[Music] Welcome to another episode of Tech Unhinged, where technology meets human. Today I’m joined by someone who doesn’t just talk about the future of AI. He’s helping build it inside one of the most influential companies in the world. What sets Satard isn’t just his technical background. It’s the depth of experience he brings across industries, streams, and technologies. From mentoring students to leading enterprisegrade AI projects, he’s built a reputation for turning complex ideas into practical high impact solutions. Welcome to the show Sad. Thank you so much. I really appreciate being here. Well, Sad, I want to remind here for a second. You started out as an executive engineer in telecom in Islamad. Today you are building intelligent agents at Microsoft. What shifted in your mindset or worldview that led you from traditional systems to intelligent systems? Thanks for this question. You know, I’ll go a little bit about uh the origins. You know, I’m originally from Pakistan, born and raised, you know, loved being in that city. I loved Islamabad in a way that in ’90s, you know, it was uh beautiful, not so much traffic, which has changed a lot. Yeah, it’s been amazing. Regarding your question about from the journey from EU to Microsoft, how I see is like we’re like the world is getting abstracted, you know, in some ways, right? How I view like even engineering disciplines is like you start with mechanical engineering you you go on top having electrical engineering um electronics computer science and now we’re abstracting that to AI right so the it’s been a journey it’s been a journey from last couple of decades that every engineer is working on different layers of their infrastructure that goes from even starting from civil and mechanical engineer all the way to what’s coming now is AI, right? So, working at EUON versus what I’m doing right now, it’s not pretty much different. At the end of the day, you want to help business succeed in some ways, right? Interesting story. You know, while I was doing that, I was part of uh core network planning at Euon at a very, you know, was like it was just getting out of college and I got this amazing position where I was part of the roll out of 3G and 4G. So if you remember back in 2013 2014 that was the time when 3G and 4G was rolling out you know and I was that person and I’m not like this is true and everybody could say that I was that person who was there when it was getting switched on in Islamabad Karachi Lahar I covered five major cities when 3G and 4G went live at EU like in the network and I was reporting to my CTO and that was the time you know as part of the history for the first time where uh we’re working on that but from that perspective you know I feel like technology has evolved a little bit right that was a momental shift in going from circuit switching which has been primarily how network mobile networks used to work to a little bit of packet switching right and now we talk about AI and how the whole you know it’s going to get people out of job and everything at that time I was seeing in person that people who had so much experience in circuit switching wanted to adopt packet switching which is 3G 4G you know it’s like the internet was coming it made me realize at that time that that’s an evolution you know in some ways that how technology evolves markets get saturated and then you have to move to the next layer so the same way now working in AI and building AI agents you know I feel the same way I’m working on a different abstraction layer which is the top AI layer things scene. We’re going from simple systems, databases, data warehousing, you know, all of that stuff which I’ve been working with in traditional, we call it narrow AI, the use cases of machine learning to more intelligent systems, which is, you know, transformer models uh with whatever is going on in AI. On top of that, we’re building AI agents. So, it’s part of an evolution of technology, different abstraction layer, but that’s how technology works. Well, uh, moving to our next question, S, you were an MBA and, um, you know, UT Austin and hold a masters in computer science. How did those two worlds, business and engineering intersect for you when working on AI strategy at Microsoft? I mean, that’s a great question. You know, a lot of engineers think that they’re building cool stuff. And I agree, you know, we’re building cool stuff, but at the end of the day, business drives technology, not the other way around, right? If there is a business use case that is there, people want it, then we’ll build technology around it. We have not done in a way that hey, I’m going to build something cool, but there’s nobody to sell. So, you always have to think about how business impacts technology. If there’s going to be buyers, we’re going to build technology around it, right? So, that’s one thing. Second thing is on my personal journey, I always wanted to understand the business world. you know growing up I wanted to have a world view of how everything’s working how the world is working from geopolitical things to economics you know how different uh trade things you know all that stuff happens so I went through the journey through technology you know so as I said that abstraction layer in my uh uh last it’s like I’ve done electrical engineering right in my bachelors so that provided me an understanding of how the systems work beneath and then I have a masters in computer science which talks about more application layer, right? building on the hardware and that drive me a little bit of like understanding the stack and then complementing that with with an MBA provides me a 360 view of business world right at the end of the day you are trying to work I mean we call it in capitalism you work for shareholders right the shareholders for a company you want to make them rich that’s the crux of capitalism but understanding the holistic 360 view of business how what I’m doing is going to impact our customers. It’s going to impact our business outcomes and that is a key thing. So I think I mean I I used to like read uh business related books before but I think MBA provided me a holistic 360 view of from marketing to finance to valuation you know how companies are being valued, how stock market work, how a business becomes profitable from startup world to very mature enterprises like Microsoft. So it provided me with the skill set and the essential toolbox to understand business understand what engineers build impacts the business. So yeah for me I think you know uh the journey has been and it it helps me a lot when I work at Microsoft talk to my clients and partners understanding their business need rather than hey I have a cool thing to build and give it to you and that’s not how how the world works. So when you talk to a customer, when you talk to a partner, you need to understand where they’re coming from, what problem they’re trying to solve, and then you come in, that’s how we solve it, and how much business impact is going to happen. Yeah. No, that makes complete sense. But uh s do you think that in your profession this gives you an edge over other people who are, you know, just engineers or maybe they, you know, they haven’t been exposed to the business side of the world or does it not make a difference? It does a huge difference. I mean I tell everyone it’s like you know even with engineers you know with Microsoft or outside my friends it’s like hey you always need to start understanding the problem or the business that you’re trying to solve or impact right without that you’re just like uh is shooting in the dark you don’t know what you’re trying to solve right so it’s imperative to understand and you know you brought it up you know it’s like we talk about how AI is going to take the low-level engineering jobs you know for example me who has an MBA let’s say or somebody who’s a business leader with AI building has become so easy you need to be an architect you need to be a business solver you need to be a product manager who understand and can build technology and let AI agents build it so in that way that one person can do 20 plus people’s job how AI agents going to build what he wants to build so for the viewers you know for for your viewers if somebody’s like student or coming out of college it’s imperative ative to understand the business world without that building has become so easy you know everybody can I mean AI is way better than coding and building than a normal person so you know you’re also working on one of the major agentbased system for one of the world’s largest energy companies without disclosing uh what you can what’s something that surprised you while building agents at that scale that’s a great question there’s a lot of noise around uh what AI agents are what AI is and you know nobody like pinpoint. So I’ll take a step back and you know I’ll just uh explain a little bit like for me what agents are and I’ll give an analogy of like driving you know you know self-driving is coming. I live in Austin where Whimo and you know Tesla’s headquarters like 5 miles away human driving will be obsolete in some ways and that’s why I want to give this analogy. It’s like when you think of AI agent think of human driving where human is a base model having a brain that can do certain thing that is our AI model. Then you have two other things. One is skill and one is tool. So you have this model, you have this skill which is driving itself. How to drive, right? And then tool is a car. So for me an AI agent is three things. One is the base AI model. You call it a human brain, call it a driver. You have a skill how to drive in in terms of building AI agents. It’s the prompt that guides the agent. And the tool that’s got invoked in the tool is a car. So in the basic building block, this is what an AI agent is. And when you build this, combine this, it’s autonomously can do stuff like I can go and drive my car. I can do it autonomously without somebody’s supervision, without human supervision. So on the base level, that is that is that building AI agents are the easiest thing to do in today’s world. But taking them to production, that’s the hardest part. I’m going to repeat that because uh last week I heard from um you know Langchain CEO Harrison Cheese and he had a big conference and he had that on the on the screen. It’s like building an AI agent is pretty easy but taking that AI agent to production for the enterprise use case. That’s the hardest part. That is where everybody’s going to focus on. And I’ll explain you why. Building is easy in a way that you know it’s like more of an engineer mindset. It’s like hey I’m going to build this cool thing. But do you have put governance in there? You have put responsible AI in that. Do you have put security and access control? What business problem it’s trying to solve? Is it making sure that there guardrails and controls? You know what if it’s going to harm somebody, right? How much ROI we’re going to get from this AI agent? How much productivity we’re going to get? So when you think about building an AI agent in enterprises, you have to think about all these things holistically. how the system is going to interact and with humans as well right so working at Microsoft building enterprise AI agents one of the key things that we look at start from business what problem we’re trying to solve how AI agents going to solve it but then seeing the holistic architecture building it governance all that stuff that I’ve mentioned right this is a year of AI agents you know we called it 2025 is the year of AI agents we have seen enterprises big enterprises are trying to solve that Now to your question about energy sector you know or finance sector or any other sector you know like things like white collar job or like computer jobs yeah they’re going to get revamped but energy where people have to work on transformers where where they have to like fix the machines that’s like very mechanical how AI agents going to work on that believe it or not we’re going to do that adoption there like imagine a maintenance engineer where he has this AI agent taking a picture problematic transformer and asking an agent hey what should I do to fix it right so it’s not only going to get the internal documentation and everything it’s going to guide him like hey you need to do this this this this right and do it right in some ways that is the first step the second step is where bots’s going to get introduced you know where they’re going to do all that heavy lifting so I’ll give you one example you know self-driving cars are coming we know you know it’s like in one or two years human driving is going to go way, you know, it’s like they’re everywhere. Back in 2014 and 15, you know, there was Cruz is one of the GM’s subsidiary which it went bankrupt. You know, there a lot of competing self-driving companies are working. But back in uh 2014 or 15, there was an incident that happened in San Francisco where cruise car actually dragged one passenger almost 10 miles, right? I mean, the state revoked their license to operate. one small incident and you get revoked. Now imagine you’re like running a billion dollar business. One small mistake can ruin everything. So that’s the world we going to live in where responsible AI these guarders and controls are really important. AI agents can’t run without clean governed data and you must have seen this in your expertise as well. You’ve led architecture at PepsiCo and Pool Corp as well. What’s a reality check most enterprises need when it comes to AI readiness in their data? I think I would say that your AI strategy is your data strategy. Garbage in, garbage out. A lot of people say that it’s like if your data is not clean and kosher, you’ll have garbage out, right? Which is not going to be personalized, contextualized and everything, right? That’s why you know even Microsoft we had this Microsoft fabric you know that talks about the whole data strategy putting into data lake making it nice and clean and then put AI on top of it and that’s part of a user you know customer getting matures terms of their AI journey. So absolutely data is the most important thing. Data is going to be the king always. You guys all know that we ran out of data to train AI models, right? There big companies that are training the models on synthetic data which is all madeup data because we ran out of it. There’s no other data that they can train the models on. We’ve already done that, right? So data is always going to be the king. Why? Because it’s highly contextualized. You ask Chad GPD right now, hey provide me life advices or whatever. Chad GPT just puts some random stuff right but you as Rabia you know you want to understand how it correlates to your life and that’s a data in your brain that provides that contextualization and personalization. So hands down if I could convey one message through this it’s like the next couple of years where things are going to be are two things personalization and contextualization that’s going to drive everything because we have generalization general models everything you want things to be contextualized the thing that you want to ask about and personalized to how it correlates to you. So if you ask about life advisers because I brought this up because you know OpenAI just sent a huge report. It’s like a lot of people are using Chad GPT as a therapist and life. Oh yes they are right. It’s not contextualized and personalized. It’s going to tell you things about a CEO should know but you’re just a college grad. You’re just trying to survive you know and it’s not personalized and contextualized in some way. So do the viewers stop using it as a therapist and a life coach yet? And that’s what the strategy of open AI is. You know it’s like Sam Alman said that I mean he didn’t say that he implied that they want to understand your life. That is the key from Google to Microsoft is not consumer focused that much. From Google to Apple to open AI all these big companies meta big they know everything about you. All their goal is to understand you. They already know that. But putting that into a model so that they understand everything about you is going to be the key to win because you’ll have these two things personalization and contextualization. Yeah. Yeah. The other day I was reading how you know meta agents are going to take over the ad agencies closely by 2026. So you know we are seeing a lot of changes in in this landscape. Um so you’ve also saw you’ve sort of um touched upon this but if we specifically um you know talk about this that it’s generally seen that businesses care more about contextuality and accuracy than creativity. So how do you design AI agents that reduce hallucinations especially in high risk domains like finance or energy or other sectors? And that’s a great question in a way that for businesses I mean it depends on the business. If you’re working in film industry, you know, they need more creativity. But if the business is working in, you know, it’s like finance domain, they want to make money and they want to be very quantavy in some ways. So AI agents are a great example. You know, last year when nobody knew what AI agents, you know, AI was coming, hallucination was a big problem. AI agents are in some ways the answer to mitigate hallucinations. But there are different techniques even in the agentic world. I would bring up three things you know first is human in loop you have to have human in the loop because for edge cases for ambiguity and vagueness you need to add human second is you add a reflection agent right it’s one agent making a decision another agent is a critique or a reflection agents understand the decisions of the output you have made and made sure that it is cool so you add flection agents and feedback you know it’s like you can add a feedback agent that reviews critique the answers and see verify one more time. Right? So there are different strategies in in in the world of hallucination. I would not say that it’s never going to be the case. But even with AI models, you can change the setting of a temperature. Temperature is a gauge how the model would go more creative versus more accurate. You can gauge that all the way down to more accurate wherever the use cases. but add these reflection feedback human and loop agents to make sure the outputs are are good. Since you’re operating from inside Microsoft’s, you know, cloud and AI group, what is one underrated strategic move Microsoft is making in the agent space that you think most outsiders miss out on? I’m very bullish on Microsoft even before joining Microsoft. I bought a lot of shares 10 years back, you know, because of their strategic position and leadership. I’m going to give huge credit to the CEO Satan is one of the best CEOs in the modern world that we have seen operated from empathy to understanding really well where he needs to take the company from where it was after Steve Balmer. We had like three CEOs from Bill Gates to Steve Balmer to Satya and I’ll talk a little bit about his personality his perception. I’ve read this uh his book uh hit refresh you know if you want to check it out great you know coming from India very simple background in that way you know his parents were in civil service philosophy and all that stuff that that created a little bit of his worldview of like where things are and empathy was a core part of it and I think that leaders ideas and culture impacts everything I always say that you know there are two things that impact the the company It’s the mindset and it’s a culture. They’re two things. Anyway, you can get the smartest guys in the room, but if there’s no culture, there’s no mindset, nothing. He’s one of the godfathers of capitalism. He said culture eats strategy for the breakfast. I heard that one. Yeah. Right. It’s like if you don’t have a good culture and a mindset in a company, your strategy is nothing. So it’s very imperative to understand what Microsoft core principles are based on growth mindset based on think win-win and go up the abstraction ladder. I’ll give you two examples what they did. They embraced one of the huge upset in the Microsoft history when Microsoft which is a closed source company embraced Linux open source. They’ve been competing for a while, right? Microsoft whole strategy was compete, compete, compete and go up the ladder and you know it’s like make sure your enemies are are not winning. He changed everything. He collaborated with even Apple to build apps that are focused on iPhone doing collaboration with Nvidia, you know, open AI. It’s like collaborating with everyone. Everyone go to the platform level. Go to And what’s the Microsoft mission? It’s empower every individual and company on the planet. When you say empower, that’s a huge message. So what Microsoft did is went down the abstraction layer and telling we’re going to empower everyone from customers to partners to everyone, right? We’re going to collaborate with you. We’re going to make you successful because when you’ll be successful, we’ll be successful. Think about with the whole AI boom, which is the company that got the most out of it? It’s Nvidia. They’re making they work at the low level. They work by making these AI chips which everybody’s going to use and they’ve got their business. So Microsoft in that way in enterprise scale is the same level. We’re building platform for developers. We’re building platform for customers from low code to pro code anything you can come and build build your business and take the most out of it cuz it’s the same thing you know like I always said life is not a zero sum game. A lot of people think that life is a zero sum game. Life is never a zero sum game. Life is a positive sum game. You give you get the it’s not a pie. You don’t take a piece of pie, it’s going to go less, it’s going to go bigger. So once you think about from that mindset, you know, success gets you in some way. So in in that way you know if um talking about Microsoft strategy even AI strategy what we are building is we’re building for the developers for the customers for the partners to make them successful from copilot copilot studio to Azure AI foundry in AI space these three things go from low code to a little bit pro code to all the way custom pro code and you can mix and match and that’s the architecture that we were trying to build as well you know it must be really exciting to be working with all of is a lot of people throw around the term the jargon we’ve been listening digital twin for AI agents but realistically speaking how far are we from an AI that can actually operate on our behalf while we are asleep yeah I mean that’s a great question I think uh we’re pretty close in that way I I’ll give you one example uh Reed Hoffman from he’s a co-founder of LinkedIn Reed Hoffman last year created his own digital twin and you can check it out you know he talks talks about he talks to him refining ideas and everything it’s like you know he has put a lot of videos you know so it’s called read read Hoffman AI you know with video and everything so he trained his model gave everything that he knew the books that he has read the worldview that he has his pictures and everything and created this model or fine-tuned that model for his personality and he talks to him like check it out and that was last year right in my opinion we all will have one AI clone with us all the time. We’ll have a minions, we have multiple clones, you know, one would be for work, one would be for the party clone, the the work clone. So, we’ll have different personalities of these clones that would work for us, right? And you can pick anybody and you’re like, hey, talk to this guy, you know, and that’s part of you. That’s the world we are going to live in. It’s going to be the same for enterprises as well like uh Satya mentioned that in the future we will have humans and the swarms of AI working together with different personalities and everything. So we’re pretty close. I feel like in a year or two I have if you would ask me I have created one GPT that I put a lot of personality of me around how my writing styles are from LinkedIn to Twitter to a little bit of Instagram and everything and sometimes I ideate like hey how would I write to make sure that you know it’s like I’m fast on like my writing style my thinking style everything and somebody helping me. So we’re going to give all this knowledge to AI of course to build all that stuff our personalization and that’s the great way of like uh how you know Sam Orman said that we want to understand your life right we are willingly giving that information the same way Instagram is taking our information of like hey we want to put cool pictures on Instagram and they’re learning Google you know it’s like hey we have to use Google in searches but Google understands you way more than you do probably in some ways right oh So with AIE1, we’re going to give all this information to these companies so that we can become productive. There’s no way out. But in that way, we’ll create these digital twin, you want to call it, AI clones, you want to call it your virtual things. And guess what? They’re going to live forever. That adds the cherry on top. I was having a discussion with my dad, you know, who was a teacher and he does not know what I do in some ways. And I’m like, “Dad, you’re going to be fine. You know, we’ll create a digital clone of you. you’re going to live forever in the cloud. He was like, “What?” You know, I don’t understand what you’re saying, but for your viewers, you know, I would appreciate if you guys watch Black Mirror on Netflix and movie her, you know, where this guy fell in love with the OS and that was back in 2014 2015 these things came out. There was we didn’t even gener was not even a thing you know at that time that’s reality now people say you know there have been reports that people are falling in love with child GPT oh yes they they give more credit to charge GBT in comparison to a human sitting next to them human relationships are not working lately also actually you know my my next question where we can del more into the side of things so you know we we’ve seen Hollywood has long imaginated AI agents that comfort evolve and then love but in real life especially in an enterprise agents are measured in KPIs and not feelings. Do you think we’ll ever see a convergence between um emotional intelligence and business intelligence or are we building an entirely different kind of intelligence? a lot of industry experts you know the head of meta lon you follow him you know and a lot of other people including you know like Samman who’s building open AI but he’s not a researcher but a lot of researchers say that LLM’s you know large language models might not be the way to AGI in some ways that would have that emotional understanding uh they can mimic pretty well so you see a lot of fine-tuned models there’s a model by anthropic I think Tropic or Google deep mind Salomon Musafa Solomon you know who’s now the head of AI at Microsoft uh his previous startup was this bot pi which was emotional intelligent AI you can check it out you now talk to him it’s like pretty emotionally intelligent to understand feelings to understand the tone and even with chat GPT now you can talk it understand human emotion in some ways like the pauses if you are sad it’s going to stay different if I’m excited I’m going to say something else right So, we’re going to train these AIs to be as emotionally intelligent as possible in what use case we’re using. That’s going to be the key, right? Not every model is going to fit, right? It’s the same way people have different preferences. You know, you like certain foods, you don’t like certain foods, you like certain cars. The same way there are models built for different things, right? We have 01 series, 03 models which are, you know, by OpenAI they’re like more reasoning models you can say. And we have 41 42. They’re generic models. So there’ll be a lot of different models. It depends on your use case. From businesses will have more accuracy from Hollywood perspective, you know, from our personal lives will have a different fine-tuned model that would understand human emotion. Arthur C. Clark, you know, fiction writer back in early 90s, you know, so he said any advanced technology would seem like a magic and it is now. You know you talk about fonts you know seemed like a magic to my uh parent generation right you have uh spoken at UT Austin Texas and even NUST in Beria back home what are students asking you about AI agents that seasoned professionals are not and what does that tell you about the generational shift in technology a lot of people want to be nice and they don’t want to put that fear into the world that hey you know it’s going to take your jobs and anything but would have to be very realistic and I’m going to play my part in being realistic and telling the next generation that things are not looking that great for you, you know, if you’re college grad. And that’s not me. It’s all the industry expert. You know, Enthropic CEO just recently said, you know, it’s like probably at the end of this year or next year, you know, we will have models and agents that’s going to perform way better than entry- levelvel jobs if you’re coming out of computer science background. So, it’s a shift, right? I’ll take a step back. you know, like we we talked about a lot about AI, AI agents, you know, where who are we humans and what are we doing in this world and you know, all that stuff. 200,000 years of human evolution that we have gone through to build these bodies and brains, right? This is what humans have been. We’ve been through multiple pivotal evolutionary moments like the invention of fire that helped us make our brains bigger. We had the invention of language right that changed everything in a way. We had an invention of writing that we can understand the history right the hunting domestication of wheat one of the most pivotal moment in our history is like going from hunter gatherers to domestication of wheat which is now you have an abundance of food at your home and you can procreate population going great. So if you see from that grander lens you know we call about 15 billion 13.7 billion years old whatever but human evolution 200,000 years we’ve been here on this earth multiple cycles of evolution. This right now is one of the pivotal moment in our human evolution. We’re going from more humanlike intelligence to another form of intelligence and eventually we’re going to merge with that intelligence. We will have some remnants of humans that’s going to remain with us and we have already been merged. You know it’s like from phones that become extension. We have neural links that’s going to get planted and then we’ll have that we’ll have our bodies augmented with bots and eventually we’ll we’ll upload ourselves to the cloud and live forever. Right now there’s talk about consciousness non-conciousness what makes us human and everything. In my point of view human body is not perfect. There’s a lot of flaws you know you fall you know I had ACL surgery you know I’m not the same afterwards. There’s lot of limitations of human. Why not to augment having another eye on the back? Why not to have another forearm, you know, why not to have another leg that that’s going to help you? Why not to have a a wheel with you that you can travel anywhere? So all these augmentations going to happen. So that’s why I took a step back and you know provided like in a grand scheme of things we are going through another pivotal evolutionary moment. It’s one of the kind that we have an emergence of another intelligence that we’re going to merge with eventually, right? So, we should see it from that way. It’s part of evolution, you know? It’s like I’m a little bit existential in that way. 60 70 80 years of human life, we’re just like fireflies. We are fireflies and that’s it. Now is the time where next part of evolution with that particular context. You know, if you’d see and zoom out the grand reality of things, your relationships, the things you’re doing, you’re concerned about every day. Think, you know, we’re going through a big shift in human evolution working. And then, you know, you talk about displacement of jobs and everything in a small scale, how many people before three or four decades, how many people were working on their computers, right? It’s just like three decades, four decades, right? Most of the jobs were in factories, manufacturing and things like that. The whole economy went to knowledge work back in 90s when the computers came early 2000. It’s just two or three decades. That’s it. It’s nothing compared to 200,000 years. You know, it’s like it’s nothing. We’re going to another level. We just had a good run on the knowledge work. Now AI is going to kill the knowledge work and the white workers. The white collar jobs killed the manufacturing jobs because we created streamlined manufacturing in some ways. Now AI is here to kill white collar jobs and we’re going up the abstraction ladder. Yeah. But then you know where does it take it from killing you know the white collar jobs to taking the jobs itself right? I mean when we talk about you know entrylevel jobs being you know on the hit in the upcoming timeline and I think it’s still facing a backlash for most of the grads right now. There was a time you know where we were entry level or intern and then we are where we are today because of those experiences. But then is that generation now you know just relying upon their skill how they present themselves how they are more handy with the AI but they’re not going to have this edge of an experience. So you know what are your thoughts on that? I agree you know I’ll give you my example back in 70s 80s civil engineering you know if you talk about engineering civil engineering was like hailed upon you know it’s like hey you did civil engineering after that mechanical engineering came you know and then afterwards electrical engineering came when I was going to college at that time in Pakistan during that time everybody was talking about phones telecom it’s like hey you need to go to telecom engineering when I was there it was the top engineering selection in the whole Pakistan you n being on the top and if you make it to the telecom engineering you are the top top so you know if you remember you know people know that you had to select certain things where your priorities and everything would choose telecom engineering the top minds at that time was that within four years of my degree that telecom industry went there to a saturation when I came out it was already getting saturated right so it’s like a small cycle of how things change but again we have to realize is business drives technology, not the other way around. So if you’re making money there, it’s going to happen. Nobody’s going to give, you know, we’re going to live in the realistic world. You know, a lot of uh young kids, they come out of college and I’m sorry to say, but you know, it’s like entitlement is a thing. You know, it’s like, hey, when I come out, I should have a job. Why don’t you create a job? Right? Think from that mindset. It’s like nobody’s going to give you anything on the platter in life. Life is hard. Humans have been on this survival mode for a while. Life is hard. You got to adopt. You got to get better. You got to know if somebody would have a job for you then they will have you know unfortunately you know in Pakistan when I was there maybe now I’m a little bit disconnected from the politics everybody blames their government right yeah it’s like government should have new jobs like how why don’t you create a company that would help create jobs for other people as well right? So would have to shift a little bit of mindset. You know, it’s like rather than looking for jobs, create jobs, create a product. This is hands down, you know, now we talked about a little bit of pessimism about how things are moving. Let’s talk about optimism. This is hands down the best time in human history to build products. Why building has become so easy with AI? You have your co-pilots, build it, write code, whatever. Go build things what customer wants anything. So to all the young entrepreneurs and people go start building even content creation you know is itself an act of building something your lifestyle or coaching or anything but you’re doing something which people want to see. How capitalism works is like it’s buyers and sellers right it’s like demand and supply these two things there are only two things demand and supply the more demand the supply is going to so so you have to balance that what people need you have to build you know coming to another question which is very relevant to this so sad given that you know we’ve talked at length about AI and everything is is software engineering debt I mean how do you answer this question to or how would you answer this question to a 22year-old CS this grad who’s terrified that their skills are going to become obsolete and you know software engineering is just a piece of cake. Well, um I would say that if you are like a really top level software developer, you’re still there, but soon we are going towards that. It it’s not looking that great in in that way. We need to reskill a little bit and go to the core values, right? In my opinion, mid to low level software skills, Python and everything, you know, if you are like coder, uh AI is going to do a better job. But if you think about being an architect of the application, being building products end to end, being a person who is solving the business holistically, that’s something that’s going to be way better, right? So that means that if you are someone who can leverage AI, AI agents to build products holistically, building the whole architecture together, how all the pieces come together, that’s the key. how to make it a product that is what we need in this market right now otherwise you know AI is going to do a better job in coding as well this is the first time ever in human history where computers have understood our language which is English or UDU or whatever now we can communicate with that previously what software engineering is software engineering is more about development and talking in computer language right now we understood that language now we can communicate in our language to them and they will respond. So in that way you know we would have to think about like going to the core skills critical thinking problem solving logical thinking even coding has these concepts you know it’s like decision tree now you for example make AI agents there’s a whole architecture that you have to build topologies like I’m going to put one supervisor I’m going to put three of these agents this agent is going to talk to so you are becoming an HR of these AI agents that are going to talk we’ve we’ve talked about you know how it would affect the students But you know let’s shift on the other side of you know how people with your exposure with your expertise and at your professional level you know settling with this idea of change you know that has come with the AI because now we all know that change is the only constant but recently it has been quite frequent what are your thoughts on that I think for me you know there there’s certain core principles that you have to follow the same way you know the companies operate on certain core principles for me as well lately you know since everybody’s learning and things like that I follow certain things like having a growth mindset be a lifelong learner adopt change right and this is the part where you’re asking is like we have known one core principles of human survival is like you got to constantly adapt to new things and adopt faster right you could go wrong in a lot of different things you know social media and everything the world is becoming more volatile who would have thought that uh putting your videos on YouTube and social media and sharing your life is going to make you rich. But people who took that opportunity are there. So there’s no definitive answers like hey this is the guide that to live a productive life or whatever. Everything’s at stake right now. Human knowledge has has decimated everywhere in the world. Everybody knows whoever’s going to take an opportunity they’re going to do it. And that’s a great example of like they’re talking about open AI and I’ll give you like one dramatic uh story to it. It’s like, you know, one night these two guys, Elon Musk and the CEO of Google, Larry Page, they were friends, you know, they were very close friends. Elon used to crash at his place in Palo Alto, like in Silicon Valley. And this is a story told by Elon Musk, although he’s been a little bit all over the place lately, but you know, that was before. And they used to have these existential things that, hey, you know what intelligence form we’re going to work in. At that time Google was deep Google is everything came out of Google. All this AI everything’s came out of Google. Google was a forefrontier back and deep into research 2005. I won’t go way deep into that but it’s just like Google acquiring deep mind deep mind and there was another guy you know who a co-founder Londonbased startup they started solving these problems and the first was figuring out AlphaGo. AlphaGo was the first model that they used to beat human. Lee Sidal uh from South Korea, you know, one of the top Go players in South Korea. They beat Alpha Gobi. And that just changed everything in the world. These people of like the top uh mindset. They’re like, “All right, this is possible. This seems possible.” That was the first point in in the history of us. We thought that all right AI is going to be something now. And then Google heavily invested in AI. They acquired deep mind start doing all that Google brain they had these two divisions and still the head of Google AI uh Demis actually he was the founder of deep mind you know working both with Mustafa Salomon to build all these he’s still there at Google they have Google brain and deep mind that was embedded into that and they were working on the these AI challenges and everything and they took this step you know I was like hey let’s let’s figure out the next step and one night Elon Musk and Larry Pageige were having this argument that and this is a real story of course you can go and check it the whole it’s pretty interesting since then they haven’t spoken with each other and Larry Page said that AI is going to be the future human intelligence is going to go away and Elon according to him he’s a humanist he said I want humans to succeed in some ways his proclaimed stories and Larry Page told him a specist and you can check you know he he he rant a lot about on TW Twitter about this that how dare he call me I am a specist yes I care about our species you know I care about humanity and everything and he’s like no AI is going to take over and after that day that night they had a huge argument and a fallout and they have not spoken they are not friends with each other since that day Elon went back and started this new company called open AI talked to Sam Altman Alia and Greg Brockman they put this team together to solve that was back in 2015. Call it an ego, you call it whatever. That’s that was the starting point of AI, generative AI recently 2005. They were just burning cash. He put 15 million of his money there in open AI Elon Musk and then you know raised some money. They had GPT1 didn’t do well. GPT2 didn’t do well. GPT3 was the point 2019 where Sam Alman is a is a big Y combinator startup guy in Silicon Valley. So he knows how to build products and everything. After four to five years on this journey of having no idea what they’re doing. They wanted to solve this AGI problem, artificial journal intelligence problem, right? Who would have thought in their mind it’s like I’m going to build like something, you know? They took chance on that, right? They were going to burn billions of dollars there but it was existential to them. 2019 GPD3 came 3.5 came with Chad GPT and rest is history you guys know but that was the the starting point of how everything started of these couple of knuckleheads uh going in and uh starting something which nobody believed in. So when you talk about your life, you know, your viewers are like, “Hey, things are working in some fashion. There’s some rules to the game. There’s there’s no rules to this game. There’s no rules to this life game.” You’re going to listen to a lot of like these gurus where they teach, oh, you should live life that you should live like this. No, nowadays, you know, if you make a video and you are like uh it goes viral, you are rich. It’s like who would have thought. So in that way there’s no rules to life right you’d have to take chances things go big you go big and you create a history and you know people say that you know it’s like history’s been written by victors that is true it’s like we have not seen the people who struggled and failed we only think about Alexander the great we’re going to talk about Elon Musk who who made it and the people who who have not made it follow the same recipe but they’re not part of the history and the story you know Assad everyone is talking about AI regulation from a global perspective including China, the US in IP wars, right? But when you’re working on enterprise implementations, what does real governance look like on the ground? I think governance is a huge topic which everybody’s talking about. You put regulations on thing, there’s no innovation that happens. We have seen three models. I I’ll talk about three models. The American mindset which is very entrepreneurial. We have European mindset way heavy on regulations. And you talk about Chinese model which is in my point of view authoritarian and copying. Now things are changing you know they’re becoming more entrepreneurial uh competing but it is a lot of copying that China has done you know they build they build everything based off American companies you know from IPs to everything like everything. So these three different mindsets they are there in the world of AI regulation is really important for masses for population but for innovation it’s not. So if you I’m I’m not sure if your viewers would remember you know it’s like when OpenAI came out with this model a lot of people 50 people signed this hey stop including Elon Musk because he wanted to catch up you know with his ex AAI they signed this piece of regulation hey halt any AI development for 6 months right and that was part of the regulations like hey so that I mean in business world you know it’s like he wanted to catch up on X AI or whatever his ulterior motives were but in my point of view and the world’s point of view, it’s always profits over responsibilities. The bad actors in some ways are going to win in in some ways. You can’t stop people from doing something. But you can have a rule of engagement or an understanding with each other that hey let’s let’s agree on that it’s going to impact the whole world us and you be responsible. We’ve taken a great example from nuclear technology right I mean 1945 we we had nuclear bombs since then probably like 10 plus countries have this after Hiroshima and Nagasaki we did not had any nuclear war and you saw recently you know India and Pakistan had a fight they could have both blown each other up like who who would have suffered you know it’s like the population so in that way there could be an understanding but you cannot halt the innovation that’s going to go. So in that way the governance is going to be there for the betterment of all the society but both all these three mindsets are going to go and innovate and compete you know in some ways it’s my personal opinion I reside more with that culture of freedom and free world to entrepreneur to a little bit of capitalistic mindset versus authoritarian mindset. So I can share a little bit of my views on like the west versus China is becoming emerging of huge power almost at the same level of AI development where China is Nvidia chips have been banned to be used in China although deepseek model has been trained on HD 100 which is the Nvidia chips now that’s banned there’s an escalation there’s lot of things going on including the the tariff war but Huawei which is one of the Chinese company, you know, have almost the same level of hardware that they have and their software companies like Alibaba, Tencent, you know, all these companies are doing pretty well. They’re almost at the same level where Americans is. In my point of view, we’re going to win the war like US is going to win the war this AI because we are always a step ahead innovation. It could be new models that that are becoming. I’m going to just uh highlight one thing that brings up a huge concern about what reality we want to tell the people of that country or entity. It’s like imagine AI models coming from China that would have a different version of reality. Like we know deep sea came out and if you would ask tenement square incident it would not answer it was redacted right. So in an authoritarian world you would have these problems where growth and information is not going to be the key. It’s about masses controlling the masses and live population. And in the free world it’s going to be different. So we’ll see how it’s going to unfold. who’s going to win what. We have seen throughout our human history that certain ideas took over, certain ideas died down, certain nations took over, certain nations are not there anymore. You know, the fallen of the great Roman Empire. Everybody talks about huge one of the greatest empires ever. Now there’s nothing. You see the coliseum, there’s like nothing. It’s some of the things that goes in nations and humans that make or break them. We’ll see if we’ll have a free world led by west or we’ll have a more authoritarian version of China that’s going to take over and you know see how things going to unfold. But one thing is for sure you cannot stop innovation. Both countries, both regions, both powers are going to still keep developing and that things going to continue. No, I think that was that was a great answer and pretty well summed up. Sad McKenzie predicts trillions in productivity from Gen I but you’ve seen projects up close. What’s a productivity myth you think we need to bust about AI agents in the enterprise? So shifting gears you know back to like generative AI what it impacts how it’s empowers. It’s very hard to quantify that you know you have a you create one product like let’s say it’s very quantifiable how it’s going to help. Unfortunately with generative AI there’s no correlation directly it hits productivity but how to measure productivity number of hours labor cost right productivity hits three things right labor cost time and it’s pretty hard to quantify everything you know a lot of projects you would have to go after the project to understand how did it impact so generative AI falls on that category and even when we talk to our customers it’s the same thing it’s like okay how it’s going to help me my business then we have to tell them it’s going to cut your resources or labor you’re going to have to fire 200 people but you’re going to save cost on that and that’s what this generative AI is you know when we quantify it’s like hey you’re going to save 100,000 man hours and one person is going to do five person’s job so you can fire so in that way things are going to change so when I say that you know like generative AI is going to hit these three things which is labor hours and cost, right? It all goes to when you evaluate a company, you know, the company or a business works. There are three things or two things, you know, at the core how to make a business profitable. It’s either you increase revenue, either you decrease cost because your net profits is how what you take back home. So for your viewers, you know, if you if you want to know how businesses work, there are three key things, you know, it’s revenue, COGS, we call it cost of goods sold. You know it’s like the cost. But the the third thing is the net profits and there’s like lot taxes or whatever but profits is revenue minus cost is your profit. Any company wants to increase the profits right how you do that either increase the revenue and cost is going to remain same or you decrease the cost. Unfortunately generative AI AI agents and everything is going to hit that cost thing. What is cost? Humans are cost. We have to pay people to work right. What is cost? Having all these fancy things you know that you would have to do and number of hours and all that stuff is a cost. How you cut cost? Build these AI agents leverage generative AI which is very strategic in a company and you decrease cost. Your revenues even they remain same you’ll have more profit. So once you decrease cost it’s going to go in the pockets of the shareholders. Right? So I’ll say one thing and everybody knows that with this generative AI and AI agents and all this AI revolution half of the society are going to get benefit the most rich going to become more richer and you’ve seen that trend happening it’s been happening ages it has been like that yeah and that talk you know that brings up you know it’s like if one person like Elon Musk you know having $400 billion can you imagine one person having $400 billion you know it’s like more than the the total GDP of uh Pakistan I guess you know I’m not sure but it’s one person you know owning countries how and why we come to that level do we need regulations and you want to create a a more capitalistic society where one person have so all these things are up in the air there’s no one model to have it but that’s the world we living in but if you want to be at the top go leverage these technologies and everything so that you can leverage more but the thing is going to be the investors who’s going to invest in that the people are going to use it, the rich people are going to become more richer. The society is going to become more dismantled in a way that who are going to become more poor, who are middle class, they’re going to go back and then richer and and same with nations as well. You talk about US versus Pakistan, there was a huge chunk of Pakistani market which was IT engineers and everything, right? Do we need them now? Not sure. We need high level engineers but you know it’s like mediocre or low-level coders. No. And it’s been shifting right now. You don’t see that effects right now because everybody’s building. We we are building. Everybody’s building. Once this is going to go down next year, next to next year things going to stabilize. Imagine the IT industry in Pakistan. So I don’t want to sound pessimistic but I’m trying to be more realistic in what’s coming in Pakistan. What industry especially IT industry a lot of people are in that particular field. And I tell when I went in Pakistan and gave uh these talks I was talking about the same thing is like the things that are coming in the future are going to impact you directly. How to prepare for it you know that’s you have to think about yeah that that is something that will differentiate from in between and from the outside that how you are going to cope with that change and how prepared you are for that. But then you know you giving keeping up with with the tech and the innovation and the AI agents we talk candidly about it. If AI is the devil, we are that devil’s advocate these days. Yeah. So, uh well sad uh we will now move towards the end of our conversation. I have two more questions for you. After working across continents, industries and AI maturity levels, what’s one hard lesson you’ve learned that you wish someone had told you earlier? I think one lesson is like it goes to the core. It’s like embrace change as much as you can. In this productivity era, hyper productivity era of social media and everything our brains are cooked right two things focus and time these two things are the ultimate human currencies in the world we’re going to live in attention which is kind of focus you know attention is all you need attention is the key even the paper of these transformer model actually GPD35 is based on that the name of that research paper that started everything this AI I think the name of that paper was attention is all you need cuz there were three things in that thing is like one self-arning this learning but attention it’s like focusing on the right thing focusing on on the key things of understanding the context that changed everything in genative AI and the GPD came out of that so check it out the paper that changed everything was a paper by Google brain six people by Google brain that provided The basis of GPT which is the model that came out first was attention is all you need and in this world everybody’s after your attention you need to prioritize like even big companies social media how they make money viewership attent focus and you give your focus and time that comes to attention you’re giving attention they can sell you products and you are the product you are the product in the society in social media and every you know you are the product that they want your attention everybody wants your attention So prioritize your attention. Have focus and time on things that are really important to you in the long term. Those are the things that you need to focus on. I tell myself as well, it’s like my time and focus are the two most important things in my life. Prioritize on things that are really important because we all we’ve got is these two human currencies. You know, a short life, focus on the right things. So, and then embracing change, you know, all the other growth mindset and be on the, you know, on the forefront of things and all of that, but focus and attention. I mean, focus and time is this formula for me. You know, it’s like focus plus time, attention is a key human currency. But then was it always easier for you to embrace change, you know, to be who you are today? Was it always easier or did it come the hard way with lot of lessons? I think it’s come a long way. Like if I’ll give my example, I’m not like super successful person. I’m just like a mediocre who who has gone through certain understandings, you know. But I would say you’re very humble about yourself. I appreciate that. But you know, it’s like everything’s relative, you know, is relative. It’s I would say myself like mildly successful in certain things. But what changed for me especially coming from a Pakistani mindset and the world question everything around you. I was a big I would I would ask question. I was curious. us like why you’re doing this why you’re doing that like from a very childhood and I wanted to be different in some ways to like hey if somebody’s doing that you you should not just go and follow be like a sheep you know question things be curious you know it’s like curiosity is one of the key things that you can put in yourself like be curious and then you’ll find you’ll go through your own journey everybody has their own journey that’s the unique thing about everyone they will have their own journey but be curious ask question why and then find answers and go to that path. Sometimes that’s going to be a dead end, but you’ll learn. There’s no losing. It’s either learning or winning. So, what lessons I’ve learned, you know, along the way is I think I am still following the way I’m following the the world. I’m very adaptable to change. Very adaptable to change in a way that I want to learn new cultures, you know, I want to talk to new people. And that’s maybe personality. I’m a little bit extrovert. I go do different things you know I run meet people things like that so it’s a lot of personality but it’s again your own journey you know it’s like whatever you want to follow whatever you think is best for you you follow but be curious one of the key things you know in my talks my last the the ending of that uh talk was on in one slide and you know it’s a I go on a run on this uh trail right right across you know is a ladybird trail and I go on a run every Saturday and there is um a fly over you know and there’s a thing written there live a great story and every time I go and I’m like in my own zone of like hey let’s just you know do five miles and everything and that’s what I see and that’s what reminds me life is short we’re here we are fireflies you know make the most out of it live a great story do not live a mundane life needs to be cherished don’t be depressed or whatever just keep going keep working on something. There’s so much to learn, to know, to experience and everything. Be curious about everyone, everything. And live a great story. Do not live a mundane life. And I feel like that has embedded into my mindset in a way that I want to experience new things and go. Not sure if I’m doing a great job at that, but at least I’m trying at certain level. Yeah. So, is is there anything else that excites you more other than your, you know, job or being in being an engineer or being at Microsoft? What if we take this out of the equation? Who would Sad be then? I’m really excited to be a part in this particular time and era where we’re discovering AI and I’m really excited about what’s going to be next. Although we are very small part of it, I’m more existential how I see life and that’s what I said. It’s like we are just part of the equation. You know, in in my belief as well, we we might be living in a simulation as well. I don’t believe we’re living in a based reality. Um, a lot of people do agree with that too, but you would never know, you know, it’s like where we are and everything’s going, you know, you would live, you’d be living in a virtual world and sometimes you go in your dreams to a different world. Uh, that’s all debatable. You know, it’s like we don’t know a lot of questions. Don’t know our place in universe, who we are, we’ve come from. But for me what things started was this understanding your existence and understanding the meaning of life which is I’m not going to go deep because people were like oh there’s a meaning of life a lot of people who are like working in AI researchers or whatever even you see Elon Musk gro XAI their theme is understanding the nature of universe right it’s grandio it’s like what is our place in the universe you know so I’m excited about going to Mars humanity going to Mars I’m I’m excited about AI taking off. I’m excited about maybe my next uh summer six-pack challenge which I have not started but you know I’m excited about learning new things and meeting new people meeting you and being part of this uh is a new thing and you know seeing the feedback I’m trying to start a series on on uh LinkedIn about building enterprise agentic AI as well I’m excited about that but it’s a journey it’s a journey towards something big and I don’t know and that’s drives me the curiosity how it’s going to land. I I figured that you’re an existentialist. We’ll end this conversation on this one question. So that what is the meaning of life to you have to say it in like a sentence or two. How how would you answer that? I don’t believe there’s an inherent meaning to life. You create your own meaning and I’m creating my own meaning. What my meaning is is my journey, my experiences, the people I’m surrounded with, the place I live and the impact that I’m going to have on people. And if if I could have a huge impact on people, making other people’s life great, that would be a huge success. The meaning to my life is understanding the nature of universe, as you said, our place in universe. What’s next in humanity? Are we going to merge with machines to have a different kind of species? Are we going to be like this? But these are all the things that excites me and my meaning to life is still unfolding in that way that it’s a journey I’m trying to create. Yeah. Yeah. Well, well, sad in this in this simulation of the world, I think I’m absolutely, you know, you grateful to have you with me over here to share these invaluable insights with our listeners and I would just want to thank you, you know, for everything that you shared uh you taking our time for this and thank you so much and then I believe that you’re going to do great out there. Yeah, I appreciate that. Thank you so much. [Music]