Transcript: How to Market a Database Product? | CMO of SingleStore Madhukar Kumar

My guest today is Madhukar Kumar, he's the chief marketing officer of Single Store. Single Store is widely known with the name MemSQL before its rebranding. It's a proprietary cloud native database. Single Store announced their Series A funding of $5 million in 2013. Since then, the company has raised $300 million plus from various investors, including Khosla Ventures, Excel, Google Ventures, Dell Kaplan, and others. Madhukar previously worked in product marketing and growth at Davre, Nutanix, Redis Labs, and Oracle, and has a background in journalism and has been in an engineering. Uh, so this episode will talk about product marketing, growth marketing, and how to market your product to a technical audience. Um, and if your customer is a developer, how, how do you reach them? We'll also talk about career in marketing, product, and a lot more interesting things. For for this episode, and for all the future episodes, please subscribe to Start Project on Spotify, Apple, YouTube, or wherever you listen to your podcasts. So with that, Madhukar, welcome to the show. Mr. Raj, good to be here and it's good to have a conversation with you. Uh, so where are you joining us today, Madhukar? I'm in uh California Bay Area. So we have an office in Sunnyvale and San Francisco, but today I'm in Mount. I think Single Store has an office in Seattle as well where I'm. Yeah, we do have an office in Seattle, London, Portugal, and India. Uh, so I think to the conversation, I think it would be a good place to start is as a chief marketing officer of Single Store, how do you think about marketing Single Store or what your general thesis of what a chief marketing officer should be doing? Yeah, I think that's a very topical question as well. Uh, I'll tell you where I started from, but I'll also tell you that things are changing as well. So where when I came in, I tried to structure my strategy under three pillars. One is branding, the second one is product led growth, and the third one is product led sales. Product led growth, as we all know, it's bottoms up. It's going to the users and getting the product in front of them as quickly as possible and providing value. Product lead sales is top down when you're actually going towards the buyers and you are telling your story, but it's a different go to market scenario. And those were the three things. So the first one for brand, we have a metric that we measure. For PLG, there's some clear KPIs that we measure. And then for product sales, it's pipeline. Now, over the period of time, that has changed. Clearly, the brand is what drives a lot of PLG and PLS. But now, of course, the PLS part is changing. What used to be earlier, you build a pipeline with marketing qualified leads. And then the marketing job is done, that is completely changed. Now it is somewhere in the middle of the funnel where marketing has to drive growth all the way to where it's some there's some revenue as well. So from today's perspective, I would say a lot more focus on PLG and PLS, and brand, of course, continues to keep happening in the background to provide that air support. I think to double click on branding, uh you know, Single Store is a database company. It's a technical product. Um, its customers are the decision makers in, in one sense on the top down level, and on the bottom level, bottom up level, it's developers who would start with experimenting with different databases and have different use cases. How do you think about branding in the context of a technical product like a database? And in specific for Single Store. Yeah, so the way I think about branding is uh it stems from some of my experience of working in other companies where I've seen a lot of branding and marketing, and this incidentally, this is what I posted in my LinkedIn today as well, is devoid of the product piece of it. And what I mean by that, it is often around, you know, uh some catchy one liners, it is around good visuals, color typography. All of that is extremely important. But when it comes to developers, they already know that what they like has to first go through their muster of whether it fits their need or not. So from that perspective, the trial, or the product lead growth, or the open source version is extremely important. Because no amount of branding could convince somebody to go and use a database for a product that a developer is building. So from that perspective, the main job of branding is to be in the peripheral vision of the users as well as the buyers. But that place itself is so noisy these days that almost everybody has the exact same tactics. They are doing the exact same thing. There a lot of them have a lot of funding and they're large companies, they can pull a Super Bowl ad and things like that. So from that perspective, what matters is how memorable your brand is. And to become memorable, there are certain things that you have to do, especially with developers, is there is a sense of humor. The developer audience has a very general strong sense of humor. The other thing I've read is, and this is true for me too, is a majority of the developers also do some sort of a creative hobby on the side, whether it's playing an instrument or something else. And they are also very product centric. They they don't necessarily really care about all the fluff around the wise, but more on what is this product and how is it going to help me? And you know, is there technical details around that? So if you combine all of that and add the psychology of memorability, the goal is then to have developers go try it out. So, the messaging has to be really crisp, it has to be direct, it has to be a pattern breaker, it has to be memorable, but at the same time, when you have a call to action, that call to action cannot be come and talk to my sales team. It has to be go try out my product. And now when you try out the product, then the second part of the marketing comes in, which is the adoption of the. Well, in terms of product led growth or, you know, what you refer to as bottom up marketing, as a database product, um, I would assume you're trying to reach developers, you know, across different channels. Um, how how easy or hard in this environment is to reach a developer through a bottom of marketing methods. I think a good example would be that, you know, especially whenever the bottom up marketing sort of scenario comes for a technical product is, famously, like Slack was installed by people in their organizations and it was so widely used that at a point that, you know, buyers said, yes, okay, we should buy Slack and make it official. Right? There are examples of these kind of stories with a lot of developer products start with, you know, one engineer starting to use their product in their team and sort of expands to the enterprise level and eventually the decision makers in the buyers sort of uh buy that product for their company. So, talk to me a little bit about where we are in the industry right now in terms of product led growth and what sort of initiatives that you at, you know, Single Store are doing. Yeah, so to to answer your question, first of all, there are over 320 ish databases today. And when you think about it, there are, you know, it's almost like a visual where there are big circles of very large companies like Oracle, um hyperscalers, and so on, right? Then there's Snowflake, Data Bricks, and then there are smaller, maybe two or three. And then the rest is all small dots. And the small dots are very specific to a specific use case. For example, a vector only database, used primarily for semantic search. Or it could be a graph based database, or it could be a postgress in the cloud, or it could be something so niche that it only serves a very small portion of the market. And over the years I've been in this market long enough that it has become heavier and heavier towards the top, the bigger circles. And the smaller ones keeps coming in and churning and moving away. Some gets acquired, for example, Roxset got acquired and so on. So, from from that perspective, if you think about what is it that Single Store does, which is differentiated, and is there any validation for it? I would say there are a couple of different things. One, it was known as MemSQL as you said, when it was founded. And then over a period of time, it did certain things, like it went from in memory to memory first. And it did transactional analytics, both, and so on, that uh was patented, so it made it differentiated. But it has survived 11 years to have about 400 different customers, majority in Fortune 100 companies, and I would say it is growing consistently. So it's said there is definitely product market fit. So now the question is, how do you reach the developers? And the way I look at it is if you, if you want to put the databases into primary categories, there's the transactional databases, MySQL, Postgres, and then uh MongoDB. And then uh and of course SQL light is now also catching up quite a bit. And then on the analytics side, which is columnar databases, you'll get Snowflake, you'll get Data Bricks, you get Vertica, there's a whole bunch of other pieces. Single Store is one of those that fits right in the middle. It can do both. And the reality is it can do petabyte scale at a few millisecond response time. And we could get into much more detail there if you want. So now the question really becomes if I have to go and make sure that the developers use us, they first need to know about us. And since we are not open source, as a developer, this is true for me too, before I joined Single Store, when I'm building a full stack application, what do I do? I do what everybody else does. So I would go and Google it or, or because I've used this in the past, I know there's a Postgress installer, which installs locally. Or maybe a MySQL, which will come with a lamp stack. I install it. I have a bunch of code that I've written in the past in my SQL, use that uh as a starting point and then go work on it. So there's that capturing that. And then there is the enterprise market that is so well known to, you know, the large players, that they don't know of Single Store. So that's a true challenge, and that challenge doesn't exist, there's no doubt to that. But what we've realized is if you want to really go after the developers, you cannot market to them, or you cannot even sell to them. But you have to be in the same communities and the watering holes that they typically engage in. That would be Twitter, that would be YouTube, that would be Stack Overflow, which is questionable now, but that would be subreddits, that would be Quora. So now if they are going and searching for problems related to what they're trying to solve, and if Single Store is now available, it is now up to their volition to go try it out. So that is kind of the overall I would say go to market strategy of bottoms up. But it has to be attached to a very strong product. I cannot just go weave there and then overtly be trying to sell Single Store when the product is not good. So there has to be an absolute 100% faith in the product, and that we do. That is the conviction that I have. I use it for my day-to-day experimentations and applications. But that's also a challenge for us to go out because now most people are used to open source like MySQL, Postgres, SQL light, and so on. So, um talk to me a little bit about product led sales or top down sales motions. And you have a lot of experience working with, you know, other enterprise companies like Oracle and Nutanix, which sort of fall into that, also have these enterprise sales motions attached to them. And I've seen closely how that sales endeavor works, many of selling large enterprise customers. What is the value proposition look like when you're trying to sell an enterprise decision maker or buyer? And how do you basically, you know, position or differentiate yourself from all these players, right? There's so many especially when you have hyperscalers with large sales teams vying for the same customer. Yeah, and I think that's where I get very excited because when it comes to value proposition, a lot of that comes from our existing customers. So it so happens that if I put one or two of our customers with some prospects, I don't even have to talk. They will go and tell the others why they switched over or why they're using Single Store or, you know, what are the things that they're doing with AI with Single Stores that has made it possible for them. So from from that enterprise perspective, again, the idea is yes, there has to be brand awareness. So you need to be in the peripheral vision of the buyers themselves. But at the same time, you also need to have a strong validation that others have used it. So, you know, what I talk about top down is for buyers, it is mostly people do what they see other people do. This is an example that I'll tell you, right? Like uh we live here in the Bay Area and there's for years and years, somebody would come to our door and say, would you like to find out about putting solar on you? And you know, nobody, I never would engage with them and walk away or close the door. But one day I go out and I see four of my neighbors have installed solar. Next thing I do is I go out and ask for this vendor to say, hey, or I go ask one of my neighbors, which one did you use? And then I'm the one having a conversation. So in the world of marketing that's more of an inbound, and the inbound is driven by some sort of a network effect. So, again in that buyer market, you have to or we think about it two ways. First, if the developers in that account have used that product, they like it, then they're the ones who's going to influence the overall buy when it goes to enterprise sales. And then the other piece is you need to figure out your exact customer profile by doing a look alike of what we have, and then using things like referral or using things like account based marketing to be in their view when they're about to make a decision. And when they are, that's when the rest of the gear comes in and it, you know, kicks in. So that's how I would categorize both the product sales as well as you know, what is zero touch or near zero cost acquisition PLG. So in that, you also have to involve humans, which is where the account executives and the sales team comes in. That's involves a CRM, that involves a pipeline, that involves a process and that's a whole different ball game. So it's almost like you're running two different companies. One company is meant for developers or the users, and the other one is for buyers. And there is a seamless overlap somewhere which is where you need to, you know, really focus on the product adoption and making sure that it's actually giving value to your users. So, you moved, I think you changed your career multiple times in terms of what you're working on and eventually came into marketing. Talk to me a little bit about those transition points and in your career from, you know, being a journalist uh to an engineer, to a product developer to a product marketer now. Talk to me a little bit about those pivotal points and you know, what drove those, you know, each pivotal point. Yeah, I like to say that I always had a attention deficit disorder. I grew up in India in a small town. Everybody that I grew up with either went to on to become a doctor or they went to IIT and I was not really good at taking exams or, you know, going in a competitive academics. So what I ended up was doing was I tried a bunch of things that I liked and writing was one of it. And so that what got me into journalism. I love the whole idea of building something from scratch, based on an idea that you have an abstract that you have in your mind that you could then give shape to. When the internet came about, the medium changed and it was a similar thing with making web pages. You have content, you have some ideas, but now you also have to think about the visuals and the web application kind of is an extension of that. So I went very, I would say organically from being a writer to then going into a developer for web applications. And then if you tie the strings or threads, then you realize the next step is to come up with the ideas of the product and building it out and that's where product management comes in. And marketing in many ways is just putting it all together because as a marketer you're doing five or 10 different things. You're thinking about analytics, you're thinking about product user experience, you're thinking about audience mapping. And you know, it's it's it's like running your own mini little ADHD camp where you're doing 20 different things. But for me it's not as if I went and said, well I want to become a CMO. I was building stuff. I really enjoyed it and then somebody would come along and say, you're good at this. Have you tried product management? You should go do this. And somebody would take a risk and give me that opportunity and I would like immediately say yes to any offers or any opportunities I would get. And then I would just jump into learn because it's something new, again with my attention deficit. I would learn, have to go figure it out. And then somebody else came in and said, well, now go do marketing because you did a great job with the roadmap. Now go take it to the market as well. So that's kind of a more of condensed way of saying that it has been uh, it has been a journey by itself. And a lot of that came in because a lot of people took risks. So today, a lot of things are changing, you know, because of AI and what it each, I think knowledge work, every role is sort of changing like what it means to be a great developer is slightly changing, right? If you're great at using cursor or GitHub copilot then, you know, from you can step from an average developer to a 5x, 3x developer quite easily. Uh or at least you're about to. And I think in marketing also, they're seeing a lot more new tooling which will basically remove a lot of hassle and sort of increase marketer's productivity in general. But my question to you is, if somebody is, you know, in product marketing or trying to get into product marketing, what are the one or two skills that you look for or you know, within your team or that you've observed that make them great at product marketing. Yeah, I've been thinking about this quite a lot because this is true even for development or coding as you said, right? There are a lot of people today, whether it's using cursor or Windsurf or clients or Bold do new, you can just go in without knowing anything about building an application and build a prototype or you build even a mini full stack app. But unless you have gone through the suffering of making a mistake and troubleshooting and asking people, and then getting that burned into your brain, you you never, you would never understand the nuances of building stuff. So to give you an example, I can go in and I can use cursor, I can use windsurf to build something. But when it tries to do certain things in a different way, because of the learning experience I've had, I can stop it and say, no, don't do that because if you do that, it's going to have an effect somewhere else, because I've done that mistake before. And over a period of time your AI gets better and better. But we went from basically in programming telling a computer how to do things to what to do things. And it's now kind of circular, where if you don't know how to do things, what you tell it to do could be rife with errors as well. And this is exactly true in marketing as well. I have gone through several, several trial and errors to know what works, what doesn't work, what is specific for a specific audience versus a specific kind of a product versus a channel versus a content. And yes, a computer or an AI can do multiple different permutations of it and you can do it, but unless you know how to do certain things at the basics level, you are not really uh maximizing the use of AI. So from my perspective, there are really in this world two ways to do 10x or 100x with AI. One is if you are sort of generalized and you have breadth over multiple different things then you kind of understand the landscape and so you know exactly where to connect and how to do things in a much more efficient way but also in a way that actually moves the needle. And the second portion is if you're highly specialized, extremely specialized in one specific thing. So you've spent 20 years in just performance marketing, then you can use AI to do all the repetitive stuff, but that performance marketing experience that you have now makes it 10x. So what I look for when I'm looking for a marketer is a is the product is technical, then is this person technical as well or not? That is table stakes. Can you use the technology, can you use the platforms that's available to you. And two, are you a generalist that can do multiple different things and now with AI you can do a lot of that like 80% of the level. And third, which is a huge benefit is if you're also a specialist in one specific. So, I know it's it's a little bit of like trying to find a unicorn, but those are the marketers or for that matter any employee that are really shining in today's world and in my belief will be in the near future as AI helps you do a lot more of the mundane stuff. And so what you bring in has to be highly differentiated from the other employees and the applicants. Wow, I have a uh thesis around marketing and I want to you to react to it. Um is because I'm sort of in the startup and uh venture capital world. Um I have this thesis that if for every $1 venture capital that's raised, uh 25% goes to the cloud. So Amazon AWS and Google, another 25% goes to marketing. There's Facebook, Amazon, and Google. I mean Google and Facebook mainly and then you know, a little bit to Amazon if you're a direct to consumer company. And that's my thesis around even B2B companies or like enterprise B2B companies. Based on your experience, is that still happening or is the spend on primarily dominated and you mentioned performance marketing on Facebook and Google or are you guys exploring different channels and spreading your ideas across other platforms. So, I would say um it has changed and it is still changing. There are certain companies that will go spend definitely 30% or more in sales and marketing, but at least in the last couple of years what we have seen is it's no longer growth at all cost, but it's more around growth at sustainable cost and predictability. And so with that, at least my personal experience has been that I'm not a big fan of paid campaigns. And so when you think about paid campaigns, obviously it involves a bunch of different things. In fact SEO to a large extent, a lot of what companies do could also be considered um paid campaign because I used to work for a company where they were spending some of their budget in getting back links. And now to me, I feel like that seems so wrong in so many different ways. So from from a paid perspective, what I feel now or what my experience has taught me is often times they are not as the returns that you get, whether it's in terms of leads or whatever you want to call it, is not of the same quality as you get when you get from inbound. So to answer your question, if I were to do this which I'm doing right now or even in the future, I would rather spend more on building a brand that creates the inbound versus putting more money on things like SEM or display ads, which sure it gives you the impressions, but the quality of the you know, the leads that you get are extremely poor. What are some of the brands or companies as a marketer that you really admire and think that are doing great work? Either it be in terms of like a technical product or in general a brand or you know, that's doing great marketing. Yeah, there are quite a few. Of course the OG of all of this is Apple. But if I think about and I'll tell you what I like about Apple is that the marketing is not separated from product in any way. If you look at their website, if you look at the events that they do, if you look at uh whatever they post out in videos on YouTube, to the actual product that you touch in a retail store, to making that purchase, to opening that box, to the support and any returns that you do, it seems like marketing or whatever is that team, has touched every single piece of that and they are, you know, extremely extremely detailed and very delightful, again an overused word. So there are certain technical companies that I am a big fan of that does this. Webflow is one of them, then it's not a technical company, but D Brand, they sell decals and they sell stickers for laptops. They seem to be doing extremely good exact same philosophy. They have a very funny and a good sense of humor in the emails they sent out in the social media post. But it's both pre purchase, post purchase, and even in support. So those those are some that come to my mind when it comes to like really good branding. One, I look for if it's very product centric, and two, if that product centricity is cohesive across the entire user journey, not just for branding and not just for ads, but goes all the way to the support. I think well, I mean, to your point that about Apple, and it reminds me of uh something I've been thinking about in terms of how communication has changed in the last three to four years. And I feel like Apple is slightly taking a backseat in in that terms because it's not just Apple, communication, traditional communication, you know, which has gone through multiple iterations of, you know, PR people checking it, uh multiple you you write an email that's, you know, looked by 100 people, reviewed by it and then uh you know, it's sliced and diced and to an extent that it takes a shape of which it lacks authenticity, some communication lacks directness and it comes across as too mechanical. and that's why people now, you know, are starting to realize that even though AI can write a lot, it should it will still see like AI wrote it. All right? And people are focusing that these are good for generating drafts, but not the final version, because you still need to add your own touch of what makes that, you know, what makes it tick. And I think communication in that sense has changed and like the PR first sanitized communication has taken a backseat in general on in the society. What what are your thoughts around that? Yeah, I think it goes back to the brand. It goes back to the brand identity. And what I mean by that is I've seen this being done differently by different organizations. Now, if I was Apple, I wouldn't think that Apple with such a large organization and such a large customer base would be doing personalized or very human based communication. To me it feels not scalable and it also feels kind of unnatural. But if I look at certain companies, like much more smaller companies, for example, you could tell, uh you could tell that, you know, that communication is coming from somebody who owns the product, whether it is the CEO or the chief product officer, or whoever it is. And you see it all over the place. It's not just in one single place, like Airbnb is a great example. these days replic with Amjet Massad. So you kind of get to feel that there is a human sight to that organization or that brand identity. And then if it flows through in the communication, even better. Now, I'll I'll 100% agree with you like in even in my organization, I told people not to use any kind of AI to generate content. But it's almost stupid to not use it for research. You have to go and do keyword research, you have to go build an outline, and all of that definitely use it. But what I saw early on was a lot of people were gaming the SEO system by creating workflows that would go and create hundreds of articles and publishing. Sure, your traffic will go up, but what kind of traffic? Is this the traffic you want? And is that the metric that you want for your overall revenue growth? So from that perspective, I think things has started to change. Now the last thing I would mention in this area is I'm still not 100% clear whether every email communication has to be one to one personalized. I feel that is kind of the reason why quote unquote the pre AI AI version was weaponized in marketing. It was all about personalization. You saw it in Instagram, you saw it in YouTube extremely extremely powerful, but it didn't lead to the same actions at a society level that we would have all expected. So I only hope that we don't make that same mistake with Jenny AI. So I do feel the authenticity should be there, the human angle should be there. But my personal opinion is that maybe one on one customization is going too far. A because we as humans can figure it out if it's been written by human or not. And two, I think it is it's it leads to other consequences which we're probably not ready for yet. Let's shift gears and talk a little bit about AGI. I saw this interesting video that you're breaking down your thoughts on AGI. Um, I think Sam Altman said AGI is a couple of years away. I think complexity CEO said the same thing. I have my own take, uh I'll share it in a bit, but I'd love to let you talk about, you know, what's your perspective on AGI and how how are you thinking about it? So, when we think or when I think about AGI, there are different, you know, it's like touching that elephant and blind man where everybody thinks of it maybe slightly differently or at least I do. So when people think about AGI, they think it is something that is identical to a human. And when you say identical to a human, now that becomes very subjective because what part of human are we talking about and so on. But in general, I think it's an established fact that today it is very mathematical, it is very permutation based. And AGI then at some point will be slightly more advanced where it's nuanced, and then super intelligence would be you would be able to do nuanced plus mathematical at a much higher level than humans. But if I think about, let's say, what is an intelligent human being today? An intelligent human being today is somebody that has a lot of knowledge, which could come through reading books, consuming media, or just straight up experience, brute force. And then the other piece is uh the ability to learn and then use tools that are available to them. And for us the biggest tool today is a computer, which is connected to the internet with your knowledge and the tool, primarily the computer, you can now primarily do anything you would want to get your job done. And then the last piece is user interface, where a lot of people like if I want to do stuff, I can't be doing that on my own. I need to communicate that to hundreds, thousands of people through voice or through natural interfaces to be able to explain, motivate, inspire, and get other people to do stuff. So if I break it down into all three and then I look at what we currently have. We actually have all of that. We have large language models with exceptionally large corpus embedded into them in terms of knowledge and data. And you can continue to add it with rag and Laura and West fine tuning and so on. Two Anthropic released computer used tool, which basically now gives the large language model access to the computer. So anything that you can do now it can do as well. So that would be reading and writing files, it would be going to the internet and everything else. And third with open AI realtime API, now you can engage with it your large language model without having to convert your speech into text and then embeddings. It just happens in milliseconds. So if I package this all together and I I haven't even touched all the audio, video and the images that you have, you can literally create a quote unquote agent that looks like us, talks like us, can be on a Zoom meeting, can go to the internet and do exactly what I can do. And has uh what is now called reasoning at inference time, which is the O1 model. So you're not just parroting what's in your knowledge base, but you are doing a chain of thoughts and then you are continuously evolving on it. So if you want to ask me a slightly non controversial take, I would say AGI is already here. But if you want to think of this or maybe you think of this in a way that what I want to see a personified version in a robot, which is all packaged together, then sure, I would say that's a year a couple of years out. And if you think about that personification but with like omnipresent and watches everything and all that, then sure, I would say that's a few more years out. But yeah, I I feel we have all the ingredients and it is here and it's actually a good thing that people haven't realized that because it's unimaginable how destructive it could be or disruptive it could be. I think you subtly touched upon or like my takeaway from what you said there is we we are about to see a change in definition of what AGI itself means because I think in public perception or pop culture perception of AGI means yeah somehow we've replicated human conscience. I think that was sort of pop culture definition of you know AGI meant for a lot of time. And as you said, you know so eloquently here, like we already have most of these tools already there and this is also one of the reasons why uh chat GPT wants to do uh desktop app and not just be on the browser because then being a desktop app really gives you that power for the digital agent or AGI whatever you call it to actually go and do things on your behalf with, you know, admin level control. So I believe that AGI will be there, be here in two years, but it will not be in a sense that pop culture thinks AGI is going to be looking like. It it is not going to be uh you know something from a science fiction movie, but it's going to be like uh in con that you hired. Um at least in the first and then that in turn sort of, you know, becomes a more mature employee over time. Uh that's what I see it and I think you you're absolutely right that most of these elements are already there and it's it's part of how you put these together in in an application format and it will also become even really relevant if these application port from your desktop to mobile to even meta AR glasses, right? And then you're constantly talking and the real pop culture phenomena of AGI like the Jarvis of Iron Man really comes together when it's on desktop, it's on your mobile and it's something you are saying. Now all this data is with it, it reasons with it, it plays around with it and you can instantaneously ask and talk to it in audio, video and you know, text format, that makes it really powerful and actually everything I said is actually available for someone to create. So it's a matter of like putting this together um, you know, all this together. I would say there are two things that need to happen which I would say two chat GPT moments which will make all of that packaged together. One is the large language models becoming so small and efficient that you can run it off of any device. And not just GPU device, because that's not going to be sustainable. And then the second thing would be an ability to do self-learning based on observation. So, if I let it listen to my conversations, if I let it see through my glasses, I don't need to fine tune or retune or or retrain it. It should be able to pick up the nuances and be able to update its knowledge and so on and be able to help that person in a specific domain based on the instructions. So, if these two things happen, then yes, that's a different definition of AGI and we are still some distance away from it. But not very far if I'm not mistaken. And I think the third inflection point I'll add to that is when we really bring all the physical data and make an element understand the physics of the real world and then you can sort of deploy robots and I think that would be a hardware uh inflection point that we might see in the near future. That's when all the Black Mirror episodes start becoming true. I think once we add the LLM on meta glasses, I think you know, you have a couple of Black Mirror episodes. That's already true there. That is, yeah, that is. Um so, I want to get back to your career a little bit and um, you