Snowflake VP of AI Baris Gultekin: No AI Strategy Without a Data Strategy & Why Skills Are the New Apps

I absolutely believe that there's going to be a ton of jobs and ton of opportunities at the same time. I don't think change has happened this rapidly before. So there is going to be a period of adjustment and that period of adjustment may be felt in a way that is uncomfortable. I believe the opportunity is massive. I feel like all of our sci fi movies became real within a span of a couple of years. Like, my daughter, for instance, will go and practice a topic to prepare for an exam and she'll say, hey, I'm learning this. Give me a series of flashcards and then gives you a series of flashcards. She practices on that. Whenever there's something she doesn't understand, she asks questions on that specific topic, immediately get answers. This is a different world. I wanted to collect all of my health information in one place so that I can do some analysis on things. It turns out it's very, very difficult to do. AI is now at a place where I can just go tell Claude to use computer use, go to all of these interfaces, scrape my own data, put it in my own computer so that I can run my own analysis. So if I were to do this myself manually, I'd have to go and click hundreds of times to try to collect all this data. It would take a time to do this. Whereas when I instruct the computer to do it, I was able to do, if this was a project, a weekend project, I was able to do that, you know, within an hour or so. Hi Baris, thanks for joining me today. Thanks for having me. Baris, I want to start with a moment. It is 2022. You've been at Google for 16 years, one of the longest tenures of anyone at that level. You build projects, products used by hundreds of millions of people. You have equity, you have impact, you have safety. And you walk away from that and you walk away to start an xyz, a blockchain data infrastructure company. And in the middle of, some would say in the Crypto winter and 18 months later, you sell that business to Snowflake and take on the most ambitious AI infrastructure challenge in the enterprise world. So my question is, what led your decision to jump into TRI and accomplish most Cs as impossible? It's a great, great deep question to start with. Yeah, I've been at Google at that time for a while and I loved my tenure there. Google is a phenomenal company. I got to work on really interesting things. Just loved it. And at the same time, I've always wanted to do something on my own. And you Know, this is something that's always been ingrained in me to just try it out. And I had a lot of ideas. I was really fortunate to be able to start a lot of new products at Google. And you know, around that time I've been thinking about what's next. I've spent a lot of time in the AI space before AI was hot and at the time it didn't quite work. So I felt like maybe it's time to try something else. I was really fortunate to have worked with Sridhar Ramaswamy, who's now the CEO of Snowflake at Google, when he was running ads and really, really enjoyed working with him. And I've kind of seen him as a mentor. So when I was talking to him, he suggested like, he, he's been thinking about Blockchain as well. And I was really interested in that space. So it was just the right time to try something else, something new, work with amazing people in a startup environment. So that's what I got to do. And again, I was really, really fortunate to work with these awesome, awesome group of people as we did. This AI boom started with, you know, ChatGPT launching and history at the time, very, very clearly. So Snowflake saw that as an opportunity to double down on AI and acquired our company with lots of AI experts to come join and build out AI products at Snowflake. And then that begins the point of that's the Snowflake journey and starting the product. And you've seen the Google now, the assistants, the voice agents back then and then now they've taken into a different form and a different shape. How do you see that right now? Especially seeing what is it like before and now that it works? Yeah, yeah. And now that it works is the interesting thing, right? Because Google now was a precursor to Assistant and it launched with a lot of fanfare. People loved the idea of getting information proactively before they ask it, and they were very helpful. Tips and information then came to Google Assistant and there was a lot of expectations with Siri, then Google Assistant and Alexa about having intelligence at your fingertips on your mobile device. The challenge was that the technology wasn't quite ready, so each use case had to be built by hand, one by one. When somebody says this, then do that, and that is by nature brittle. When you talk about voice, you're making intelligence more natural. And when you make something more natural, there is expectations of human like intelligence, intelligence, human like behavior. So when it works in a couple of your use cases, but not in most of your Use cases, that expectation breaks down. So that was the biggest challenge with the Google Assistant era and now with AI being as capable as it is, we're in a completely different world because there is clearly a deeper understanding of what's being asked and the capabilities are much more robust. So it absolutely feels a lot more natural and it kind of matches the expectation. And what's also interesting is, you know, the likes of Google Assistant started with consumers first. ChatGPT is still very consumer oriented product, but there's a ton of opportunity in the enterprise space, which is relatively unique. I think having enterprise almost exceed the pace of innovation, which is fascinating to see from the workflow aspect of things too. Like before you have to design each of the logic and then the business workflow. Now that the LLMs can kind of understand that through skills and then just the training data. So one difference I think in being at Snowflake is that the AI is coming to the data versus the other way around. So how is that architecture critical for the future of generative AI applications? There's the governance aspect of things, but is there also different advantages that are not obvious to people today? Yeah, I mean, maybe if I just zoom back a little out. Ultimately agents are incredibly powerful, but they are as good as the context that's given to them. So data is incredibly critical and data is the most important asset of any company. So therefore companies platforms like Snowflakes, where there is a lot of you're starting with a solid data foundation, allows companies to build on that data foundation. So when we started, our customers essentially told us we want to bring AI to run next to data versus to bring massive amounts of data to AI. And as you called out, one of the biggest benefits is security and governance. So when you do that, all of your governance that you put on your data, all of your access policies that are already set are by nature respected versus you having to replicate it in multiple places. So it's substantially simpler to build as well as more secure and governed. A huge benefit also is by being close to data, we're helping our customers build AI ready data so that all of their data is much more easily usable with AI. And then we help them do that by bringing the semantics of their business onto their data platform so that AI can understand how their data is laid out, what are the things that they care about, what are the metrics, what are the terms and synonyms that are used in their business and so forth. So that becomes incredibly valuable as well. So at a high level, ultimately it's all about Trust. For companies to be able to scale AI in their organizations and with their customers, trust is incredibly important. And trust shows both in terms of kind of are you governing the data? Is only people who have access to this data, you know, able to use it? Also are you able to build high quality AI solutions? High quality, quality AI which boils down to the deep understanding of their data and business. So I feel very fortunate that Snowflake, being very close to data, is very uniquely positioned to take advantage of this opportunity with our customers. It is also very interesting because so the customers have different system of records that they've been signed up over the years and then think they had applications built for like Salesforce saps and Workdays of the Day and then each of them have their own database UI like custom business Logic just hidden and they're unable to speak to one another. So like in an agentic future, do you think that the application layer kind of disappears somewhat because then instead we'll have kind of agents directly query a unified data execute actions and maybe the agent of a system of record, like a system of agent then speaks to a system of agent of another system of record provider and then that builds like that agentic future. Like how do you think that is going to evolve? I mean it's a good question. It's a question that's on top of everyone's minds. How does the world change? How do these platforms evolve? I absolutely believe all of these platforms will evolve to a point where majority of the interactions are done by AI versus humans. Today these platforms are built for humans. But that is absolutely going to change. And in that world when those interactions are mostly driven via agents, the interface basis change you're on, you know, you, you, you've made another assumption in, in that question which is do these companies disappear altogether? I don't necessarily see that to be the case. You know, if you're using your example Salesforce, if you're already using Salesforce, you, you might interact with Salesforce differently. You, you might interact with an agent that captures that information there and then an agent can talk to a Salesforce agent and so forth. However, it's just that each of these solutions, each of these providers will evolve to support this new paradigm versus not. I am sure there's going to be a ton of new opportunities that will emerge as well. When you're able to kind of break the barriers a little bit so that there is level playing field in natural language for agents to interact with one another, those silos will absolutely start breaking down. So that there is tremendous opportunity for consumers to be able to more easily that kind of get, get work done across multiple different platform solutions and so forth. So it's hard to imagine exactly how that's going to evolve. Absolutely, it will evolve and it's moving very fast. Like it's like even I've seen the multi agentic framework online. Like the agents people are downloading, I think the open, open cloud and then on a Mac and then just running it and then do things. And then that's also an interesting feature too in that future. Let's say I have an agent and then there's another agent someone else has unleash. My agent could negotiate meeting times or approve purchases or shopping behavior. What do you think will be more of the protocol layer for that agentic web? Do you think that's still going to happen as an API calls or payments manage? What's happening is the LLMs are now conversant in natural human language. There's a lot of benefit to series of APIs. There are, you know, great protocols emerging that's making this very kind of easy for, you know, for agents to talk to one another. You know, you have MCP for just getting contacts from various places. You have other agent to agent protocols emerging for agents to talk to one another. You have protocols for commerce emerging. I think it's in the early days but you know, protocols will emerge. And the beauty of all of this is because it's natural language, I think the adoption will be much faster. Yeah. And also in a voice that we understand can speak every single language around the world and that makes it much more accessible. And among this, what is one case that's exciting you the most that you see in day to day? I mean right now it's not quite a single use case, but I am really, really excited about the power of these reasoning mod when they meet skills just to expand on this, these you know, skills is essentially instructions and some scripts to take certain actions. And you can imagine many, many skills that will get developed that will get open source and you know, for these agents it's like in Matrix. I'll say now I want to know karate and I know karate, right. So being able to go tap into these skills and then use, use it for various purposes expands the kind of surface on which you can take action. So I'm incredibly excited about, you know, that the power of that more concretely, you know, on the enterprise space, you know, you have, let's say for Snowflake we built our own agent for our Sales team. It's a very powerful agent being able to look up any customers information, any open issues they have, whether they have any renewals, what are the use cases that they're working on. Really, really powerful. But you could imagine expanding it to, you know, you can add a skill to go, you know, plop information from some repository, you can add a skill to draft marketing campaigns and so forth. And each of those can be really well customized. So excited about that potential and it is huge because then, because in each skill customers can, can include their, their business workflow, business logic that is like very niche to them. And then there will be a different AI. So personalized, tailored, just living to automate their processes. Exactly. And that applies to both processes within the company. But you could even personalize it for your own processes, for your own workflows and automating many of your own needs as well. So because this is super easy to create, I expect a lot of personalization to happen here too. And again looking back, you seeing the Google now, the Google Assistant time on the user behavior of interacting with AI agents, especially at the time it's been very voice driven, right. Like you were asking questions. Do you think that as UIs are changing that we will adopt maybe one agent that has thousands of skills can do my legal for me, can do my accounting for me. And I just say it to the. I definitely do think that the user experience is very nascent right now. It's all text based and if you go back to how computing evolved you can go back to the terminal and it was all text which I feel like we're now back there and we love using the terminal for coding assistants and so forth. But from that we evolved to graphical user interface which was kind of more natural. Vision has a lot of, you can pack a lot more information in it and then other modalities are coming. So I do believe that interaction will absolutely have a lot more visual components than what we have today. Like being able to just ask a question and rather than answering one question at a time, just seeing a form, quickly tapping on something to, to fill out a form fields faster, stuff like that will evolve. Voice is already had made a comeback. You know, it's quite popular. You know I love talking to my AI agent. You can brainstorm, you can, you can talk about while you're driving. And in other ways I think that is getting much more natural. You know, whenever things are natural they'll get adopted more. So I do think that voice is an important modality but I think the part that is going to develop more is the user experience of interacting with these agents with visual, you know, visual interactions and each of them. Because it makes me think about a little bit of the data layer as well, because each of the UI will be connected to an agent and then that agent will use skills, will use the data to connect and make a judgment or decision based on the agency that we give to them. How would you see that? Each of the agents sees the same source of truth. So for example, I may ask for it to negotiate pricing for me and my agent goes grabs a pricing from somewhere and another agent goes, grabs a price elsewhere. So there could be some discrepancy there. So how would we, how do you see that same source of truth be. Be accomplished at the data layer? Yeah, I mean, I think we're talking about these agents getting more and more human like in their capabilities. So you. We could absolutely look at how we do it and expect that is one way to do it. Right. So like when you're negotiating a price, you have a sense of value for yourself, you have a sense of price of similar items in the market. And those are things that agents can absolutely have access to. In terms of common grounding, I am not sure if that's necessary. You know, you have the common grounding which is the world knowledge and the web. And then increasingly, what I expect will happen is my agent will have a lot more knowledge and information about my own needs and preferences. And that will be different from your agent by nature. But also both of those will be also kind of grounded on all of the accessible information that's out there. Each agent will defend, in a way, their master softcut. So I'd love to talk a little bit about the economics of intelligence because in a way, we are now in the era of similar to renting compute, we are renting intelligence. And then of course, training a frontier model costs millions of dollars and inference at scale costs a lot as well. How do you see that evolve over time? So especially the architecting a role where compute happens, where the data already lives. Does that data proximate compute model democratize AI? Or how from cloud providers to data platform providers like Snowflake evolve over time? One of the philosophical questions. Yeah, yeah. You know, clearly there is democratization of data as well as democratization of intelligence happening. Right. You know, democratization of the data is happening because you can now access with natural language to massive amounts of data, sift through it, get insights and use it. And democratization of the information intelligence is happening by these models. Getting both substantially better as well as cheaper over time. At the same time because the capabilities are increasing, people are using them more and more and more. So. So the overall cost and usage is increasing in terms of, you know, I don't really think about balance of power between kind of data platforms and AI platforms and so forth. There's clear, you know, stack and I also see that everyone is going up the stack as well as down the stack at the same time. You know, when you see where the value has been accruing. Looking at the likes of valuation of Nvidia for instance, huge amounts of value is accruing at the infrastructure layer. You have hardware providers like Nvidia, I would say then the LLM providers are capturing a lot of value and creating a lot of value. Then you have the data layer, then the application layer and so forth. Increasingly many companies are focusing more and more on the application layer because ultimately that's when you create a lot of consumer value, a lot of customer value. And I see that continuing the pie is growing so much that it is not a fixed pie. We need to all kind of protect the share or it is that the opportunity is massive. So everyone is running to build high value, high quality products to be able to serve and capture this massive potential. So I see the kind of tide rising, all boats versus some tug of war between platforms. And in a way everyone is collaborating with each other as well. So that brings the improves the footprint everywhere. It is interesting because the so we said UI is changing, user adoption are changing, people do want it more, so they do you want to use it more. But at the same time costs are going down because we're able to produce better models closer to data. But at the same time user behavior is changing as well. The way maybe in a way the jobs and then at the enterprises are evolving over time. I'm not a big buyer of that. Like oh, jobs are going away. There will be jobs of course and there'll be even more. Looking at from your lens and then the industries that are shaping, what do you see? And there's one to pay attention to which is I absolutely believe that there's going to be a ton of jobs and ton of opportunities at the same time. I don't think change has happened this rapidly before. So there is going to be a period of adjustment and that period of adjustment may be felt in a way that is uncomfortable. I believe the opportunity is massive. However, in terms of how this is all going to shape up, I. I don't really know. I see everything is evolving. I look at education the way my kids are learning is changing and should change the types of tools that they have access to. How they use these solutions to deeper learn, understand, customize, personalize. Their education is a massive opportunity. How much that levels the playing field in education across the world is a massive opportunity. So you have education on one end, you have clearly a lot of transformation happening in technology on another end. Health is changing. Now everyone has an AI solution that they can quickly go ask questions to about health related issues. And AI is incredibly good at giving no health advice for legal reasons, but pretty good at giving you information that is helpful. So I think the change is not really focused on one industry because it's. So we're talking about intelligence. It's a very kind of human term. Right. So it affects I think many, many industries. All industries. From the education aspect that's very interesting too because from kids they like when I was growing up I was still connecting with a normal phone to dial in so I had to so my parents wouldn't be able to use the phone because I have the Internet and right now like kids have just chatgpt can do so many. How do you see how are your kids for example looking at and using AI? Yeah, no. I feel like all of our sci fi movies became real within a span of a couple of years. You have self driving cars on the streets. You talk to your computer. Like my daughter for instance will go and practice a topic to prepare for an exam and she'll say hey, I'm learning this. Give me a series of flashcards. And then it gives you a series of flashcards. She practices on that. Whenever there's something she doesn't understand, she asks questions on that specific topic. Immediately get answers like this is. This is a different world. Right. Versus you going through your notes and your books to try to understand something and have to have a human to ask questions if you don't understand something. Yeah, it's in a way that they only learn what they need to learn and how it's portrait and interpret it is perfectly designed in a way which is amazing. There's also this hyper personalization that will I think get realized in all sorts of different ways. If I want something done now, I can just go create a custom app just for myself and I can do that within tens of minutes. And now I have exactly what I need. You could apply that to everything. You can have a very hyper personalized education just for your needs, but apply it to every single situation. Do you think that will like Result in more internal tools, internal applications getting built like it could be even a student building like a flashcard application to an enterprise building their own expense management software by wipe coding or a. We can managing it. Yeah, certainly, definitely. I believe in that. And these products are also easy to stitch together as you called out earlier. Right. You know, because the interfaces are, you know, mostly natural language based. You could imagine I've just built this and I'll have it talk to this other system that somebody else built. Exponential set of capabilities that you get because you're constantly building on top of one another. And it's also you have in full control as well on the design and improvement of the process. Right. In terms of the alignment side like of course the foundation model, companies do a ton of work on post training like trying to align the model to human values. When it comes to enterprises, I think it's a little bit different because every enterprise in a, in a competitive world and everyone is trying to win for themselves while again their partnerships and those sorts of things are going on. And aligning enterprise AI with corporate values is an interesting space as well. And I like to ask a few questions around that. Especially like if an agent, for example, optimizes based on profit but then accidentally does something like forgets there was an agreement or the legal side of the things and if that happens, who's responsible? Is that the model provider who couldn't do the post training well, is that the enterprise? They didn't do the agent configuration really well or is it just the CEO of that enterprise customer who needed to think a bit better about the accountability side of it? There's no direct answer to this, but just brainstorming. Yeah, I think building systems with strong controls, guardrails, evaluations is, is a very hot developing topic as anyone is building production ready agents. You need to have a set of guardrails. You need to make sure that those guardrails are followed by the agent. You need to have a series of evaluation data sets and you know exactly where the agent trips up you for the task that you build the agent to do. Usually the recommendation at least for wear technologies today is to have, you know, scoped agents for which you have, you know exactly what the skills are, what the tests are that you want the agent to be able to complete and build out those evaluations and know exactly where the agent is good and where it's not good. In terms of where responsibility lies. I think responsibility lies with the company and the person using these products as tools. Ultimately these are very, very powerful Tools and the judgment is still on the company, on the developer, on the person to figure out how to create it, how to tune it, where to apply it, where to use it. And it is like you have access to a lighter and a lighter can light a candle or light an entire building down by burning curtains. So it depends on who uses the tool and knows exactly what to do with the tool. Setting those guardrails is really important. It's an exciting space and especially everything is getting rebuilt from scratch. And we mentioned about education, we mentioned about healthcare that is evolving. There's also the aspect in healthcare where the data has always been so problematic. There's always stuck somewhere and some data warehouse in an on premises and it's very valuable data and it's being used mainly. Maybe a model can learn from it and then help thousands. How do you see that space itself evolve to break those barriers to entry for themselves? Healthcare is a very regulated industry for a good reason. The data is valuable, but at the same time regulation is there for good reasons. So development here needs to be very clearly respecting all of those kind of regulations and the guardrails that are put in place. So I'm not a healthcare expert, so I don't have, you know, depth of insight here. So I'll give you more personal examples of how I've been using my own health data. Just recently I wanted to collect all of my health information in one place so that I can kind of do some analysis on things. And it turns out it's very, very difficult to do. You know, you have seeing various different practices that are captured in multiple different systems that do not talk to one another. But AI is now at a place where I can just go tell Claude to use computer use, go to all of these interfaces, scrape my own data, put it in my own computer so that I can run my own analysis. So if I were to do this myself manually, I'd have to go and click hundreds of times to try to collect all this data. It would take a long time to do this. Whereas when I instruct the computer to do it, I was able to do, if this was a project, a weekend project, I was able to do that, you know, within an hour or so. And it was quite valuable data. Yeah. And then it's, in a way it's so UI driven as well. Like, you know, a agent goes and navigates through the UI and then downloads that data and makes it accessible. I did a weekend project on, I was getting a lot of spam calls, but they're voice agents. That are spam calling me. So I was looking to build a voice agent that will also spam, call back the other voice agent and then see which one will hang up first. And it's also an incredible result too. No, I didn't scale it, tried it a few times in a way, like back to the healthcare side. You're right, the data is fragmented, but in a way still being able to go with the agent. I think that's the early signs of a solution that is cooking to help us bring it into the one place. So there's a talk about of course reactive agent and then autonomous agents. And then I think in enterprises we didn't see a full autonomous agent taking ownership of the full end to end business workflow yet. What do you think is the limiting factor there for them other than control? Right. I mean, I think it really does boil down to trust. As we were saying right now the technology is at a place where you can automate a lot of things. And we do see with our customers a ton of automation happening in majority of these cases. Like there is human in the loop and human in the loop for again good reasons. You know, you want some validation. Increasingly, as quality gets to a level where it surpasses even human error rates, the automation will not necessarily require in certain cases the humans to be in the loop for those things. We see this with low stakes automations right now. So there is definitely automation happening where humans aren't in the loop. In high stakes cases still there are humans in the loop looking at it, things, improving things. But even in those situations, what used to take weeks now is taking only minutes because of the efficiencies that we're able to get. Even in customer support. I think it's when a ticket comes through, the agent can manage the entirety of the conversation and go ahead and resolve it. Right. And then we see that happening in cases where it doesn't. It can fall back to a human but with a summary and a set of resources for that human to make a judgment call. So when the agents can do changes in the system of records, like if I have one agent that's able to change the database system and an email, be my management, essentially be my AI manager. So that goes managing all my tech stack, it will interact with many different application systems and from other system side the similar things are happening as well. Like an application provider is giving their agent and then enabling me to go and access to those system of records. So it does sound like the system of record is going to stay for a Long time, of course. And then that data layer is still very, very, very important. How do you think though that data layer will change with AI? And I know that there's the, we got now cortex code as well and ga, so it's able to go in and then be able to speak to the Snowflake OS in a way and then let me understand. And that's a huge change of user interactions on the data side. So looking there, how do you think that the data layer will change? Right? Ingestion's happening, pipelines are happening, data is getting stored, but the user behavior is changing and possibly possibilities are changing. I think a massive amount of new data just got unlocked. So that that alone has implications in terms of use cases like you could imagine. We have customers that have tens of millions of documents and you can go and then talk to those tens of millions of documents and ask questions, which is now possible, which was absolutely not possible before. So it's great. But whenever you want to do something like hey, you know, tell me how many of my contracts are expiring tomorrow, let's say you haven't done the pre processing of all of that before. It would not be a good architecture to go okay in real time. Let me go figure this out instead, you know, what people do is I'm going to run an extraction job and then figure out all these different fields and I'll extract it, I'll put it together so that I can take a look at data, aggregate this data, get some insights from it. So what I see happening from a tech stack evolution perspective is a couple of things. One is all these different data sources that did not talk to one another before are now becoming available for an agent to connect to and then reason with it. MCP is a big unlock from that perspective. So I could imagine there are different data sources. And I have a question, and the question requires me to look at multiple different data sources. A protocol like MCP allows the model to go, take that information and then give you an answer. And that's great for real time lookups for something that requires you to reason over massive amounts of data that breaks down. So you still need a data platform like Snowflake to be able to have access to that data in the first place so that you could run large scale analytics and then get insights there. I think what's happening is the capabilities of these platforms are emerging. So what you can do with massive amounts of data is changing, is increasing. So one example, for instance is, you know, super simple example, let's say you have lots of, you know, customer reviews. And being able to just say, summarize all the customer reviews for me now is, it's, it's, it's literally, you know, someone's, in 10 seconds can get an answer from massive amounts of reviews. Just glean insights, understand what's working, what's not working. What are people complaining about at. That is a capability that just got unlocked, so you can imagine many more such capabilities emerging. And that requires both these systems to talk to one another in many cases. It also requires, again, the core principles. We like saying at Snowflake, there is no AI strategy without a data strategy. Ultimately, you still need to think about how do you bring all of your data together so that your data assets can be joined, can be combined when you're running, trying to bring insights across your data assets. And in a way, there's the, the, when we build the summaries or even sentiments on a certain customer review, it's like AI scientifically generates data and then that generation is already an increase in the data layer, the support for it. And then there's also the, I think the semantic layer to be able to then retrieve that data from tables. Yes. Which is, which is really interesting that in a large amount of context, millions of rows. So semantic models, semantic layers are becoming very crucial parts of the data stack. You know, there's a lot of conversation happening these days around kind of the full ontology of a company's knowledge. Right. You know, you have your own knowledge base with lots of documents. You have your structured data, and that structured data has your business semantics. You know what your metrics are, how do you define the profit, or how do you define whatever metric that you care about? All of those are crucial parts of enabling data for AI. Right. So as companies think about how do I get my company AI ready? It's, it starts with getting your data in the right place. You know, you break down all your data silos, you make sure that all your data is secured and governed so that only the right people have access to the right data. Then you have, you bring the semantics of your business onto that data. You bring your knowledge base of all your documents accessible, then you build out your solutions. That alone is a massive space and it's only going to mature as more and more capabilities get unlocked. And it's at the end of it, the enterprises do operate on their data and then they need to. The more clean it is, the better. Then the agent understands reasons and then gives them a response on it. I'd like to Ask about because you, you've been in startup world and right now the startup world is also booming. Like there are so many opportunities, there's a new startup coming out to disrupt so many different verticals from healthcare to the supply chain. And at the time when you tried your blockchain startup, and that was also taking a path where not everyone, everyone's fearful to take. Right. The difficult challenge. So if you were to go back at that time, and let's say that time is now that there's so many startups, so many opportunities, what would be one problem that you would want to tackle first? You know, with startups, the idea of, let me think of ideas, let me look at it, usually does not work. Ultimately you're trying to really deeply solve a problem and then for you to really deeply solve a problem, you need to have depth on that problem. So it's either something that you want to solve for yourself or you really deeply understand and solve. Not only that, but you also need to feel really passionately about that problem. Because startup world is a challenging ride. Right? You know, you're a small fish, you're trying to grow. Every day is a countdown until the next event, either a next raise or count down to product profitability. Given that, you really need to feel passionately about that space, otherwise it doesn't work. So that's the preamble essentially. If I were trying to do a startup now, what would I do? I don't know. I would start with a problem that I feel really passionately about and think about, hey, how is that problem going to get solved the best way. And the answer is not always I need to go start this company because no one else is solving it. There's all sorts of other kind of ways of thinking about it. Usually people who build successful products are ones that they start these companies because there is no, they cannot not do it. There is this burning desire to go make it happen. And then you go make it happen. Yeah, it is. And then one problem that a founder tries to solve turns out to be hundred more problems within that problem that they need to figure out how to fix when it comes to the ecosystem evolving related to the startups. Because now in a way like there's the concept of being a rapper and everyone is being a rapper of something. How do you see Snowflake like supporting that ecosystem? Because there's the now, the postgres, the data storage aspect, the agentic aspect that startups can take and then build an application layer on top take to market. Yeah, yeah. So Snowflake has always been a very partner friendly company and the platform to build on top of. So what we do with our partners is there's a series of products that might be helpful. Some of those you called out, like startups that build on Snowflake, are building on Snowflake for a couple of reasons. One is Snowflake makes it very easy for our partners to bring their solutions to Snowflake's customers. The tricky thing is access to data is usually the biggest blocker for startups and Snowflake creates a secure environment for Snowflake's customers to provide that access to other partners. So that's a very attractive platform for others to go to come and build on Snowflake. And increasingly Snowflake offers inference that's super easy to use across all model providers that a startup can use easily. And the benefit is it's running within the Snowflake security boundary. So it's attractive for Snowflake's customers. If you're selling to those customers, you can build agents again on the data and then make those agents available for kind of customers on Snowflake's platform. Again super powerful, being able to do more complex reasoning tasks with data and then making it accessible and shareable across the Snowflake platform. You called out postgres as a new capability that's now possible to do and helpful for startups to take a look at and use as well. And then as startups grow, Snowflake's own kind of analytics platform at some point also becomes interesting. Overall, the idea is that partners working with Snowflake within the Snowflake's platform make their products much more accessible to a large audience of Snowflake customers. Yeah, it's building on top of the platform. There are so many advantages too. And as the platform scales then they scale with that as well. I'd like to ask one last question and then around, mostly around the coding agents versus general agents and there's the conversation around with especially Dario De from Anthropic says that coding agents would be the path to lead to a very general intelligent system. And I think we've solved it. The cloud code cloud and coworker coming up as well. And on the other hand there are vision models and then there are world models. So many blending of things happening. I'm against enterprises they are adopting at the latter part of the curve. But based on what you are seeing as the data layer, your experience from startups to experience in enterprise adoption and leading Snowflakes AI, what do you think will unlock that general intelligence is the next bet. I think we both sometimes overestimate what AI can do and underestimate it in some cases. Like when we talk about general intelligence, a human like intelligence, I don't think we're anywhere close to it. And I don't think our current architecture's trajectory is necessarily going to get us there in the next one or two years. At the same time, this current trajectory is incredibly valuable and incredibly powerful and will solve a ton of problems and will move humanity forward. So I think, you know, it's, it's a bit of a, an academic thing to figure out, hey, do, do we have human like intelligence? Are we going to get general intelligence and when are we going to get it? To me, that's a less interesting question. More interesting question is how are we, where are we in terms of the capabilities and are we making the most use of it and, and, and you know, where is the technology going and then how much more opportunity gets unlocked? How does life change as these capabilities become available? So I see them more as tools and these tools are very helpful and I'm incredibly excited about these tools becoming more and more helpful and then more and more kind of cheap and accessible. Yeah, at the end of the day, it helps us solve some problems and then be tools and be helpful to us. Yeah, exactly. That said, thank you very much for joining me today. It was great. Yeah, thanks. Thanks for having me.