hey, will AI generate sustainable and real sustainable economic gain? I think it goes to that question of, will people change their work habits to incorporate and make the best of the technology? And so I think the GPT technologies are very helpful in helping you connect your story for change with the technology.

(Art Hu at 00:33:12) Why now? Right? Why me?

(Joel Beasley at 00:33:16) You know what? I have three little ones, like eight, six, and three, ages, and they're like little GPUs. Right? And it's like, how do I interact with this processing unit that's only 20% complete or 50% complete? Because I'm over here and I have a perspective of it, but I don't remember what it's like, or I don't necessarily have all the best practices preloaded for three-year-olds or six-year-olds or eight-year-olds.

(Joel Beasley at 00:33:44) And so I have found it very helpful to interact with it and have conversations about different behaviors I'm seeing from my kids and how to approach it and why they might be doing it and how to communicate better with them. And it's been incredibly helpful. It's made me a better parent for sure.

(Art Hu at 00:34:00) Yeah. And I think part of the reason for that is because when you think about it, these LLMs are really the encoding of everything that's ever been put down and talked about, right, and is openly accessible. And so, one, it may not always give you the most novel it can. But in this case, I think the breadth, like you said, because you don't know what you don't know. Right?

(Art Hu at 00:34:21) Everyone's the first parent once. But when you can get so many different perspectives, it really opens up that aperture so you can take stock. It helps you take stock, maybe be less overwhelmed. Right? Maybe take a breath of, oh, yeah.

(Art Hu at 00:34:35) Right? This has happened before. Right? I'm not the worst parent in the world. Right?

(Art Hu at 00:34:39) My kids are within the range of normal, maybe. I think that's extremely helpful in the moment. Right? It helps you take a breath and understand to get that view.

(Joel Beasley at 00:34:51) Well, I found it enormously useful to help me create option sets. Right? Creativity, creating option sets. It's like, here's a situation. What are potential options? And that's a large creative load, to which I've come to learn throughout my life that that is a muscle in which you can make tired every day.

(Joel Beasley at 00:35:13) The creativity muscle. And so I've actually found myself saying, okay, rather than expending creative energy right now and just purely generating from a white piece of paper the potential options, I'm going to leverage to get a small advantage of my energy use. I'm gonna leverage the AI to create a limited option set, and then I'm going to pick from them and iterate on that selection. And I have found that to be a massive advantage.

(Joel Beasley at 00:35:43) I don't know if other people are using it like that, but—

(Art Hu at 00:35:45) Well, I think, again, both in enterprise as well as consumer spaces, this is exactly kind of the digital and the human worker coming together. Right? From a chemical perspective, it's—and I think of it and carbon and silicon synthesis. Right? You and I are carbon-based.

(Art Hu at 00:36:01) Right? And now we have silicon-based entities that are very capable in their own ways who can help augment that. And I do like the point about creating options because I think that's fundamentally what we do as thinking people, right, as knowledge workers in particular, which is, right, having what you said is kind of the silicon basis where I can do a lot of the work upfront. And, therefore, in your case, you know, your judgment about, well, where else to take it can be applied in new and better ways.

(Art Hu at 00:36:32) Right? Because you actually can, like you said, right, you have finite energy in a day to go do things. And rather than thinking about how to get everything together, you can think about how you can orchestrate that. And I think that's important because that kind of leveling up is a way, it's a very valuable way of creating options for the future.

(Art Hu at 00:36:52) Right? And it allows you to direct your energies in ways that if you look at where good ideas come from, right, it's not that, oh, there was a—typically, it's not just one flash of insight. It's that there's a mix. Right? So the more ideas you can bring in, the broader perspective you have, the more synthesis you have, the wider options that you're able to go through and the higher level at which you operate really increase the probability that you can ideate, like, one scenario, right, that pulls together the best of all these elements.

(Art Hu at 00:37:19) So I think if you dig underneath the statement of why you and I find the ability to go pull a lot of work together to create perspective is so valuable, it goes back to where good ideas come from. Right? It's not one good idea came. It's this idea has a lineage with hundreds of other ideas before it, and you kind of beg, borrow, steal, and remix them. And now you can do that way faster, right, and you can apply your human judgment on top of that.

(Art Hu at 00:37:43) And I think that's incredibly powerful in both company and personal settings.

(Joel Beasley at 00:37:47) I like the way you talk about silicon-based life forms and carbon-based life forms, because I think about this quite a bit, and I'll tell you how I had a perspective change. So originally, I almost looked at it through a negative perspective, like, oh, silicon is this alien species that's coming, and it's, like, you know, taking over humanity. But then more recently, I've been looking at it like, we're a life form, silicon's a life form, and we're coming together, and we're kind of creating this hybrid life form. I know it sounds a little weird, but it almost seems like if, uh, it's what's happening. For example, our—if I were to play you a stop frame animation of, like, the rise of silicon, you would see it come be in these computers that are largest rooms, then progressively get smaller and more integrated into our bodies.

(Joel Beasley at 00:38:41) Right? That would be the progression of it. And I think that that progression will continue until we somehow blend in to become one.

(Art Hu at 00:38:50) Well, I think that is such an interesting framing, right, to really say it's a we continue. Right? Because that means there's a process. You're right. The carbon and silicon synthesis, if you think about a mainframe that weighed, you know, ten tons and took up half of a room, it's still a silicon form, and we still used it, but it was separated.

(Art Hu at 00:39:09) Right? The physical boundary is so clear. That's in the room. It's a batch processing mode. You know, we turn off the lights, and we go somewhere else.

(Art Hu at 00:39:17) All that's been happening is the footprint, right, and the physical separation gets smaller and smaller, and then the time separation as the lag, right, and the ability to communicate with these in real time also gets smaller and smaller. So I don't think there's—it's one of these interesting things where that's been happening since the birth of modern computing, right, after right around the World War Two till the modern day. But at some point, that gap gets so narrow, both in physical distance and the ability to be interactive, that people start wondering, oh, this may be something that feels very different, but it's just our awareness. It just switched on that said, oh, it's really blurring the lines in a way that before it never did because the gaps were so wide. Right?

(Art Hu at 00:39:59) So I think it's a testament to the progress we've seen, but it is a continuum. Right? I do think it's continuing to—and even the way we talk about it, like, Lenovo talks about having a personal AI twin and an enterprise AI twin. Right? And that's absolutely the direction things are going because we are. Right?

(Art Hu at 00:40:16) Because we're able to augment so much. Right? That's the value. Right? We're actually able, through this synthesis, to do so much more than we could otherwise and before.

(Joel Beasley at 00:40:27) Well, you made my day today, Art, because I have shared that idea with some bright people, and I've never got it articulated back to me so clearly. Like, you've really grasped what I was trying to say there with these life forms. And, Josh, we can mark this for, like, a clip because I wanna listen to Art's response a couple times because you said it so well because it's—it's a hard thing to talk about because it's so easy to get weird, like, with the language and, like, I guess, heady would be the word. Right? To get kinda out there.

(Joel Beasley at 00:40:59) It's really easy to get weird. It's really hard to, like, keep it professional and to try to say, hey, I'm seeing this pattern happen. Do you also see this pattern happen? And a lot of people that I've pointed that out to, they don't like it.

(Joel Beasley at 00:41:13) I don't know if it's just a human, I don't like what I don't know type thing, or if they're not seeing it, or if I wasn't communicating it correctly, but today, I feel pretty good. Yeah.

(Art Hu at 00:41:23) If I could—well, thank you, Joel. That's very kind of you to say. But I think this also goes to the very real homework and task that we all have as technologists to help society understand. And I do think language and framing it properly is super important. And this is the change management that needs to happen everywhere, right, commercially and in the personal sphere, because I think that's why what the way you framed it, which is, hey.

(Art Hu at 00:41:49) There's a historical arc here. We're just part of this. Right? This has been an ongoing process. People probably agree with that.

(Art Hu at 00:41:56) And, yeah, okay. Now it's getting really close, so we have some additional questions we didn't necessarily think of. But right? Versus if you had asked, hey, Art. I think you're like the Terminator.

(Art Hu at 00:42:05) Right? You're gonna have this weird laser eye with this gun, and right? Then I'd be like, no. That's very scary. But if you say, hey.

(Art Hu at 00:42:11) Look. This has been happening. All that's doing is, like, the distance, the time is getting smaller. Oh, yeah. Yeah.

(Art Hu at 00:42:16) Okay. Well, let's think about what that means. Right? Because now we can have a more meaningful discussion. Right?

(Art Hu at 00:42:20) I think that's part of the change management. So I love your framing as well to help enable that.

(Joel Beasley at 00:42:25) It's important because we're humans. We're consumers, and we're all out there building products for each other. And so it would be interesting to acknowledge this pattern and say, okay. Well, how's that probably gonna change what commerce looks like, or how is that going to impact the business long term?

(Joel Beasley at 00:42:40) So I think it's—at least it's not worth freaking out over, but I think it's worth acknowledging.

(Art Hu at 00:42:45) And I think that's also, right, in my role because that's what Lenovo is trying to do, which is how can we have personal AI twins, not because we wanna be cybers or cybernetic organisms, but because they're generally going to be useful, right, that the exchange and that the coming together allows us to do much more, whether that's convenience, whether that's creativity, whether that's enjoyment, or even something more abstract about happiness. But, right, whether it's for enterprises or personal, having that enterprise AI twin, right, kind of being able to sense what you sense, see what you see, right, think what you think, right, they're legitimately going to have real applications that I think are going to continue to emerge over the next decade.

(Joel Beasley at 00:43:23) Is this something you're building in your labs?

(Art Hu at 00:43:26) Yeah. So on the both the personal AI side as well as the enterprise AI side, right, we're continuously trying to push. And that was back to our earlier conversation about what does computing look like in the future, both for consumers and for enterprises, because it's absolutely going to have a lot more possibilities. And I think this is going to be one of the most exciting decades yet in tech history without being—and I really say that without being hyperbolic.

(Joel Beasley at 00:43:52) Oh, I believe it. I mean, it's happening all around us. So I just wanna get a little clarification. So you guys have, like, a digital enterprise twin product? Like, I can train it on all of my enterprise data and it can help me make decisions?

(Art Hu at 00:44:03) Yes. That is the direction we're going. Now enterprises are a bit more vertical, meaning we have a bit more domain-specific. So, for example, in our digital worker, right, we have a platform that will help around persona identification, configuring that persona, and then helping orchestrate your work around that. Right?

(Art Hu at 00:44:20) We have also a, for kind of IT platform operators, right, so that they can look at all of their hybrid cloud assets and manage that, right, around the persona and kind of the target workloads that they need. And we have vertical solutions like intelligent stores. And so it's, I think kind of the ultimate goal might be a super agent that could run everything and really be the true enterprise twin. But what we're starting is making sure that these specific domains that we can really instrument and use the enterprise context, we have those. Right?

(Art Hu at 00:44:57) And then we're building towards where more agents are able to actually work together across those domains in a more joint way.

(Joel Beasley at 00:45:05) Okay. Let's switch. You—I didn't know this about you. You guys have a robotics lab.

(Art Hu at 00:45:11) Oh, yes. Absolutely.

(Joel Beasley at 00:45:12) What are you doing in the robotics lab?

(Art Hu at 00:45:14) Well, I think this is our fundamental belief on basic research. That isn't just going to be a product in the next necessarily next school year or the next two years. But I think the robotics lab is exactly what it is, which is for a long time, we've had this investment, which is also part of exploring how can compute become more ambient. And specifically now, I think we see an interesting combination, both in industrial capacities, but also in combining with AI to make them much more to have kind of physical AI, right, because a lot of people now associate, especially on the consumer side, with GPTs and LLMs and chatbots that I can talk to online. I can either get answers or interact with customer services, but in reaching from the purely on-screen into the digital realm.

(Art Hu at 00:46:07) And I think there's a lot of interest there. One is because in healthcare and also health, right, the ability if we can actually become much more attuned to the environment. There's a lot more that we can do. Today already, for example, we can use drones, right, we can use robotics around industrial inspection. And this goes back to, Joel, what you said about why the kind of human and digital synthesis works well because you take advantage of the characteristics.

(Art Hu at 00:46:38) Right? Because robots, right, metal, right, silicon, they have different tolerances for heat, for temperature, for tight spaces, for things that otherwise human might find dangerous, endurance, right, and we're able to extend and use the intelligence and project that into the real world, right, so that we're actually having impact on both the digital and the real world. So the robotics lab is really an instantiation of that in our belief that we can have that as part of pushing the frontier forward, right, as part of new form factors and as part of integrating AI, not just for our digital lives, but also our real lives and being able to affect the world around us. Right?

(Art Hu at 00:47:17) That's actually—and absolutely gonna be a part of the personal AI and enterprise AI twins in the future. So that's what our robotics labs do, which is kind of at the intersection of actuators, battery life, right, and instantiating AI in the world by putting the brains and the compute from the large language models in a way that can interact with the physical world. So it's super exciting.

(Joel Beasley at 00:47:42) What's the coolest robot that you're allowed to talk about?

(Art Hu at 00:47:45) Well, I think just from a enterprise technology practitioner, I think the coolest ones that I really love are industrial robots that can do inspection, right? Because it just doesn't get tired. It can do all the dangerous, dirty things that we no longer need a human to go do anymore.

(Art Hu at 00:48:03) Right? And I think that's such a perfect distillation of the sweet spot, right? Because it's actually up-leveling, and it's more capable than a human could be. It's actually safer for everyone involved, and the service gets better because you can inspect much more. The robot never gets tired.

(Art Hu at 00:48:22) It never gets something wrong, right? So I think better service, better coverage, better safety aspects. It's just on all aspects, it's a win-win-win. And I think those are the things that I think from a roboticist plus AI researcher perspective that hit all the buttons.

(Joel Beasley at 00:48:40) Oh yeah. Well, from an engineering perspective, it's cool because of the efficiency. Yeah. All right, let's wrap up with leadership, a couple leadership questions for CTOs out there listening in. What should CTOs be doing right now to prepare their organizations for AI?

(Art Hu at 00:49:02) So I think the first thing is you have to understand with a very clear-eyed view where you actually are, right? Not where you wish you were, but where you actually are, meaning you take stock of how strong is your data foundation. How well equipped are you with platforms?

(Art Hu at 00:49:22) Are your teams willing to focus on this? Because if you don't have some, from a starting line perspective, if you're not there, then trying to run a race without stretching, without getting the prerequisites in place, I think is a good way to spend money without a lot of impact. So I think that's the first one. Now, that being said, I think it's time to jump in, right?

(Art Hu at 00:49:43) This is not the time for analysis paralysis. I just heard a really good analogy that I'm going to just kind of repurpose and steal here. But if you think about what you want to optimize for, at this point, and this is no longer 2023 or '24, you don't need to say, "Well, is this technology going somewhere? Is it real?" It's real.

(Art Hu at 00:50:05) So the second thing I would say is jump in. You want to get, if anything, speeding tickets. You don't want to get parking tickets.

(Joel Beasley at 00:50:13) Oh, I like that.

(Art Hu at 00:50:14) You don't want to sit around. It's like, "Well, should I do this?" No, no, you want to go. And so when you go for it, I think the final part is how do you do so responsibly, right? Because I think it's the final thing I'd say in my top three: you absolutely still need to do so responsibly, but you can do so, right?

(Art Hu at 00:50:33) You don't have to falsely trade off speed and going for it with the safety and compliance side, right? There's enough technology there now that you should be able to have the guardrails in place, that you can follow the compliance, that you can do so in a responsible way that protects your intellectual property, that protects your company's data through private cloud and through other ways, where you don't have to sacrifice and say, "Well, you know what? I'd love to go fast, but now everything needs to go through legal and compliance and procurement and ethics." You can do all of those things and still be able to move quickly.

(Joel Beasley at 00:51:09) I love that. There's a cost to go fast. There's a cost to go slow. There's a cost for either one of them. I'd rather pay the cost to go fast.

(Art Hu at 00:51:16) Exactly. Speeding tickets, not parking tickets.

(Joel Beasley at 00:51:20) I love that. Thank you so much for listening. And if you found this episode useful, please share it with a friend or colleague who you think would get value from it. And if you have topics that you'd like to hear discussed on the podcast, either add me on LinkedIn or send me an email: [email protected]. Every time I get an email or LinkedIn message, it absolutely makes my day and inspires me to keep going.