0:08

Hello, everyone, and welcome to another episode of the Leadership Growth Podcast. I'm your host, Daniel Stewart, along with my brother, Peter Stewart. And we have a fantastic guest and topic today. Bala, welcome to the Leadership Growth Podcast.-Thank you. Thanks, Daniel. And thanks, Peter. Really lovely to be here.-This is fantastic, especially as we talk about bridging that gap, and bridging the differences between the aspirational view of AI and the skeptical view of AI. I think many of us are on one side or the other or trying to figure out a little bit of both. How do we bridge from aspirational to skepticism related to AI. So let me share a little bit about Bala's background as well. Bala Muthiah is the Director of Engineering at Lyft, where he leads teams that power real-time decision systems for millions of users. I'm sure many of us on the listeners as well have used this and benefited from his guidance. An immigrant from India turned Silicon Valley leader, he's also a startup adviser, nonprofit board member, and a mentor across multiple platforms. And I'll also just mention, as we dive into this, the opinions that are expressed here, especially by Bala, are his own and not necessarily representative of his organization. So Bala, as we dive in, give us a lay of the land right now of AI. What's happening? What's going on? Kind of what's the sentiment? And we'll use that as a starting off point.-Yeah, definitely. This is something we have not seen before. I'm sure we have heard this so many times, but this time it's really, really I can internalize and feel it. When people... when I talk to people and people ask me about, oh, what's....what is happening or what is the definition of AI and what is the lay of the land I tell them this, by the time I start explaining and finished a sentence, already things would have changed. So my sentence is already outdated. That is the rate at which things are happening, whether it is technology innovations in AI or usage of AI.

Quick reference: 2:25

it took almost, I think ten years or something for AI, sorry for Facebook to reach their hundred million users. For ChatGPT it took I would say two months or something. So it's in the order of months where before we're talking about order of years. Of course, there were a lot of innovations happening before, but today there is a lot of energy and adoption, and that is in turn feeding into this innovation cycle. Everyone has adopted, they want more out of it. Every week, somebody is releasing a new model that is better than the previous ones.-Mm hmm.-So it is, I would say, one of the most exciting times to be in the tech space, to see this all happen in front of our eyes.-It is. It's such an exciting time. And Bala, we're thrilled to have you here to dive into this topic a little bit deeper. As we think about AI, and specifically we're thinking about it in the use in the professional scenario within organizations for leaders. As you've given that kind of brief lay of the land of the rapid adoption by AI and how it will continue to just evolve so quickly, let's look at the different perspectives or attitudes that different levels within an organization might bring. Looking at C-suite versus engineers who are down on the ground, like coding away everything. How would you characterize those different perspectives?-It is more like A Tale of Two Cities here. Like an example, I recently read a quote which resonated with me so much about this specific AI topic. This is nothing to do with AI though. It is about um universe and aliens. So I'll start from there and I'll come to this topic. So there is this quote, I think it's by James Clark or somebody. It talks about the universe and are we alone or not in this universe, right? Earth is Earth alone.

So the quote goes like this: 4:20

There are two possibilities. One, we are alone in this universe. One, the other one, is we are not alone. And both are equally scary. Just to see, oh, we are alone, we are the only one in the entire universe is scary. And the two, oh, we are not alone, there is somebody else there is also scary. Like, oh my god, who is that? How is it happening? With AI there's... the thing is both are equally scary. Same with if you take AI, there's one side which is saying, A splat of energy, oh, AI could change. We hear on the news, it could change, replace jobs, we could achieve one hundred X productivity. It could immediately turn how we do things in a very different level. On the other side, people who are keeping their hands on the keyboard, like you mentioned, the engineers, there is some skepticism. Because, hey, this can't even count from one to five properly. How can I trust this system to go into production? And to be fair, both of them are right. Both of these camps are right to an extent. Like that's where I feel leadership and community of leaders should come together and bridge this gap. Like, there is some reality to the excitement camp. These are the dreamers and we need them. And there is a realism when a code goes to production, when something is getting shipped. There has to be so much diligence and rigor. You can't just divide code something into production and let it go. Imagine you are going to a hospital and then something is given to you using this. Would you trust it? So we need to have some ground reality. And leaders have to bridge these two gaps because there is benefit overall. I would say it's a net positive transformation.-Building on that, how do you see kind of the top of the house, top of the organization different from the employees? And is that gap... How do we manage that gap? How do we pay attention? How do we recognize that gap? Because typically, what I find is the C-Suite, they have this mixture of excitement, and yet we are behind already. And we got to be able to see around corners. We got to anticipate. We got to make big bets. And we don't know what we're talking about necessarily, but we know it's important, and we need to then just keep going and going and going. And yet, for employees and engineers, it's often like,“This is cool and interesting, but hold on. What's really the impact for the technology and then the fear?” So it becomes more personal in some way. How are you seeing this by level? And how are we managing some of this aspiration and fear?-True, no, very true. You mentioned it's personal, right? Like that's the thing. There's a human element to it, and we should keep it in the center on both sides. And if you had asked me few months ago versus now, things... the both these camps are coming together now, like, the distance is reducing because technology is improving, people at the C-Suite or like lead... across the whole industry, people are coming to see, yes, there is a lot of great value, we can achieve great things, but they know studies are coming out talking about what is a realistic gain we are looking at. What are the downsides? And we see news where both sides are happening. So the C-Suite or leaders across in this industry are more open now to adopt the reality. Like they're thinking about it. Yes, I want to dream, but yes, I can see there is a clause. Same way, engineers and people who are on the ground who were once talking about what we just said, right? Like, hey, it can't even count one to five. How can I trust this? That is now getting a little bit okay. It is maybe it is starting to count better now. Things are, technology is evolving now, like recent releases have been better. So engineers are getting comfortable. Oh yeah, now I see. Now this is a point where hey, there are a set of things and there are few things it is really good at, and few things it is not good at. So this is the kinda framework leaders should put together and know where to apply. Just because you have it, don't use it as a hammer and hit everything. Right? The same analogy, it's just a lesson from history. It should not be treated as a single tool for every use case. So now it's where like whichever team, company, country, whatever organization is figuring this out rightly. This is an area I want to put in, this is the area I do not want to put in, they are the ones who are going to benefit. You can't blindly apply, and you cannot blindly not apply. Both of those sides will be at a disadvantage.-That's well put as you think about those core messages that senior leaders can share to help, you know, the reality of the emotions that many of the employees are feeling, and I think we've all felt those things as the confusion, of the change, of the rapid adoption, of the fears of well, what's happening? And so it's helping to ground that in the reality of what AI is good for, what it may not be as good for now, how we are going to leverage it. So there's a specificity that I think is helping to squelch some of the initial fears where it feels like when AI first came out it was kind of like this is going to be the universal answer for everything. And you just dump everything in it and it will give you the nice little package all perfectly wrapped and done. And we're finding more and more that to really use AI properly, it takes a lot of work. And there's a lot of pre-work, even before AI touches it, to give it the proper instructions, access to the right data, showing what the output is. So it's that work. So how do you, I guess as we're thinking about this human side and the reality there, how do we help those feelings, those emotions of people who are feeling left behind in all of the churn of it all?-True, unfortunately, if you look at across the world, even though we talked about how adoption is skyrocketing, right, in a very short time, there are people who are left behind. And there are people who are left behind unintentionally, meaning they don't have access digital and so on. There are people who are left behind because they are not sold this idea properly. They are not bought in. So to me, as a leader, even without AI in general, it's always getting your team's buy-in, right? You need to... work with your team and, like, team or people, company, whatever, like, get their buy-in. That's where they believe they are part of the decision making, they believe they are part of the transformation versus a top down mandate or a law or a policy where they don't have a say. So bring them in, make them part of the 'solutioning', because this is actually not a problem, this is actually an exciting thing. You want to apply the solution and bring them in. We will be mind-blown on how diverse perspectives will come and make this easier for everybody.-I love that idea. And yet, when you bring people in, that takes time. And time can be one of those things that we're all feeling that we don't have because things are changing so rapidly. And it's this inner need of, I'm already behind. I don't know what else to do to be able to try to catch up.

And I'm wondering: 12:09

what are the mindsets that leaders can have to be able to manage that perception, the feeling that I'm going to be perpetually behind? And yet, in order to adopt something new, we also know it's best to actually involve and collaborate, and to work with others to increase the buy-in, increase the accuracy of the potential solution. How do we manage that ourselves as leaders?-Yeah, it's true, right? Time is not in our favor because things are happening so fast, so rapidly. We have tried this across different even startups I consult, and by the time they put together a system, that system is already outdated. So no need to go... The rollout process, the traditional rollout process, adoption strategy, is not going to work in the new world. Having said that, we still have to follow a principled approach. Otherwise, it's going to fail. It's more like the crawl-walk-run methodology still is going to be true. But letting people know that yes, we are crawling now, and then we are going to walk and then run, will get that buy-in or make them understand. Because everyone wants to run from day one and you will stumble. This is another example from the past. I keep going to the past because I feel all the answers are there. We don't have to reinvent. Like when steam engine was invented, it was one of the biggest innovation step change in the industry, in our history.-Mm hmm.-They had to build the railroads. They couldn't put a steam engine on the road and let it run. It's going to fail. It's going to either damage the road or it's going to just fall. You need the rails. You need to put the rails so the engine can really do the job that you want it to without damaging anything. Similarly, for AI, you need to prep. Need to build the foundations, have a more structured approach, like running pilots, learning from it quickly, but give a timeline. Like, do not open as like, oh, this is a three-month pilot. Nobody is going to buy a three-month pilot anymore. We're talking about weeks. We're going to run multiple pilots. That is the beauty of it. We can afford to run multiple things now in parallel. Run multiple pilots, get the learnings and quickly too. And build in some frameworks. Like we talked earlier about identifying the right things to do and staying away from the wrong things. It's very vague. If you just tell the team, hey, figure out what is the right area to apply, what is not the right area to apply, they are going to be confused. Because it's also personal. Like Daniel, like you mentioned earlier, what is easy to me may not be easy to you. What is easy to you may be incredibly hard to me. So as a team, we need to come together and have a framework where we are going to apply on these areas. And then this, and these are the areas we will never go. That intentionality will make it easier.-That analogy with the steam engine, I think, is so powerful of recognizing the effort needed to lay the track so that the engine could do the work effectively. And as I'm hearing you describe the change process with it, there's an emphasis on timing and speed, and that we don't necessarily have the luxury of, as you say, three months of trying things out. It's like, no, we need to know in a week or two. Let's kind of pull on this thread a little bit more as you are talking about how typical change management processes in the past don't work as well now. Besides the speed component of it, what else do we need to be aware of as we're working in today's environment to change?-Like, the speed is not going to work. Another thing to keep in mind is the data that we are dealing with, right, the sensitivity, meaning you can go fast, but if you break or if you cause a damage, it's going to be very hard to recover. Like it could be a... it depends upon the company, industry, there are so many things. If it's a sensitive data issue, like it's a very different scenario, right? So that is where talking about areas you want to touch, you don't want to touch, like skills you want to deploy, we do not want to deploy. Like, I have a little framework which we have used is just to identify isn't the question we had was what tasks to put in? Like we were trying to do traditionally, right? Okay, left to right, how do we develop? Let's go start with that. Okay, first is getting a requirement ready for building something. Let's start from there. Then doesn't work. We unlocked this framework and really helped us move fast, which is anything that we do, like it could be categorized into two buckets, right? One is whether it is fun or it's boring. It's very personal, like that's why I like it. Right, it's hard to put it in, but, like, we can take a task, it could be fun or boring, easy or hard. So if you put this into an axis, easy and hard, fun and boring, right, then you can go after the first one, which is easy and boring, right? These are tasks which are easy but very boring. It's like me sending a status update every day, or somebody putting together something that some other person can read or validate, a documentation, classic example. Easy, boring. No one wants to do it. Start there. Get the excitement, right? This dopamine, oh, this used to take me two days to finish the documentation and now with AI and proper guardrails, I can do it in like couple hours and validate it. So that gives them confidence. So, then from the easy boring you go to easy and the fun part. Sorry, yeah, easy and the fun part. These are things you really want to do. It is fun, it is easy, but hey, I can give it to somebody, right? Like, then go there. And then you can go to the hard part, hard and boring. Unit test case. And like those are... I wouldn't say it's easy, but things are hard, but I don't want to do it, it's... I... we preserve intentionally hard and fun because engineers are... Like they came... they became engineers because they want to be challenged. They are curious. They want to solve problems.-Mm hmm.-It's hard problems, fun problems. Do not outsource that to AI. Do not move that to AI. Don't let your AI design your system. Like system design, engineers love it. You go to a room, their eyes light up when you talk about system design. Oh, what can happen? How can it happen? What are the things we can bring in? And it can go forever. So, the hard and fun part, preserve it, do not give it to AI. And that's your thing, and that's where creativity comes in. That's where you need to use the brain, and you don't have to outsource that to AI. So this framework helped us quickly identify where we can go and run it all in parallel. Like that's the beauty before... running it in parallel was impossible. Every pilot or experiment would take time, but now it does not. You can kick off multiple things and quickly learn and share and do. So that's a beauty. There are so many benefits we can tap into and make this process seamless.-Love the simple approach to be able to then look to see easy, hard, fun, boring.

And I love also what you're emphasizing: 19:38

that hard, fun stuff, that's usually the most complex, the most challenging, and what we as humans can get into. And can dive into. We could always use support and assistance in various ways, but reserve that for what our brains can do, not just on our own, but together as well.-Mm hmm.-And there also... I also love the idea of having multiple pilots instead of just having one kind of waterfall approach pilot and then we learn, then we start. But no, we can have many going on. It seems to me that there's a foundational piece under all of this, and it's this AI literacy. This AI understanding. And so, Bala, how can leaders be able to have enough? And keep up to date related to this AI literacy? Because I'll say, as we've had several fantastic AI experts, so to speak, on our program, they are seemingly the ones who are most comfortable and not kind of as scared of AI as those who maybe don't have as much understanding or background, and therefore it is... it's less certain, and it's both exciting and scary a little bit more because they don't have as much understanding. How can leaders gain and increase their AI literacy so that there can be this sense of calm? They have a sense of what's going on. How do we get there?-Great question. It's mainly for leaders. It's very important because they are in a high leverage, high impact position, right? But it this is common to everybody too, like all the way up in the chain, down in the chain, everyone. I have fell in this trap before, like excitement. Like few months ago, when things started coming out, pretty much every day I want to go try something, I want to do something, I want to learn something, read something.-Mm hmm.-At one point, it... came or pushed me to a burnout situation. Not only at work, but even generally outside happening on like, hey, there's new technology here, there's new... they built this browser over a weekend, things like that. What helped me ground is to realize it is impossible for you to catch up. So stop doing that. Number two is you don't have to catch up. That was my mental space that I got back. Okay, things will happen faster. It is all right. The core principles are always going to be the same. So pick something and then continue to build your foundations and learnings and stop catching up with everything. It's going to be very, very hard. Like another example is back in the days when streaming was not there, cable was very minimal, you would probably watch one or few shows and you are already good. Now when streaming came there were so many shows. Suddenly, you wanted to watch everything because your lunch conversations could be from anything, right? So every team member would have a different show to bring in. Before you would probably have two or three shows that everyone is going to talk about. Then you started watching everything and then you will know, oh my god, I don't have time for anything else, I'm just catching up. And then at one point you stop like catching up on those things. That's the same thing. Like stop worrying about being on top of everything. You don't have to. If it's really important, it'll come to you. If it is so consistently, oh my god... The fear I had was, oh, the FOMO of am I missing something? Am I missing something very important that I can catch on early if I miss. Like I've come to realize if something is that important, it will come to you because everybody around you is going to talk about you, talk about it. You will know and you will get a signal, and then you can catch up. Have more principled learning if you're starting to learn now. Like, let's say you're... it's like drinking out of a firehose. You cannot, you will like suffocate. Have a plan, what do you want to achieve? And then pick a lane and go. It's if you learn one, it's going to be transferable. You don't have to worry about staying on top of every single thing. And that's where as a leader another thing is you can deploy your team. Like everyone in your team can do something, then you collectively learn. Like you don't have to be the only one in the room. Now new people coming out of college are much more advanced with some of these tools because colleges started adopting now. So they are using it day to day than a company is able to catch up.-That's such sound, sound advice. And as you were sharing about the, as streaming came out, and you know, feeling that, oh, I'm still behind, I can't watch everything. I'm reminded when Netflix came out and kind of opened this streaming platform. You know, I was in grad school and there were colleagues who were like, their goal was to try and watch everything on Netflix.-Yeah.-Like, to watch the entire catalog. And you hear that now and you just think how ridiculous of a notion. Like it's an impossible task to do. But I think that quick little analogy helps to just calm the thought of, I need to know everything. And it's picking a place to start and then building on it because there's so much analogous in terms of tools and skills and tasks. If, for example, if you're learning to write good prompts on one LLM, odds are the principles behind that are going to help you as you transition to a different LLM or working with an agent or whatever it might be. There's a lot of transferable knowledge there.-Totally. Totally.-Yeah. So let's think for... from an organization perspective as we're trying to help reach some of those skeptics or those that may have less AI literacy or experience, what can an organization do to start to promote AI literacy?-Yeah, we'll start from what an organization should not do, right? Like and then we go into what they should do. What they should not do, which some of the companies are pretty good at doing, they should not do's are... in like having a top down mandate about oh, I want everyone to do this by this date, right? Without proper systems, without proper training. There's this Goodard's law, right? When a measure becomes a metric, it ceases to be a good measure. So stop measuring the wrong thing. Don't measure things like adoption without having impact, without having a focused outcome. Just adopting is not going to work. Like I've seen where, oh, you have to use this tool, then everyone opens the tool and closes the tool. We have seen this in corporate learnings. Oh, all of you have to go through this ten corporate learning mandatory courses. We all know what happens there. People open the window and do their own thing and then they click next when it's time. You are not doing a... You are not doing justice to the training and also to the individuals. So stop mandating adoption without outcomes. Number two, what they should do is emphasize on learnings, especially when you are exposing someone to something new. You need to create a safe space. Tell them it's okay to fail. Hey, we will pilot this tool or this way of doing thing in this project and it's okay if this project get delayed by ten percent or fifteen percent because you are taking time to learn, you are sharpening your axe before you cut that tree. So it is okay to fail here and then we learn on the next one the dividend will pay itself in the next second, third, fourth projects. So create a safe space where they should know it is all right to fail when they are piloting and experimenting. And then acknowledge public learning. Like we don't do that. We as professionals, we always want to come and perform. We don't spend time to learn. If you look at sports nobody goes to Super Bowl and says, oh, we are going to run this play which we have never run before, let's do it. This is all like practiced plays, rehearsed plays And that's what wins, right? So when you are trying something, make some space so people can learn and then you get the learnings and then you go perform. Within organization, we don't emphasize learning a lot. So make space so people can learn and then you don't have to keep selling. People would be already coming in with that mindset of I want this and I want in.-Let's continue this one more step. If you had to say the one thing. As we're kind of wrapping up here, what's the one thing that leaders need to keep in mind to be able to fully optimize their understanding and utilization of AI? What would that be, Bala?-I would say be curious. It could be on either way, right? One, be curious about what's happening, how things are being done. Be curious about failures. If someone says, I don't want to use this tool, be curious and ask why. Like why are they not able to use? Is it the tool is not good or they are not trained. Like being curious without worrying about oh, I have a target I'm going to do. That is... yes, as a leader, you have targets. That is what the job is for all of us. But be curious. That will help in so many different ways. It will help you bring new things to your team. It will help you listen to your team and understand the pain and then you solve it. As a leader, the number one job is to help your team be their best, and tie that to a business need. So being curious is the only thing. It's not new for AI, but that is always going to be the winning formula. Like, if someone is curious, they're going to be open, they're going to be learning, they're going to be improving and continuing to do that.-Yeah, well said, well said. I feel like we could continue this for a few more hours. Bala, thank you so much for the tremendous insight, practical tools and tips, as we're trying to leverage and bridge that gap between the aspiration, the hope, and the skepticism and the challenge of AI. Thank you, Bala Muthiah, for being part of Leadership Growth Podcast.-Thank you. Thanks, Daniel. Thanks, Peter. It was great to be here.-And to all of our listeners, please like and subscribe. And thank you for joining us for today's episode. We look forward to having you in the future. All the best on your leadership journey. Take care, everyone. Bye. If you like this episode, please share it with a friend or colleague. Or better yet, leave a review to help other listeners find our show. And remember to subscribe so you never miss an episode. For more great content or to learn more about how Stewart Leadership can help you grow your ability to lead effectively, please visit stewartleadership.com.