EP 104 - Future Proof Your Career in 2025

Speaker A: Welcome to Digital Value Creation. This is the channel for business professionals like you and me to try to stay ahead of the latest AI trends. And today we'll talk about how to future proof your career. There's a lot of, um, business trends and AI trends summaries out there. So we thought my brother and I could do the same thing. So I'm Tomas, uh, I'm from Philadelphia, I'm an AI software company.

Speaker B: And I'm, um, Arpad from San Diego working for an AI hardware company.

Speaker A: And today we'll explore three of the most game changing AI trends. More importantly, actionable steps that you and I can take to adapt, thrive and lead in this AI native era. Uh, which is coming in 2025. So let's get started. We'll cover three trends, three broad trends. One is how AI is becoming more capable and what that means for you and me. How AI becomes a core enterprise system and what it means for you and me if you are in it or if you're not. And the third big trend is how AI is reaching transformational impact for business. So let's get going. Arpa, take us away with the first trend.

Speaker B: Absolutely. So first of all, when we say AI becomes more capable, what it means two things. Uh, we talk a lot about agents in previous discussions and AI is becoming an integrated part of how we do business. And AI is also getting more physical in FY25 Tomas, we'll talk about that. So what do we mean by augmented workspaces? What do we mean by AI will be a collaborator as you think about why people are excited about agents? Because FY25AI will m move from just helping us, writing emails, creating content to actually do things, execute workflows autonomously or with human supervision. And imagine it's like having a partner, having a sparring partner, a thinking partner, or somebody you can delegate things to and it gets it done. So if you think about how can I prepare for that, what does that mean to my job, why you are still needed, what is the right balance between you and an AI agent? And I'm a passionate advocate for strength based leadership. And for a decade we teach people that focus on what you are good at and try to delegate things that you are not good at. But in a funny way, with AI, this might be a little bit different recipe. Yes, you can still delegate down simple executable tasks for AI or areas where AI is already reliable and and tested and the guardrails are working. But the real true opportunity also lies on the other end. An area where you are strong. What is your natural strength? Because those are the areas you can actually use AI as a research partner, as a thinking partner, as a strategic partner, see what the recommendations are. And so you are still an expert in the field. You can assess whether those recommendations are implementable today. Include hallucinations, are they effective? And you can focus on iteratively perfect that. So this is going to be a very unique skill. And if you think about great, what do I need to do to get ready for 25? Uh, first make sure you play with these technologies. You use this technology, you test these technologies so you understand where the limitations are, where the strengths are. Because those leaders who can start to act like project managers or product managers, who can leverage this team of AI agents to get things done, are going to be the one that get ahead. This is going to be probably your best move to become the architect, the manager, the leader and think about these AI agents as your extended team. So that's one key area where AI become more capable. The other is when AI gets physical. So Thomas, what do you think about robotics?

Speaker A: I mean just commenting on what you said about strength, finding strength. So both you and I believe there is a strength finder. So we, throughout our career we always look for how do we leverage our strength versus just improving our weaknesses. And I think that whole dynamic is going to change. Not only you look at in a team and how your co workers are stronger or weaker in certain areas. Now you have to look at the, the skills of AI and to your point, physical robots. So I think for most of us we've been at, I think almost the whole world has been at hands on AI mode for the last two years. I think many people have not actually dealt with physical robots, physical AI. And uh, unless you work in a factory, definitely for Amazon, or you're in an operating room in a hospital where surgery robots are operating. And this is, we believe this is going to become more important. I think. And one way to get ahead of the game is in any industry, if you work in an industry where physical labor is present, you're going to see physical robots, um, and how you can learn to deal with them, how you can learn the software piece of it, there's going to be equivalence of prompting, not just coding, uh, around those robots training, which is similar to happened with AI systems, going to happen with robotic AI systems, that would be very important. So understanding how in your industry, in your company's robots will be deployed, how they're going to be experimented, you've seen the curve, it's going to take a while for the companies to figure this out, get ahead of the curve. You can be an engineer and learn the deep engineering parts of it, or you can just be a business user, understand the business workflow that these robots need to fit into um, or patient care if you're in a hospital, and how it's going to improve patient outcomes. So that will be a very, a learning opportunity and a big business opportunity and a career opportunity for many of you. So that's a big trend with robotics and um. With that let's move on to the second trend which is how AI is becoming a core enterprise system.

Speaker B: Arpa as you think about, we treat AI today many times as a point solution. But AI is a game changer, especially when it comes to analytics. I mean the main reason a lot of times we deploy systems and applications to enable better decision making. And uh, with generative AI there's a whole glow up on the field of analytics because AI showed us what is possible. It can look at data in a chart and explain the reason behind it, make it interactive, let people with no coding experience create uh, complex analytical tools. So, so there is a lot of excitement about it. And uh, one of the major gaining factor of AI effectiveness is actually data readiness and data curation, whether the data is reliable. Because as we all know these AI systems already have some challenges with probabilistic architecture which can lead to hallucination. If the data feeding and training these systems are not effective, we just exhibit the problem. So uh, as a key career focus area, AI might change some of the analytics careers, but it also open a big opportunity. And as a business leader or as a technology leader, this is really your chance to number one, get familiar with the latest iteration and evolution of these tools. Whether it's power, bi, tableau or whatever toolsets use, they do get AI, uh, capabilities. But probably equally or even more importantly invest in understanding the role of data platforms like Snowflake DataBricks and either AWS's. Google is evolving data platform. Number one, this field is evolving in a rapid space. Number two, this is the way how data gets liberated from the applications and, and become AI ready and you can use it for AI or machine learning tools. And understanding the limitations about what these AI analytics tools are good at today and not so good at is going to be equally critical because again similar than what we discussed in the first round, those leaders who can show up in a boardroom, in an executive meeting, in a customer meeting and can leverage these tools, get faster answers, better analytics Better decisions. Those will be the game changers. Now in addition to the importance of analytics, there's another significant shift which is how we shift from focus on training to focus on inference. Thomas, I know that's a passionate area for you.

Speaker A: It's important. And just pausing in the middle of this trends conversation. Some of this may sound very technical, but I think it's part of the message. I think all of us need to become more technical. I remember time, I still remember the time some of the people wouldn't type on computers. They had typists I think. And now everybody is typing on computers. AI is going to be a similar skill while it's going to be at a technical proficiency level that we all need to get to. And to that point many of us missed the um, um wave of AI large language model training or didn't have the wherewithal or the finances or the compute power. But some businesses experimented with trying to spend um, energy and investment on training but realized that um, the large language model and the frontier models will outperform everybody and anybody and that's where the investment is going. We're also hitting a curve, um, where the incremental improvement in training is stopping. If you noticed, um, it's taken a while for OpenAI to release. Um, uh, we're still waiting for five. Um, right. So what businesses are shifting to and even the applications are shifting to is inference which is making sense of the existing model for real world applications. Um, so that's what it means. So there are many, many techniques out there but getting familiar with figuring out how the existing models can be applied better with and there are many techniques from racks to vector databases to other other uh, tools how to apply it better for real business problems. And this is very important because uh, and one of the ways is teaching or prompting better um, the large language model to solve the problem in a certain sequence. So um, um, you probably noticed that OpenAI just increased their um, O1 Pro model pricing from 20 bucks a month to 200 bucks a month. And why? Because that model can solve problems step by step. Uh, so it's able to have better inference. It can have a logic tree that it's following to arrive at a solution. In fact it's following multiple possible avenues to arrive at a solution and it can apply to better business problems. So how can you learn these techniques and how can you stay with it? So obviously experiment with O1 but also uh, see how other vendors are coming up with very specialized model that can reason through a set of problems instead of just looking for the next token and inference and step by step reasoning is going to be the next wave of um, AI beyond training. So what about the next um, observation?

Speaker B: Arpad, you did a great job setting that up because I will talk a little bit about AIOps and as these AI system become enterprise solutions, it's no longer just creating an experiment and making it go live, but you introduce or reintroduce some um, techniques about training versus inference. I think this is super important that as you said, we as leaders and as practitioners need to reach a level of uh, understanding about what is the benefit, what are the pros and cons to use a model and what type of models we want to put together. And that's going to be a critical part of AI operations. AI is not a typical enterprise solution. Then you implement it and just run it. Being a probabilistic system, being a system that is very sensitive to the data and that is adaptable for the use case. It requires continuous tuning, training, monitoring the performance, monitoring the cost and uh, also the pace, how new models getting released, monitoring the marketplace about is there a better opportunity, is there a right time to maybe shift to another model? So what does that mean to you and your career? Number one, I think we all need to tune in on what's happening on the marketplace. Whether you are listening to a favorite podcast and we will send some links about more research oriented and uh, um, also application oriented did podcast that we are also following. Just keep up to speed about what's happening in the marketplace. Number two, familiarize yourself with tools. Whether you are a leader, an investor, manager or actually technology practitioners, the space of managing IT application is very rapidly evolving and there are many tools out there that are helping manage your inference, uh costs, your AI environment cost, uh, manage the guardrails that you put in place to monitor the performance of the AI. Um, manage and control how these AI tools work. But that's a very rapidly evolving environment as well. Those leaders, those practitioners who understand how that space work differentiate themselves. That will be one of the most sought out skill as we go from experimentation to enterprise grade applications.

Speaker A: I would just comment in aiops because it seems like such a tactical and technical area. Um, I think as AI becomes enterprise grade, that's a necessary step for enterprises to scale. You sort of have to have the underpinning at enterprise grade of systems like an AI system for business to actually get out of experimentation and move to trend 3 which is AI reaching transformational impact. So I'll touch on the first uh, um, element in that category, which is we're seeing IT everywhere. So assume you have this IT foundation, you have this enterprise grade system, you have the AI ops team that's keeping um, um, the AI technology up to speed and innovating. Then you can look at industry specific applications. So you bring in new use cases from your industry. Because the best use cases will not come from technologies, it will come from um, practitioners in various industries, whether it's agriculture, health care, entertainment, that are trying to find ways of improving their processes. So it could be better patient diagnostics. So if you follow healthcare, actually healthcare has some of the most amazing applications of AI, um, for real high risk problems, which is um, pathology and diagnostics. But you could be crop monitoring, crop monitoring in agriculture we're all used to Netflix recommendations. They're going to get even better. TikTok for sure and the possibilities are endless. But if you're in your industry, how can your industry get better with AI? Uh, and you have to pay attention what others are doing. Industries have this funny way of thinking, you know, what somebody else is doing is not relevant for our world. I think when it comes to AI, it is such a universal capability you have to look for ways if they can do it their industry that way, why can't I do something better this way? So whatever we used to do is not going to apply because somebody's going to disrupt. So the strategy for career wise for all of us is to keep thinking in our job jobs, in our business, in our industry, how we can apply something that somebody else is doing and challenge our teams or IT teams or AIOps teams, saying hey, we'd really like to see this play out, um, in our cases, in our industry and that may take us to industry conferences, we may follow thought leaders, podcasts, videos, um, and start listening for those weak signals because that can definitely accelerate your career if you're uh, in business. Or it could be definitely if you're in it, in your industry.

Speaker B: Since we talk about use cases, as you mentioned, one of them is every industry will come up with a set of use cases that are unique to them. But the technologies continue to evolve and as the cost of these large frontier models becoming so high that only a couple of labs can really create these foundational models, there is definitely a need to have more specific models, maybe smaller, more agile models that can be available in the edge, whether it's your laptop, whether it's your phone, and next year is the year where the second generation of AI optimized chips are getting into these devices, whether it's an industry sensor, whether it's a laptop, whether it's a phone. They already have AI chips, initial generation. And now we see more and more companies will release their next generation chips. That enables use cases that are very hard in a cloud, uh, AI setup, uh, context aware real time translation or context aware real time image processing where you might not want that context to ever leave your device or context aware personal assistant where you don't want your calendar content, your photos, your location leave your device, but you would like that to be considered in a safe, secure, privacy relevant environment. So great, what can you do about it? What does that mean to your career? Similar that Thomas called out. Be aware of these weak signals for industry specific use cases. Look for the signals for your domain or industry specific edge use cases. Because the first people who get used to identifying how to use these tools effectively will have a chance to emerge as leaders. Because the cost equation will force more and more companies to try to adopt edge AI tools. So the question is, do you see where they can create real business value? But for me, what I'm really excited about is an even bigger game changer use case. And uh, since I grew up as English as a second language and I always worked with global teams, I know the challenge that sometimes when we cannot communicate with each other, we lose time or sometimes we lose access to amazing expertise. So Tomas, I know you are very passionate about the global game changer real time language interpretation.

Speaker A: Uh, it's funny, my brother and I talked about conducting this in two languages and see how that works out. Lose half the audience at least. Um, I'm a huge fan of um, real time language interface and we talk about inference. So part of the problem with AI, the better AI gets, the bigger the models get, the slower inferences, which is the way it generates the output and it's getting better and better. So language is getting to a point of real time meaning you and I could have this conversation and you could be on the other side not understanding a word of what I'm saying and it would translate real time contextually and understanding the meaning and semantics. And that's really powerful. And it's powerful in many ways, in many situations. One way which my brother mentioned is in the workplace. I mean we or many of us work in global companies. Imagine that. And part of this capability is already in Microsoft Teams. Imagine that you can just conduct your zoom meetings and everybody's hearing and listening to the conversation in their own language. There's um, a lot of nuance gets lost for all of Those whose second language is the main language on the call. And it could happen if you're part of an American company, listen to English. But I used to be part of a German company, happened to me quite a bit. I sort of got every second word, uh, write in German, right? So complete fluency and cultural context and understand which is what AI needs uh, to bring to the collaboration. And then you can actually expect this kind of systems to provide insights saying there was an intent, somebody in the room didn't speak up and maybe you could have those kinds of EQ elements uh, to collaboration which I'm really excited about, which is, which is really pointing towards um, a better and more compassionate use of AI. So that's another big way we can scale because communication is everywhere. And that will be one of the biggest scales I believe of AI. In 25, um, you add one more about user experience. Arpa.

Speaker B: I don't know if you are a user of uh, tools like Perplexity. There's a lot of demand, a lot of discussion out there because, because there's a lot of demand for search and AI. A very intelligent search. And Google tried that with their first iteration of extending Google search. Now OpenAI integrated that with ChatGPT with their search. But Perplexity is still growing mainly because they made the experience amazing. How they rank, how they curate, the responses are still the best. Not because they have a best model, actually they are just using one of the other models, you can even configure it. But because they really thought through that. As a researcher, what would be the best way to share this information? And you see that many enterprise software leaders are thinking through how can they integrate AI into their experience. So one of the challenge, what does that mean? What that means that some of the winners might not be the ones that just have the best AI engine or, or a best data pipeline, but might be the ones who figure out how can they make AI available, how can they provide transparency about where the results are coming from in an easy, simple way. So what does that mean to us? How can we prepare for 25 when that will be a transformational differentiator? Some of the focus on user experience, user centric design, customer centric design, design thinking. When we think about end to end AI solution I think will be a lot more important. That goes back to the initial trend. Comment about Think of yourself as the architect, the designer and how can you help your teams who are experimenting and expanding AI solutions to figure out how can that solution seamlessly integrate to an existing workflow because that will have a huge impact whether something become a shelfware, a science experiment or it will be adopted in the enterprise. So for me the key takeaway as I step back of these three major trends, that AI is maturing and reaching great capabilities. The fact that we need to treat it as a true enterprise solutions and that it will drive transformation is that AI literacy is becoming even more important. I mean I recall we did our first joint session almost a year ago and it was about AI literacy as a key career accelerator. Just want to emphasize that still is and probably even more. The second is uh, focusing on understanding data and understanding what is the role of this data solution will become critical. But what's unique and what's a unique opportunity as AI will break down boundaries between business functions. People who can understand cross functional problems and can foster cross functional collaborations can step ahead and change management. I know sometimes it's overused, but I think it's going to be even more critical in FY25 as we go from experimentation to massive enterprise adoption. Because these skills, collaboration, uh, understanding data and literacy and driving change is fundamentally what digital leadership means. So as we wrap up Thomas, what are your takeaways? How you prepare or how you want people to prepare for FY25?

Speaker A: I reflect sometimes uh, about the Gen Z generation and we all talk about them as digital natives, um, and it feels like whoever's not Gen Z missed being digital native. I think there is an opportunity, I honestly believe there's an opportunity for all of us to be AI native. We're right here, this is happening now. So it's, you know, nobody should, 20 years from now, nobody should say well um, um, I was too late to it because there's a whole new generation that grew up uh, that was AI native. We all AI native. It's all happening to us, it's rapidly. So I think it's onto us to stay on top of this technology and apply it every day. I think compared to prior technology, I mean it was hard to learn blockchain and there were many other technologies that required a lot of technical knowledge. Because of the interface in AI it's actually learnable because you're primarily interacting through a chat interface most of the time. So staying ahead of the application, staying ahead of the application of AI, uh, and the use cases. I think it's all of us to stay uh, up to speed with this technology. So it's important for us to think that way. It's not up to the next generation, not up to the IT team. It's up to all of us. And, um, if you find this useful, please subscribe, share it with your friends, and, uh, send us any questions. Uh, we're getting some questions through various channels on podcasts and on YouTube and, um, email, so please keep them coming. And thanks for tuning in to Digital Vibration. Until next time, Philadelphia is checking out. San Diego.

Speaker B: Stay curious, stay adaptable, stay ahead. San Diego is checking out.