Ep45: Open Source, AI and the Skills That Still Matter | Christopher Warren | Head of Business Strategy and Chief of Staff (Asia Pacific) | Red Hat

This is JCU Conversations, a podcast show from James Cook University, Singapore. Tune in as we ask experts in the industry more about their lives and their approach to success. Subscribe on Apple Podcasts, Spotify, or wherever you get your podcasts. Let's listen to today's episode. Hi, I'm Lance Dubos, lecturer at the Singapore Campus of James Cook University. Joining us today is Mr. Christopher Warren, head of Business Strategy and Chief of Staff for the Asia Pacific region at Red Hat, a leading enterprise open source software company. If you're curious about what it takes to thrive in a rapidly changing tech landscape, or how AI is reshaping early careers and creativity, you're in for a great conversation. Welcome, Christopher. Thank you, Lance. You've set them up very high. I think a lot of people are going to expect a lot out of this conversation. So I hope I live up to the expectations you've had. I'm sure you will. You've got the Red Hat pin. I know you've got the branding ready. We're ready to go. All right. All open. That's where open source company. So I'm here to be open with you. Okay. Terrific. So I see your title as Chief of Staff, which in my brain suddenly goes to like the White House and the Chief of Staff. What does a Chief of Staff do in a large multinational like Red Hat? That is a great question. And you'll probably get, if you got in five, six, hundred other Chiefs of Staff, they would all have a completely different answer because there are so many varieties of Chiefs of Staff out there. But it really comes down to what they call the principal, your boss, who you support, the scope that you agree with them and you're related to your skills and capabilities. And the business needs that the organization has. So my day as Chief of Staff can be anywhere from writing meeting minutes one day, very, you know, basic sort of task. And you think, well, why are you doing that all the way to driving a strategic program on AI or digital sovereignty that we do. And it's one of those roles that whenever thing goes right, it's as if you were never there. But when something goes wrong, you definitely know about it pretty quickly. And in my role, I'm lucky enough to have a really good principal. And I've been doing this role for about four years now. And it really is like an MBA in a role. You get to see all of the elements of the business right next to the leader of that business. See how they operate all the different challenges, help solve some of those challenges. So if anybody ever gets a chance to do it, I'd highly recommend to do it. Okay, that's great. That's quite a wide portfolio. It can get very wide. I had never really heard that term before in corporate America. I always thought of it as a government type of position. But lately for some reason, I'm now seeing it all over the place. A lot of the organization see it as sort of a leadership sort of launch pad, I would say. So getting future leaders into a role, see the entire business before then going down into a specific area of that business to lead certain areas. So it's a really good opportunity if you can get it. And is that your plan to move into something specific? We'll have to wait and see on that one. So we're here to a little talk a little bit about AI. How long have we got? We might run out of time. While we're working, let's see what we can talk about. Many fresh graduates worry that AI is taking their job or performing these tasks that they used to perform at an entry level higher. What's your take on that? So doing an AI is like the industrial revolution. It's like electricity. It is going to change the way companies do business. I think everybody is aware of that and knows fully about that. And with that comes changes in roles, you know, businesses completely change. And things turn over. We are still, I see in the very early stages of this. So companies have gone out there and said, OK, I can see the promise of AI. I'm going to invest in it. I'm going to give my employees access to some of these tools. So they'll buy their enterprise, you know, chat GPs, Gemini's, co-pilots. And they're going to spend money putting these tools in the hands of employees. And that costs that costs money. And it's sort of getting those employees now to test those tools, play with them, what use cases can they build? What value that they can create. But it hasn't been defined at the beginning what value will actually be created yet. And so then you've got, you know, these leaders who have invested money and you pointed out grad roles, you know, potentially declining. Well, it sort of does because an organization has just spent all this money on technology. But sort of not promise. So wait a minute. I want to, I need to cut something somewhere or slow something down. So I'm going to slow down my hiring until I figure out how this AI is going to help boost my employees productivity. I think once those use cases and that value becomes clear, I think that will change. Because again, I think AI in the way most people use it today and they're a deeper use cases that are being very specific with small language models and so on. That get it very specific in a task. But the general AI for drafting emails, researching topics. That's, I see as sort of an employee plus tool investment in HR more than I see as a technology solving a business problem. I think when yet to see the deep AI being applied in a very specific use case to solve very specific problems that unlock the real productivity that they're expecting out of that. And I think that's where you're seeing the slow down in graduate hiring at the moment because leaders are spending this money and they're sort of holding that throttle back a little bit as they get that value. Well, some unfortunate news for our graduates. It will change. I also want to come back a little bit to when we talk everybody's talking about AI, AI in this and AI in that. But there's really a lot of different things underneath that umbrella and from reading academic literature from my side that there are a lot of areas things like coding and such where that it's much more of a machine learning environment where they're finding a lot of advances in productivity and AI. But most of us when we talk about AI, we're thinking about the LLMS, the Tatchy PT's and there's a lot of concern that that's not really enhancing people's productivity because it doesn't really create and the people who are responsible for creating have to check everything that comes from the AI with their own expertise. What would you say about that in terms of from the strategic position, how are we going to get more productivity out of AI? You're absolutely right. The yes, Tatchy PT can craft you a really nice email or document or something, but you still have to check it. You still have to critically think around it. You still need to get that out of it. You need some level of knowledge and expertise to make that happen. And that's sort of some of the similar concerns I have because earlier on you'd always have people talking a lot about imposter syndrome. People who are actually really smart, but they're not sort of stepping up to the plate that they think they're not good enough. AI is creating the reverse of that, the stunning Kruger of that of people that are sitting there going, I just Tatchy PT something and I'm an expert in this topic and you don't know what you don't know sort of scenario. So I see that as a challenge of people aren't using it to really unlock productivity. They're sort of using it to get a direction of where they want to go and filter down that path. It also creates challenges I see around when you create something with AI at the moment. There's a and how you then share that with other people. I think it's out of the University of Hong Kong and speaking they looked at if I create code with Tatchy PT and then I hand it over and get it inspected and I say I used AI to help me build this code. They found that if you say that the perception of you drops they go, you used AI you mustn't be a smart the AI is doing all the smarts. The more concerning when I have as well because I have two daughters the perception is even greater for women rather than men. And so you've got all of these huge tools out there which I think are a value add and you still need skills you still need creativity to get the most out of it. Chattity and all those they struggle with metaphors they struggle with cultural specific elements they struggle with deep storytelling and storytelling human connection is in my mind never going to go away. But the content that you use to craft that story you can definitely get productivity advantages of synthesizing a lot of content earlier and then your creative element just comes at a later stage in that process. Sure as a lecturer I can tell you I'm getting a lot of chat GPT content for students and I can see where it fails and I'm trying to coach them to use their own thinking but I do use it sometimes I was just using it today to help me pop up a couple of different themes and ideas for an event. But then again I have the expertise to know what's right or wrong about it and can go and fix it but in my field of law we're seeing so many examples of lawyers getting sanctioned by the court for producing a brief that has fictitious case like hallucinations. And hallucinations are very funny word because it almost gives the the chance to be more sentience than it really has there's a there's an academic paper and several in this field that have said that AI is a I won't use the word since we're on a podcast but a B S or that there's an old academic theory about the B S or the person who produces B S is unlike a liar because a liar knows the truth and is trying to deceive you for some purpose. The B S or doesn't care about the truth they just want you to think they're telling you something accurate and that's really what the large language models do they produce a simulacrum. of English language that's got good grammar and the words go together and they make sense but they don't necessarily make sense and I think that's where I'll pivot to my role at Red Hat We've definitely across trying to look at the smaller models to deliver and so I was referring earlier to getting if you get smaller models targeted at very specific problems and you're using the larger LLMs to do that sort of broad-based element but the real value comes out of all the smaller LMs and are very specific tasks. That's where you're going to unlock more of that value So today yes, everyone uses these large LLMs But when the you come to a business scenario or a specific use case to unlock value You're going to have to get more specific to avoid You know what you're talking about the hallucinations and the B-Sing of the situation And it's going to have to get very very specific and that's where a small model becomes more efficient in the way that it You know calculates the automation and the AI intelligence It doesn't take up as much compute power as much electricity and all of those things So it's much more efficient and cost effective to deliver a solution a smaller model than these larger models Okay, so you would say then that that especially large organizations should stop trying to do a wide enterprise chat GPT for everyone And really start to work with experts in the creation of models That are much more tailored and focused to what they do. Well, you can architect it in different ways You can start with that broad model at the surface But then connect it to your models underneath. So you may Structure your organization and break down your organization into different functions or areas and say okay I'm going to create a small model around this topic this topic this topic and you connect it all up into that broad model as your sort of entry point So you can piece the two together and architect it your organization the right way and this is all comes down to you know How do you want to build AI in your organization and what is the infrastructure you need and the platforms you need to make all of those small models connect to the larger models And that's a lot of what Red Hat does today. Okay a recent business insider articles said that CEOs and C suite are Adapting AI at 80 plus percent managers in the 50s and frontline workers down in the 20s So do the do the C suite need to listen a little bit more to their line workers about whether this stuff is useful for them or not I'd like to know how they're using it Yeah, I can use it a lot drafting an email So it depends on how you can get the productivity the joke was made that well now we see whose jobs really can be taken by AI It's it's those people in the extra vice president says Okay, well, let's let's look ahead. Okay What qualities are mindsets do you believe will define success in the next five to ten years? so I think regardless of where you are you need to learn how these tools work you need to learn how AI works and the good news is today if you look at You can get a subscription to chat chibit for you know 20 20 dollars a month You know 100 or dollars a year when you think about the infrastructure behind to deliver the power that comes from that That's pretty cheap when you think about all of the data centers the compute power the GPUs the power That amount of money to then have access to this tool to then learn how to use it and and create value with your own skill sets You need to know how that works, but as we already touched upon you still need critical thinking you still need storytelling You still need all of those elements to come together and I think the most important thing a lot of People need to get across is you need to actually deliver and prove and sort of get stuff done No, yeah, sure as opposed to constantly talking about getting stuff done It's like let's just it's an action and an output and a delivery not a Discussion about oh, you know, this is what I think should happen just make it happen sure I need you to come talk to my students to stress that critical thinking part But particularly the ones who are giving me this chat gbt Output for their for their assignments But I think that's very true that That especially when you are an expert in a topic so for me when I'm dealing with legal topics and I and I Use chat gbt to help me generate some things I can see where it's going wrong and How do we get people to understand that the magic box is not really spinning gold all the time? It doesn't spin gold all the time. I was testing it the other day saying okay Well, I hate looking at numbers. It's like okay Here's my budget. What does it think about my budget and my spending too much hero there and it's bad out different outcomes And I was like that's not even the number I gave you So it can't just work miracles. That's for sure and people who are really are more critical of the tools and are really inspecting them We're talking about how easily you can get chat gbt or the other models to say oh, I'm sorry You're right that didn't do that properly. Here's another one It's also wrong. You get them to talk to each other and they just keep creating more and more nonsense between them Right, we talk about the the dystopian future of our students producing chat gbt Product that our professors are using chat gbt to grade and produce the learning content So our AI will be teaching your AI how to deal with air AI. I think there's a there's a movie I saw once I think you'd really enjoy it Idiocracy. Oh, yes, if you haven't seen it definitely see it I think that sounds a lot like what you're described a lot of people are finding that movie to be very Prescient in our current environment So in more than a summary what would you Stress to young professionals to make themselves less replaceable by these tools Well, I don't think they're going to be replaceable because We are in a lot of developed mature markets. We're looking at aging population less workforce Coming into the market. I think there's always going to be roles available for young people Because that's it needs to then help society Operate moment at tax revenue perspective and pensions and retirements. There's sort of a cycle that needs to fulfill So born broadly. I don't think Then into way about being replaceable It's Putting in the the work and building the skills that are going to be relevant for that Future that we're heading towards and going in sort of eyes wide open in terms of well. Do I get Invest my time in learning this skill or that skill and Okay, this is fun and I like it But is that going to really give me the rob roller job in the future and there's some of the hard decisions that I think younger people today Gonna have to think about Because it's not the same as it was 10 20 years ago. No certainly a very new environment that that Graduates young graduates are coming out to in the workplace So all these different shifts in tech AI automation open source Innovation what major trends do you think will shape this year and beyond? So there's probably two and that I'll touch on one is open source itself Which is what red hat the organization of work with does um open source I think is Something I wasn't really fully aware of until I joined red hat because out there in the open source community There are hundreds of projects out there where anybody can contribute to these projects And essentially what red hats businesses is taking those pieces Packaging it into a subscription with services and support and then providing it enterprise grade to organizations So you can take the code out of those projects now and use it you'll have to manage it maintain it and do all of those things yourself And what I like about open sources that it is transparent and and open you can see okay inspect the coding Okay, I know what that's doing. I know how that functions working You know, it's not sending random data off down here And then if there are issues with the code or there is a security breach found there's you know Hundreds thousands of people looking at it ready to fix it and so patches and get done faster And identified and pulled in and and fixed much faster proprietary software where you can't see the underlying code doesn't have those same Advantages associated with it So I think open source is going to be a lot bigger More broadly because you've even seen like meta Announce their LLM as open source so everyone sort of going on want to make it open here and there because they see those advantages of opening it up From from that perspective Open AI is really They're going to talk to their people about how accurate the title of their organization is to their It seems they're struggling with their identity right now. They were starting out as a non-profit And now they've decided they'll actually redo it a bigger profit So that one will leave us out but I think open source is a big theme because if you can get an open source solution Often times you've got more people looking at it can get more efficient So a lot of the open source models that are coming At the moment which are based on so the bigger broader models. They're actually cheaper more efficient, you know, there may not be as as higher quality potentially But they'll give you 80% of the outcome but at 10% of the cost, you know They're open source models that you can sort of look at and go, okay, well, I don't need the Rolls Royce second do with the the Toyota or the Hondale or whatever The second topic that I think is going to be really important is digital sovereignty and sovereignty with the straining sort of Geopolitical issues and if you think back to a time where everything was invested on a just in time basis. I don't want to have inventory there. I want to just in time sort of supply chain. We're now investing for a just in case. So what happens if AWS turns off a switch? Won't do that, but what if? And so everyone's looking at what is there redundancy, their business continuity from a digital tech landscape. Where's my data residing? Where's the infrastructure located? Where are my operations people sitting? I think that is a really big area that will become front and center this year and beyond as well. Speaking of open source, I think any of us who use Microsoft tools or other well-known proprietary software and deal with their frustrations with it would welcome what sounds like a more open structure. I would analogize it to the idea of right to repair. So many companies that make machinery, for example, in the farming industry, that farmers who have a bailer that isn't working, but they're prohibited from repairing it and getting under the hood in time for harvest. And I think a similar thing perhaps with a lot of the proprietary software and so much of it is unknown what's under the hood and how do we deal with it? Is that really now a room for that? I know in the past, it's always like, oh, you should use Linux. And yeah, but okay, well, how do I install Linux? Oh, well, you have to set the this up. I'm not doing that. I'm not a computer nerd. I want my computer to do the things I want to do. Are we getting over that now? I think it's a choice that every company and individual has to make on what they want to choose to buy. And you buy one product with one set of rules. You have another product with another. This trade-offs in every decision you're going to have to make, whether you choose a proprietary solution or you choose an open source solution. So I think it's a choice more than a right or wrong positioning in what you want to do in the trade-offs with that. Well, we've covered quite a bit already. We're getting to the end of our time. My last question, if you could travel back in time, what advice would you give your younger 21-year-old self coming out of-- That if I could go back in time. I think about it, if I go back in time, the world was very different back then than it is today. So if I could go back in time, I would think the advice I'd give to myself is you get what you deserve. And what I mean by that is when you work hard, you're prepared, opportunities come up, and you can get them. If you're going to be lazy for a day and not be as prepared, you're going to miss something, and it won't be there. So my advice to my past self back then would be you get what you deserve. But if I was 21 years old today, I'd give a slightly different advice. And the advice I'd give my 21-year-old self today would probably be more around the lines of run your own race. We have so much out there today with social media, with Instagram, LinkedIn, people sharing these amazing lives of everything's perfect, again, to have everything and trying to keep up with the Joneses has sort of gone hyper, you're comparing yourself not with your neighbor, but with somebody a million miles away who's a billionaire, you're running your own race. So focus on just getting better every day and how you're improving and adding value. Don't worry about what other people are doing. So run your own race and you get what you deserve. Well, this has been a fantastic discussion, Christopher. Thank you so much for being with us. And where can our listeners find you online? LinkedIn is obviously the place to find me. All my other social media is private, so I keep that to myself. So LinkedIn, if you're interested. All right. Thank you so much. Thank you very much. [MUSIC PLAYING] [MUSIC PLAYING] [MUSIC PLAYING] [MUSIC PLAYING] [MUSIC PLAYING]