Transcript: Learning Generative AI, Product Management & Upskilling your Career with Nataraj Sindam

Host: If you give five people a task of writing a blog post, they'll all use probably chat GPT, but one of them will still be better than other four.

So there is still that human element which sort of a taste making ability that changes when one of the person has the ability to understand what actually makes it good post.

Welcome to the learn all podcast with your hosts, learn CEO Damon Lemby and SVP of product Darren Bridget. Our mission is to explore how continuous growth shapes the path to success.

Our guests are high achievers who've learned there is always something new to discover. They'll share their stories, strategies, and insights with you.

So whether you're a budding entrepreneur, a seasoned executive, an aspiring athlete, or simply someone who believes in the lifelong pursuit of knowledge, this podcast is for you.

Join us now for the latest episode as we continue the learn it all's journey through a world that never stops teaching. Welcome to another episode of the Learn it all podcast.

Three quarters of all knowledge workers are now using generative AI and that number has nearly doubled in the past six months. We are Damon Lemby of Lernet and Courtney Richie, Director of learning experience.

Today we are joined by our guest Nataraj Sidham. Nataraj is a senior product manager at Microsoft. He's host of the startup podcast.

Today he had a me on as a guest and he's also an angel investor and author of 100 days of AI and educational series on AI. Nataraj, good to see you buddy. Nataraj: Thanks for having me on the show. Host: It's great to have you here.

Can uh can we start off? Can you give us a little background on on your journey? Nataraj: Yeah, sure. Um so I'm Natraj. I grew up in South India um in a small town.

Uh studied uh math and computer science in India and then I got hired by this company called Epic Systems, which is based in Madison, Wisconsin. Um that's how I got to move to the US.

So I sort of moved from a tropical 25°C weather to a minus 10°C weather in, you know, in November 2013. Um, so directly went from, you know, very normal climate to a snowfield uh Wisconsin. Uh so that was an interesting change.

Um, you know, worked at Epic uh developed healthcare products. So Epic is one of the largest electronic healthcare record company. So worked there as a software engineer and product developer. Um got to do a lot of interesting things.

Um one of the things Epic is known for is they always send their employees uh to observe their customers. So got to visit a lot of uh interesting places that I would otherwise not visit in my life.

Like, you know, I was in um hospitals observing how doctors used, you know, software. Um, how healthcare professionals used software.

So from front desk to inpatient therapy to uh physicians, surgeons, um so it was an interesting observation of how, you know, technology can impact sort of one of the most fundamental aspects of humanity which is healthcare.

Um, so we did a lot of interesting things um in that space, you know, developed from products from scheduling, payments um to, you know, in interactions with patients and documenting that uh in software.

Um, and tried to do that in a sort of a customer friendly way. Um, so after, you know, working at Epic, then I moved to Microsoft. Uh right now, I work at uh in a product called Azure Files.

Uh, you know, where we store a lot of unstructured data in Azure Files and unstructured data is where pretty much, you know, all the AI buzz is happening right now.

Uh before that, I worked in a couple of products, you know, including Microsoft 365, Azure stack, which is sort of like Azure in a box where, you know, customers want Azure on on premises.

So Azure stack is sort of like a hardware plus software product that gets delivered to customers and they can use in parallel with Azure Cloud. Um so after doing a bunch of things now I'm currently uh a product manager in Azure Files.

Before that I was also a software developer within Microsoft. Uh so yeah, that that's a little bit of background about me. Host: I I love your diverse background and story.

One of the things that I I first noticed when I went to your uh website, uh was the 100 days of AI. Can you share with us a little bit what inspired you to author that and what key insights have you come up with through that journey?

Nataraj: Yeah, I mean, um you know, like everyone when ChatGPT sort of exploded, um, you know, I was also sort of blown away with the technology um and I have done and if you are a software engineer by profession, uh like pretty much all software engineers, you know, in Big Tech do machine learning.

Um and there's this very famous course by Andrew Ng who's now currently on the board of Amazon and Andrew Ng is sort of like the iconic figure in the field of machine learning especially as an educationalist. He founded deeplearning.ai.

And he has this amazing courses, very long form in depth courses to teach machine learning. So you see almost all software engineers who want to pursue or exposed to machine learning take these course in deep learning.

So I took that course about five maybe five years back when I first joined Microsoft. Um and understood like what is the cutting edge of Microsoft sorry machine learning is. Uh but then I didn't really you know focus a lot on it.

But then after ChatGPT, uh I I realized that this is you know truly important and truly foundational. Um, so I wanted to sort of take time to spend more time and to understand what is happening at a cutting edge level.

Uh so my background in machine learning and mathematics obviously helped.

Um so I said okay in 2024 I'll spend 100 days just two three hours a day spend on learning and you know going deep into LLMs, large language models, what sort of applications can we build?

What what is the impact going to be? and just do experiments and play around and come up with ideas and also share those in the form of blog posts.

So I call it 100 days of AI and just started documenting those and it's mostly an exercise of curiosity plus you know just learning and because, you know, I think when when engineers were doing machine learning or when I was doing machine learning five years back, I couldn't squint and see the future beyond what was there.

Um you know, natural language processing is, you know, what our effort it was referred to which today is referred to as, you know, large language models.

Um but we couldn't see how it would improve applications beyond what we were already doing in Google search or beyond what um, you know, we are doing recommendation systems for example in Netflix was pioneering recommendation systems which got eventually into Amazon and all the other websites where you can see, you know, objects being recommended based on your behavior and different data points.

So, when you when you squint really hard, you couldn't see a future beyond that, at least I couldn't.

Um, but with large language models and if you squint hard enough, you can see you know, the future around the corner just with the current technology that we have without any new innovations. And obviously we'll have some more new innovations.

So that sort of led to, you know, do this exercise and uh it definitely sort of reinforce the fact that this is a very foundational transformative technology and it's a horizontal technology. So it will affect all industries and sectors.

It's hard really to think of a sector that it will not affect. Um, so that that's some of the background where, you know, why I started 100 days of AI and why I was documenting it.

Host: Let me ask you just uh take a step back uh for us novices out there. Can you explain a little bit uh what large language models are? Nataraj: Um, sure. So what is essentially happening?

I mean I I'll explain the process and maybe that sort of explains what they actually are. Um it's actually a probabilistic equation what is happening behind the scenes.

So for example whenever you're talking to chat GPT you're talking to a model behind the scenes and when you're saying, you know, uh they're actually a predicting machines. So they're not actually um intelligent in a way that we are intelligent.

Um they're actually predicting machines. So for example you ask them uh if you give a series of text let's say what is your first and you give a blank.

What they're trying to do is they're trying to predict okay what is the next token or word that is of highest probability?

And the way the way the model actually tells you that or returns you something or completes that sentence or words is basically behind the scenes, it is building a large probabilistic equation. Um and how does it build that equation?

That's where the training that you often hear. Uh you know, for example I give it a book of text, right? Or a paragraph of text to the model.

What it learns is if I if you take a sentence and from that sentence it learns that okay for if I if I've given the first two words I have to predict the third word. If I have given first three words I have to predict the fourth word.

Now that's the information that it is learning and creating a probabilistic equation essentially behind the scenes.

Uh so when you train it on just one paragraph, its prediction capabilities are, you know, pretty much to zero because the equation from one paragraph, how much can you learn from one paragraph? But you take the entire text on the internet.

Um you know, I think GPT3 was trained for around 45 terabytes. You can say why is internet only 45 terabytes?

You know, internet is actually in petabytes, but when you compress this pure text that they actually want to train on, it actually came about to 45 terabytes and they train on this 45 terabytes of pure text, then it starts to get good at predicting or completing a text.

That's what essentially a large language model does. Courtney: Um I have I'm obviously selfishly curious about this concept of of knowledge work and how AI is going to be reshaping that.

You know, Damon mentioned at the beginning that over three quarters of the world are, you know, using AI now. So I'm curious from where you sit and all the knowledge that you're trying to glean on this technology.

How do you see AI reshaping knowledge work and specifically the product management space I know you specialize in. Um what are the impacts going to be and how is it going to shift the entire model of knowledge work?

Nataraj: Yeah, I mean uh so any big wave, right? when when there is any technology wave, I think there are two ways to look at it. One is it could become a threat to an existing sort of status quo.

And then you could look at it and say okay this is a great opportunity. So, I'm sort of looking at it as a second part like it's a great opportunity to rethink a lot of things.

Um, and whenever we've seen that a new technology has come, we've shifted our sort of abstraction level at which we humans work.

Like there's this great example that, you know, 100 years back humans were hired to sit in a room and do mathematical calculations. So, you know, people were actually hired just to do additions and substra subtraction uh subtractions in a room.

From there we went to went on to create, you know, abacus and then you know calculators and then computers. We're still doing math, but we're doing it at a different abstraction level.

So I think that's the way to best think about, you know, knowledge work as well.

So, yeah, we were still, you know, 10 years back, 20 years back, we were still writing blog posts. we'll still be writing blog posts, but it will be more computer assisted blog posts.

You know, uh if you if you give, you know, five people a task of writing a blog post, um, they'll all come up they'll all use probably chat GPT, but one of them will still be better than other four.

So there is still that human element um which sort of a taste making ability that changes when, you know, one of the person has the ability to understand what actually makes a good blog post and what actually because there's not just writing the structure, writing the flow or conclusion concluding it, right?

It's about you know what is the point? What is the storytelling? Can you say something differently? Um, so all these things you have to consider when you're writing a blog post.

So, I think we'll still do the same thing, but we'll do it at a different abstraction level. So you can apply sort of this to at different take any knowledge work, right?

If you are an accountant, you would still have to go through that training of understanding, you know, what a balance sheet is, you know, how the technical terms work, what is the legals around accounting, but maybe you'll not spend more time making tables and doing calculations, but maybe just verifying it, right?

Uh and some AI does a lot of work that you do and then you become the verifier.

Um, like even in case of lawyers and I was just looking at an application the other day, you know, immigration is such a complex and costly affair because lawyers are limited and lawyer's time is very costly.

But there are companies now which are automating a lot of work that, you know, otherwise required paralegals or lawyers to look at.

So there'll still you'll still need a lawyer's expert opinion, but you'll only need one hour instead of 10 hours of lawyer's time. Um, so it is more like you know the abstraction level at which we will work changes specifically for knowledge work.

Um, for I think product managers, I think it's a great opportunity to make careers. I think um, generator is sort of like the best way I can describe it as like intelligence as a service.

So, you have now this, you know, sort of API in the form of API you can actually get intelligence. Whether it's OpenAI, Anthropic or Google Gemini or Microsoft Phi, you're basically getting from an API some intelligence back.

So you can use that to create things that were otherwise not possible. Uh and we are already seeing that, right? You know, Microsoft launched this application called Recall where you can search it's sort of like searching from your memory.

It's not because you're browsing, you're context switching, everyone has this 100 apps open at any given of uh given time. But no one remembers, you know, what I've done there. So you're wasting a lot of time um in context switching.

So I think a lot of these second order problems are going to be solved because of this new intelligence as a service. Um, and even for product managers, I think I feel product manager's job is to provide clarity, right?

You take all this you know knowledge, all this um you know a little bit of chaos that exists right now in generative AI and extract value out of it.

You know what really matters at the end of the time is, you know, implement that generative AI in your product in a way that it's valuable to your customers.

So it starts with, you know, okay, this is all good, you know, this is all there's a little bit of hype. There's some fanciness going on, but how can you really use it to your own product? Uh and I think there's some great examples, right?

Um uh for example notion, you know, when GPT3 came they launched a lot of features, which is, you know, summarizing text, uh easily generating text because notion obviously is dealing with a lot of text. It's almost like a perfect application.

But I would say you could go beyond if you're a startup or a product manager working in a startup, you can actually rethinking how note taking should happen. Like is Word or notion really the abstraction level how we should write notes?

Or can we come up with a new form factor? I think this idea of coming about from new software form factors is something I'm really excited about and I think we'll see a lot more.

I think Recall is sort of like the first application which is like newly we are rethinking. Till now we've only saw like okay notion embedded AI in its own app. Um we are doing customer support.

I think customer support, I think it's Klarna who did a full on customer support revamp in their organization and saved like, you know, 10 million tens of millions of dollars for their company.

But I think I'm still waiting to see a more you know first principle based new products that will really you know um change a lot of sectors and industries and deliver a lot more customer value. I think we are yet to see that.

I think we'll see in the next couple of years because right now the more hype is at a foundational level layer level and not application layer level.

In a way you can say perplexity is a little bit like that because they're focused on summarizing or the way we search is slightly being changed. I think we are going from like you know searching to answering.

Uh we went from okay we'll search and we'll get 10 links back, but now we go and ask and we get an answer back. So I think there's some second order shifts that are happening. Uh perplexity on a consumer level is different.

Um, there's Jasper.ai or copywriter AI. They're sort of changing how you write um, you know, for the purposes of marketing or uh for different uh purposes where you are using writing essentially.

So uh I think we are yet to see a breakout new like fully AI first app. Um but having said that AI also has been there in a lot of apps that we don't generally associated with AI.

I think Google is getting a lot of criticism right now that it's been slow and everything, but Google actually did a lot of interesting things especially if you're a pixel phone user.

Google did a lot of interesting things in AI before anyone else and the great example for that is Google Photos.

Um and Google search answers like even though perplexity is saying that they are the first to launch, the answers in Google have been launched I think in 2016. So summarizing an answer on search itself was launched pretty early.

Um and Google also has this amazing feature where you know we all hate waiting for a customer rep to come online. It will take 30, 40 minutes, right? And Google has this amazing feature called hold for me. That's also AI.

People don't associate it with AI but Google has done a lot of interesting product features that are you know sort of submerged under the product and we generally don't associate it with, you know, generative AI or AI.

Um and in fact the best of features are also like the AI features which don't look like AI features. They just solve the customer problem. I think you're on mute. Damon: Sorry about that.

Uh Nataraj, I was reading somewhere now Google search is changing, right? Is it changing more towards conversational search?

Nataraj: They're definitely adding conversation features because right now, um and they're rolling out uh they used to this used to be part of search labs, which is if anyone is interested in like upcoming features of Google they have to go and sign up for search labs uh where all the cutting edge features are given to you first.

So in that program it has been there for last one year where almost pretty much everything you uh search you get an answer for it immediately.

So that is your first summarized answer and there's an immediate chat option where you can chat with an answer just like a ChatGPT sort of experience or a perplexity like experience.

Uh my thesis is that the second order effect is going to be that we are moving from search to answers. Um and it's interesting to see how the whole ad business model is going to change.

I think there's a lot more second order effects here because web itself has been going small. You know, the incentive to create new websites has been going down. Like there's no new incentive for creators to create new websites.

Um, so if you do that the web actually becomes more centralized as the internet itself becomes more centralized, lesser number of players.

You could say there are newer websites, but there is no newer substantial websites where people are spending time. If you look at where people spend time that has been being centralized.

Even though there are newer websites but those new websites don't get traffic. Uh, so there's this big second order problem of what will happen to web and what will be happening to web based business models.

So there's also this panic among you know media businesses that how will we get traffic if everything is an answer and not a link.

Um Google claims that you know you'll get much better answers if you sort of have a domain authority in one particular domain.

But I think it's also a hint at centralization that there are only so many niches in the world like okay let's say say education or like high school education.

I think one or two websites will get all the traffic because they have the source of truth and they have they established domain authority. So anyone new to break out, the barrier of entry is going to be much much much higher.

I think the barrier of entry of like just SEO optimization has been much much higher in any new field or any existing field.

Um, so I think we are we are going to see some really interesting second order effects in web itself because of this search to answer transition. Host: Interesting. Now, let's let's move in this direction for a second.

You know, a lot of our audience are potentially looking for a career change. What advice do you have for individuals who would like to become a product manager? Where should they start?

Nataraj: Um, I mean I I'll I'll give a take that why I became a product manager um and hopefully that helps people. Um you I I was originally a software developer and I for most of my career I've been a full stack software developer.

Um, then I realized I was sort of trying to find my person market fit. You know just like people look for product market fit, I was trying to find what is the best you know person career fit that I could have.

Um, and that led to me becoming a product manager. That sort of self reflection. So everyone has to identify their own strengths.

Um, on what they're good at, what they can what they can become good at, you know, if they are learning something, there's always this concept of flow that they you know get at when they are doing something.

So I think it's it's a it depends on you know, you have to do that self evaluation for yourself.

Um, and if you are interested in business, if you're interested in tech business and if you are relatively good at tech, um, then you can I think think of becoming product manager. Um, but product manager also is a broad term.

There are a lot of different types of product managers, right? Uh for example, I I define my role as sort of like a cloud product manager where it's you know, it's much more technical.

Uh you have to understand how the cloud works, how distributed computing works. Um you have to understand you know the technical concepts more than a regular product manager. But if you're a consumer product manager, it's slightly different.

Uh if you're a growth product manager, it's again, you have to be really interested in growing products uh and how do you think about growing products whether it's B to C or B to B. Uh have you grown any products?

Have you tried growing any products and have you been successful at it and have you found enjoyable doing that? So then you can move to growth product manager.

Um, so I think becoming product manager I would say is uh is an intersection of are you good at understanding tech? Depending upon and how much tech is depends on what kind of product management role you want to get into.

And then are you interested in the business of technology because most product manager jobs in technology and are you generally interested in business because product managers are sort of like, you know, a for CEO.

Uh people say, you know, you are a CEO of a product, but you're not, but yeah, you do focus on things that a CEO does, focus also growing the business, growing the top line, bottom line, you know, profitability.

So are you really interested in business? Um and then do you have like some of the soft skills because product manager should have some of the soft skills uh especially like um do you have storytelling capacity?

Can you you know convince people about the thesis that you have, you know, on what direction the market is taking or where the customers are on why some experiment is worth doing or some experiment is not worth doing.

So I think a combination of these three things makes you a good product manager and you have to evaluate it your for yourself. Experiment a little bit, learn a little bit and see how it goes and then decide to transition to product management.

Uh for me I was lucky because I was working with product managers. I was looking at it and I sort of through self reflection realized that how uh you know deeply I was interested in just the business of technology.

I was investing outside the company. so I was already into you know tech in a deeper way that made sense for me to move as a product manager and people around me telling were telling me that hey you know product management is a better career path for you.

And I also felt that working with great engineers, great software developers, I knew that I was not top 1% of engineers. That led me to become okay, where can I be top 1%?

Uh okay engineering, I am good, I can do the job for next 20 years, but it is, you know, it's like this. I everyone is talking about Nvidia H100s, right? The chip that sort of powers all LLM training.

Um and people often say this phrase that um, you know, work hard you can do anything. Then I'll tell them okay take a CPU and make it work hard and create an LLM. It doesn't happen. It will take infinity to make it happen.

So it is like people are born with different types of CPUs and GPUs. So you have to evaluate with your CPU what really works or GPU what really works.

You can't outwork Elon because he has an H100 in his head and you have a regular Intel CPU on your head. So you cannot become an Elon just by working hard. That's a myth the world tells you.

Yes, hard work takes you a long way, but it has its limitations and not everyone can everyone does the same hard work. And let's say everyone does the same hard work, can everyone create a Tesla? No. So it's just brutal truth.

So you have to evaluate like which skills are you better at and where you can find your superpower and you know transition into that rather than like struggling really, really hard in something.

And even like one of the founders I was talking to he said, I wish someone told me this 10 years back that successful products which will become big sort of give you that signal very early on.

Not like after two years, three years, but you get that customer signal very, very early on.

So if so it's better off to sort of like when you are early in your career experiment with a lot of things so that you really find that one thing that clicks instead of like focusing a lot and grinding it out and you know saying that you know hard work will reach you everywhere.

Of course you need hard work, but I feel that a lot of people don't focus on the direction. Um, yeah that that's my long answer of how to become a product manager. Courtney: I like the comparison to Elon's CPUs. little different.

Um for for a lot of the people that that you know listen to this and and and uh we work with, you know, they're dipping their toes in. There's a lot of um trepidation about where to even begin.

And one of the questions that I have just for the foundational understanding is the competitive landscape of AI, right? There is what people have access to at their fingertips.

I think most are aware of you know ChatGPT is being kind of something they can just go get right now.

Can you um maybe kind of walk through who the key players are in the competitive landscape that is AI right now and what their big uh movements are, what they're going to be doing and what we should expect to see from them in the next six to 12 months.

Nataraj: Yeah.

So I think in terms of like the landscape I think you could define um I think there are foundational model players uh you know like the OpenAIs, um and Cohere, Anthropic, Google AI or Microsoft who are training this large language model.

So they are um and in that what we can see next coming is this um multimodal models. We are already seen GP4O which is sort of like multimodal model. Uh I think we'll also see a lot more video and multimodal models coming out from foundational models.

Um and compute and the cost of compute and the efficiency of creating a large language model is going to um be increased and sort of the cost is going to come down heavily.

Um, like if I have to predict on what's going to happen on foundational models uh direction. Um, just continuing on the landscape thing. I think you have foundational model and about that you have sort of like a tooling/orchestration companies.

Uh you have like you know Langchain, GPD Index, you know, who are sort of enabling others to build applications. Uh so that's actually a pretty thriving ecosystem right now.

And on top of that you have like these new applications, you know, like you know the Jazz AI or GitHub copilot or Runway ML or ChatGPT you can technically consider it as a application.

And then there are like sort of existing applications within that application layer you have like existing applications who are any who are using AI to make their product better. Uh so that would be like notions and the Clarnas of the world.

It could be any application really uh who are you know actively embedding AI to make the product better. Um and even like below the foundational layer you have sort of like the operations or enabling foundational model layer companies.

Uh you have there's a there's a company in Seattle called OctoML, you know, there's a interesting company called Lami um which is started by a bunch of PhDs to enable training and inference, you know, faster with AMD chips.

Uh so here these are companies which are enabling inference layer, training layer and sort of like data orchestration for pretraining before pretraining or like collecting the data. So you could put data bricks also in this sort of layer.

Uh you have companies like Mosaic ML which are acquired by data bricks. Um, and I think below all this I think we have is the cloud layer which is sort of being used in all these layers.

Uh and which is sort of like having its own new um so you could say a new wave of cloud computing because of how much compute and storage is needed for developing this entire stack.

And again before below cloud you have the silicon layer which is uh even Nvidia and um, you know, pretty much every big tech company is now doing silicon and sort of trying to do their own CPUs and GPUs.

So I think that that's broadly I would say uh you know what is the landscape look like. Um but in terms of like opportunity, I think application is the biggest opportunity. Foundational models has really high barrier.

Uh but I think there is also this notion of, you know, foundational models will eventually reach out to the same capability. Um, but what does that mean as an investor or someone who wants to create a new foundational model?

If you are already in that is that the space is still early and you should go and do that. I I wouldn't say that just because we have five players we shouldn't do that because you know remember Google was what 16th search engine.

Um in spite of all these companies like Mistral suddenly became popular because they are much more efficient in terms of like cost and performance wise because they have like a very good balance.

They're actually they're able to achieve a lot of better performance with much lesser cost and they have like innovative way of doing it. So there's a lot more innovation in foundational models before it gets commoditized.

It might be true in 20 years, but to get there a lot more innovation has to happen, a lot more unicorn has to be created in that space before it actually gets commoditized. Um yeah, I think that's how sort of like the landscape is looking right now.

Host: So one question I have for you and from listening to all this which I think is great is obviously you're curious, you're a lifelong learner and now for the last seven plus years you've been at Microsoft which has gone through a amazing transformation under Satya Nadella and you know he advocates and he talks about this in his book, uh Hit Refresh I think it's called.

Yeah, that's right. And um about advocating for a learned at all culture, which obviously we are big fans of. How have have you experienced that firsthand at Microsoft? Nataraj: Yeah, definitely.

I think when Satya came to Microsoft, I think he or when he became the CEO, I think he did two things which sort of really set the stage.

I think one is he sort of define the vision or like what Microsoft is very well, which I think this Balmer era sort of lacked, um, which is sort of to say that Microsoft empowers everyone to their and to to do their job essentially.

That's what he defined. I'm paraphrasing a little bit, but that's sort of one line sort of puts Microsoft in the right position.

I think he understood really well what Microsoft's role in the world is and that really one sentence sort of makes it clear. You don't have to go beyond like what is Microsoft.

And I think during the Steve Balmer era I think there was a little bit of you know what is Microsoft? It does so many things, but what is Microsoft? like no one really understood it.

I think that's one thing he really did uh sort of set the narrative in terms of what Microsoft will be doing. And it also flows down across, you know why they acquired LinkedIn, why they acquired GitHub.

When you look at it from that one line, it all makes sense. So I think that's one thing he did um to sort of the set the stage for everything else. And the second thing is this uh internally I think to set the culture, right?

With this you know you have to be a learn at all versus a know it all. And that's constantly reinforced within the company. It also makes people um be more empathetic towards their co-workers.

Uh because often what happens is you don't want to have a douche who's very intelligent or who's very productive, but you want someone who's enabling. who's intelligent but also enabling others.

Uh so I think a lot of this is sort of embedded into the culture of Microsoft you know last seven eight years. Um, so it's a constant reinforcement that you have to be continuously learning, be curious.

Um, and even though like these are simple things I think, you know, my understanding is all strategy is pretty simple, uh but it's hard to do. Um, so I think curiosity or learn it all yeah everyone knows, you know learning.

Yeah everyone just learn everything. But you know it's sort of like going to the gym. We go to the gym you get fit, but it's hard because it's hard to do every day. Uh so it's it's sort of the simple philosophy of learn it all and versus know it all.

So you're always sort of coming up any meeting you go to, any customer call you go to, you're coming up with an approach of learning something new and trying to make sure that you're getting to the truth of the matter rather than coming at it with an approach of, you know, I already know this.

And that makes a difference across the stack whether you are in you know trying to get uh a product spec or whether you're working with engineers or whether you're working with customers, whether you're working with other partner teams.

Um and it's also emphasized at every level. It's not just engineers or product managers. It's emphasized at management level.

It's how you're judged inside your performance uh, you know, at whatever role on how you're helping others learn or you know how you're helping others is a big part of you know how being a Microsoft employee.

So it's sort of like not just words, but they flow into like a specific action at each individual role and sort of defines on how your career uh you can't just fake it that you can't just say learn it all, learn it all, learn it all and it doesn't happen that your career will not move forward.

You have to show it, right? So, yeah, I think it's sort of set the right culture across all teams. Host: I think what you said is incredible. You know you brought two points.

And this is important for all you small business owners out there if you manage a team or if you manage a large organization is really around the vision, focusing on what your vision is and and what your purpose.

Because I see that a lot working with our customers and some of the people I mentor is that people are just kind of confused like we go in so many different directions, we try so many new things, you know, who are we?

And um, I think that that what you brought up right there is is critical and I'm sure Courtney will agree in our world, you know, our ears light up when you talk about learning and and it's true. Learning is tough.

It doesn't happen. it doesn't happen easy and

---