Snowflake VP of AI Baris Gultekin on Bringing AI to Data, Agent Design, Text-2-SQL, RAG & More

Hello and welcome back to the Cognitive Revolution. Before we get started today,

Revolution. Before we get started today, a quick final reminder. If you dream of a career in AI safety research, the deadline to apply to Matt's summer 2026

program is January 18th. Listen to my recent episode with Matt's executive director, Ryan Kidd, for all of the reasons that you should consider applying and then get started at mathsprogram.org/tcr.

mathsprogram.org/tcr.

Today, my guest is Barish Golden, vice president of AI at Snowflake, the cloud-based data platform that now describes itself as the AI data cloud.

Beerish came to Snowflake along with Snowflake's current CEO Shrear Ramaswami as part of Neva, an AI powered web and personal knowledgebased search engine

that Snowflake acquired in May 2023.

Since then, he's been working at the intersection of frontier AI capabilities and hard enterprise realities, deploying these systems in environments where security, governance, and reliability

are strict requirements.

As you'll hear, Snowflake's core philosophy is to bring AI to the data rather than sending sensitive data out to model providers. And in this episode,

we unpack exactly what that looks like in practice. We cover a ton of ground,

in practice. We cover a ton of ground, including the massive ongoing unlock of unstructured data, which is making the 80 to 90% of enterprise information that

was previously trapped in PDFs and other documents queryable for the very first time. The current state of both

time. The current state of both texttosql and rag systems and why reasoning models have finally made natural language data analysis reliable

enough for business users. the

trade-offs between using frontier models versus smaller specialized models and when to use structured workflows versus letting models choose their own adventures. How data residency

adventures. How data residency requirements are shaping partnerships between cloud providers, model labs, and platforms like Snowflake. How AI coding assistants are changing the discipline

of product management by enabling rapid prototyping of working features. where

Beerish sees value acrewing in the AI stack and his prediction that horizontal applications will win out over narrower vertical solutions and finally why

Beerish takes the over on my timeline for autonomous drop-in knowledge workers and what he believes will have to happen first.

If you want to understand how large enterprise companies are deploying AI today and what's really working as they mature from the early experimentation phase to the ROI at scale phase, taking

all of the operational complexities and security and governance concerns into account. I think this conversation will

account. I think this conversation will be perfect for you. And with that, I hope you enjoy this deep dive into enterprise AI adoption and the future of

data intelligence with Barish Golden, vice president of AI at Snowflake.

Barish Golden, vice president of AI at Snowflake. Welcome to the Cognitive

Snowflake. Welcome to the Cognitive Revolution.

>> Thank you, Nathan. Thanks for having me.

>> I'm excited for this conversation.

There's going to be a lot to learn. I

think people, we have a very diverse audience. The number one profile is AI

audience. The number one profile is AI engineer. And within that profile,

engineer. And within that profile, people work at a lot of different kinds of organizations from solo entrepreneurs and consultants to startups to enterprises. So some people will

enterprises. So some people will certainly know Snowflake and will work at organizations that are customers of Snowflake. Others probably have heard of

Snowflake. Others probably have heard of it and don't really know too much of the backstory. So maybe for starters just

backstory. So maybe for starters just kind of give us the real quick Snowflake 101 and then I'd love to go into how AGI is Snowflake today. [clears throat]

>> Sure. So Snowflake is a data platform.

We call ourselves an AI data cloud. So

what that means is our customers bring a lot of their data onto Snowflake so that they can secure it, govern it and analyze large amounts of data for various insights, dashboards and the

like. And from an AI perspective,

like. And from an AI perspective, because there's a lot of gravity to data, our customers do not want to replicate data in multiple places.

Instead, they want to bring AI to run next to data. So that's a very high level overview. And AGI PIL is an

level overview. And AGI PIL is an interesting phrase for us. We're quite

practical. We serve large enterprises and the goal is to get to high quality AI agents that that create positive ROI for customers quickly. And that can

happen today and it is happening today.

Super excited about where things are.

>> I definitely want to come back to the bring AI to data strategy that you guys have in a few minutes. But to just double click a little bit on the before and after because obviously Snowflake

has been around for a while before certainly anything like the AIs that we have now were available. So what were people doing before with Snowflake and

what are the new AI use cases that have been unlocked over say the last I don't know when you would start the clock right do you start the clock at chat GPT or do we need it was that like not quite strong enough to actually make things

work but yeah there's a lot of different dimensions maybe let's start with before and after >> sure so I'll start with the before so a lot of our customers have been using Snowflake mostly for structured data

initially and and this is where they'll bring the data in and and then they'll run large scale analysis um either to um to have insights uh to power BI

dashboards for instance um or to do to to do various analytics to understand their business. The uh the value of of

their business. The uh the value of of the platform is to be able to you know bring data from all the different places to break the data silos uh so that you could run analysis across uh you know

large amounts of large amounts of data.

Um so what is happening with AI is um there's a big unlock of course of unstructured data. Um so and this this

unstructured data. Um so and this this plays out in two ways. One is um if you have you know thousands uh you know hundreds of thousands of documents for instance you can now extract structure

from these documents uh and then you can analyze them. If you have contracts for

analyze them. If you have contracts for instance and if you want to say hey what are the contracts that talk about this specific thing? how many of them do I

specific thing? how many of them do I have uh or uh you know what are the contracts that are expiring soon uh in this category. So being able to run

this category. So being able to run these analysis super easily is not very very feasible. Then there's of course uh

very feasible. Then there's of course uh being able to uh do a lot of again kind of large scale analytics work but with um with a lot of ease. So a lot of the

pipelines that used to be built around I'm going to go classify my data. I'll

I'll extract information from it is now very very simple to do. So uh the way uh we've done it is we've again brought AI to directly work with the uh engine uh

for analytics. So you can do things like

for analytics. So you can do things like you know classification extraction of of that data very very easily and increasingly of course uh there is a lot of interest in bringing natural language

interface to all a company's data so that you could just talk to your data you can democratize access to all that data uh for the organization. Could you

give a sense for sort of the balance of structured unstructured or then I guess you're also saying that like structured data is getting structured through the process of basically AI retro

annotation. My sense is that and I think

annotation. My sense is that and I think of oh gosh there's so many vector databases the exact name of the one the where the founder told me this is slipping my mind but one of the

interesting things that I've understood to be happening in general with business data is that structured data was kind of like the tip of the iceberg in many organizations where it was the most

usable kind but it was actually like a relatively small amount of the data. It

was Chroma, it was Anton from Chromad who said that like most of the data that was going into Chromma DB had never been in a database before at all. It was just lying around in various places. So, have

you seen like a sort of great unlocking of people like dumping more and more data into Snowflake because now they have ways to make it useful where it just previously wasn't even worth it.

Yeah, we're absolutely seeing this uh you know 80 to 90% of all data is unstructured data and and you know because there weren't a lot of easy ways to process this, it was not necessarily

seen as the most you know usable uh data and it is now very usable both from a kind of extract and then bring structure um to it perspective as well as just you

know talk to all of that data uh you know find the right information you know using you know vector DBs for instance and then and then build uh in agents and and chat experiences of it. Uh so we're

seeing this and and it is again playing out in both both ways both uh more and more data is getting structured um so that you could run analytics on it down

the road as well as you can just use all of this data in conjunction with the structured data that you have. So for

instance um you know if you want to build um you know anything let's say a wealth management agent so you still need to be able to look up what the you know stocks

are doing in a structured way but you also have all of the u equities research uh that that that's in PDFs that you'd like to be able to use. So being able to combine both structured and unstructured is incredibly important for real world

use cases.

>> Where would you say we are on text to SQL today? It's been a while since I've

SQL today? It's been a while since I've done a show on text to SQL. There have

been probably two episodes on this theme historically and I guess there's last I checked there was like a range of opinions where some people were like yeah it's just not

really there. Other people were is there

really there. Other people were is there but you have to do a lot of work to make sure that you have a good semantic understanding because a lot of the SQL databases that come in are like there's multiple columns and there's tribal

knowledge on teams that like we don't use that column anymore. or we haven't deleted it, but we don't use it anymore and it's superseded by this or there's all these little nuances that kind of live in people's heads. And so some people have said, "Oh, the models can't

do it." Others have said the models can

do it." Others have said the models can do it if they have enough of that kind of context. What does a process look

of context. What does a process look like today? And how good does it get if

like today? And how good does it get if a new customers I want to start to enable these talk to my data with a sort of texttosql kind of strategy? What does

that look like now?

>> Yeah, I mean you called right. It's been

traditionally very difficult for models to get text to SQL right and there are various reasons for it. It is first of all if you ask what's my revenue the answer is there's only one answer. So

the the margin of error is very low and and the expectations of quality is incredibly high and the reason it's been really difficult for these models is because a lot of you need a lot of

semantics to be able to figure out where to get the data from. So first of all what is the definition of revenue? What

is the definition of profit can change and then how is it modeled in the data side can can be tricky can change when we're talking about real world scenarios we're talking about thousands and thousands of tables that that have

hundreds of thousands of columns in them to be able to go and reason about so it's been traditionally very difficult what I'll say has happened in the last 6 months to a year is with the reasoning

models getting substantially better and increasingly being able to bring the semantics relatively easily onto the platform. We've had pretty substantial

platform. We've had pretty substantial gains in quality. So we now have our product to do this for instance is Snowflake intelligence and we're seeing tremendous demand for doing text of SQL

and the quality is at a place where you can now deploy them very broadly. So

very high quality and very useful because the structure data is quite useful >> and when you say deploy broadly you mean to users who are not data >> analysts. That's right. That's right.

>> analysts. That's right. That's right.

So, so for us for instance this product I I mentioned Snowflake intelligence is our agent platform and it's it's being used by business users and it is the fastest growing product that we have on

Snowflake because we're now making you know large amounts of data easily accessible to business users to ask uh you know questions and get insight very quickly. In the past they would have to

quickly. In the past they would have to go to an analyst uh who's familiar with the data to for them to kind of build some analysis and then get back to them a week later. Now they can just directly

ask questions.

>> What does that process of bringing semantics onto the platform look like? I

can imagine >> a sort of big setup one time where you go out and interview the people that have set these things up and and should know. I can imagine at runtime you might

know. I can imagine at runtime you might have to come back with questions and say, "Hey, I've got multiple columns that are ambiguous here." I did one episode also with a company you may know called Illumx where they had a really

interesting strategy that was around basically building what they considered to be the sort of canonical like abstract ideal form of an enterprise in

each major vertical that they served.

And then they built their query engine off of that ideal. And then the mapping process was like okay now how does your actual real world enterprise deviate

from this ideal let's map all that out but then we know that once we've done that mapping the the logic on top of it will be trusted what mix of strategies are you using what you find to be effective

>> so so first of all again reasoning models are now at a point where AI can help substantially in building out a semantic model and uh the

inputs to that uh are like both of course you know the data that's um in the system all of the metadata that's in the system the names of the tables and columns as well as the data underneath

them u but um we have built a series of connectors to things like BI dashboards uh that also have a lot of semantics in them that that's super useful in

building out semantic understanding for the organization also things like you know queries people have been running in the past like these are all kind of hints uh for these uh agents to go help

our customers build semantic models. Um

Snowflake has recently also announced u what we're calling open semantic interchange uh which is uh a uh and and

trying to create an open standard for sharing that semantic model across um you know across the different platforms so that you know we can more easily create these common semantic

understandings for AI to act on.

>> Okay, that's interesting. How does that can you unpack that for me a little bit?

Like how does that work? What do I do as an enterprise if I want to adopt that standard?

>> Yeah. So, uh this is it's it's still early. Um so we're working with

early. Um so we're working with companies like Tableau um Omni uh you know with BBI platforms as well as uh as well as others technology providers to

create a um an exchange uh format so that if you create a semantic model in one platform you could just use it in in a different platform. Uh so there there's there's active development now

with u all the parties in the uh in in in in the open semantic interchange to define uh what that uh interface is so that we could support an open interchange of of the semantic model

essentially. So what that would look

essentially. So what that would look like is a customer can go to uh Snowflake for instance they can go you know build out their semantic model um and then they could reuse uh that

semantic model uh in in another place that supports the open interchange.

>> For one thing I wish that would come to electronic medical records sooner rather than later. I feel like yikes has been

than later. I feel like yikes has been my recent experience there. I wonder how that going back to the first question about how AGI pill various organizations

are. I wonder how that how do you see

are. I wonder how that how do you see that from a competitive dynamic standpoint because I think one of the things that's most interesting where you know the dice are in the air

so to speak it feels to me right now in the software market is like sure seems like the pace of software development is increasing dramatically.

people are I think mostly it's outliers or just plain bluster at this point to say that people are like deleting systems of record and rolling their own in house but you've at least [clears throat] got that talk is out there

>> and then presumably everybody is gez if I can >> maybe already see a major acceleration in my software development or if I can project a year or two in the future and I can see a clear path to a major and

I've vibe coded three AI apps for family members for Christmas presents this year and the acceleration at that level is certainly very real. Then it seems like everybody is going to be incentivized to

like try to go take sort of some conceptual territory from companies that maybe used to be partners, used to be compliments. It seems like it's headed

compliments. It seems like it's headed more toward competition. So all these companies that you named, right, historically you specialize in one thing. They did a little bit different

thing. They did a little bit different thing. They work nicely together. You

thing. They work nicely together. You

got a lot of customers in common. Great.

But if I'm them, or maybe if I'm you, like I might be I might start to worry at this point. Geez, if I'm Tableau, should I be afraid of Snowflake? Are

they going to come after me with something that sort of replaces what we do? And do I want to be partnering with

do? And do I want to be partnering with them on these standards or do I have to fear that whoever is the whoever has their hooks deepest into the customer can box out and colonize these

additional niches? Now, that's a pretty

additional niches? Now, that's a pretty AGI pled point of view. Maybe you think that's just we're where I'm getting ahead of myself there, but what do you think?

>> Not at all. Not at all. I I actually really love what's happening. And what's

happening is the uh the silos are coming down. There is this is all great for for

down. There is this is all great for for customers for consumers right with all these open standards essentially the beneficiaries are our customers there is no lock in anymore and then that's I

think that's great that's great for competition that's great for innovation that's great for customers so any and we're seeing this play out across the board right any anywhere in AI the

differentiation is coming down that means everyone is doing more and more things to create more and more value which ultimately is great for the industry and great for consumers and customers.

So I'm loving what's happening as you called out the the walls are coming down the lock in is is no longer there and that that makes product development really important that makes the speed of execution really important and

ultimately it's all about creating more and more value.

But does that take us to a place where companies that used to be friends are trending toward frenemies? Because it

seems like there's only so many idea like it just seems so obvious to me in so many places that like a big platform like Snowflake would be like sure we could do like what Tableau does especially now that we can get so much

more stuff shipped on a quarterly basis.

I would say the the pie is growing. So I

don't think it's a fixed pie that the people are trying to protect. So the

types of things you can do is growing and that is super exciting. I also don't think that everyone can do everything.

Ultimately where each company focuses is closer to their area of expertise to to their differentiation. So I don't

their differentiation. So I don't necessarily see that everyone's going to do everything but I also do believe there is a lot of competition but there's also growing pie which is which is exciting.

>> Yeah. Okay. Hey, we'll continue our interview in a moment after a word from our sponsors.

You're a developer who wants to innovate. Instead, you're stuck fixing

innovate. Instead, you're stuck fixing bottlenecks and fighting legacy code.

MongoDB can help. It's a flexible, unified platform that's built for developers by developers. MongoDB is

acid compliant, enterprise ready with the capabilities you need to ship AI apps fast. That's why so many of the

apps fast. That's why so many of the Fortune 500 trust MongoDB with their most critical workloads. Ready to think outside rows and columns? Start building

at mongodb.com/build.

That's mongodb.com/build.

Your IT team wastess half their day on repetitive tickets, password resets, access requests, onboarding, all pulling them away from meaningful work. With

Servil, you can cut help desk tickets by more than 50%. While legacy players are bolting AI onto decades old systems, Serville allows your IT team to describe

what they need in plain English and then writes automations in seconds. As

someone who does AI consulting for a number of different companies, I've seen firsthand how painful and costly manual provisioning can be. It often takes a week or more before I can start actual

work. If only the companies I work with

work. If only the companies I work with were using Servil, I'd be productive from day one. Serville powers the fastest growing companies in the world

like Perplexity, Vicata, Merkor, and Clay. And Serville guarantees 50% help

Clay. And Serville guarantees 50% help desk automation by week four of your free pilot. So get your team out of the

free pilot. So get your team out of the help desk and back to the work they enjoy. Book your free pilot at

enjoy. Book your free pilot at servil.com/cognitive.

servil.com/cognitive.

That's sv.com/cognitive.

Let's go back more toward the technical side for a minute. We kind of went debun text to SQL. Let's do the same thing for rag. So we've got all these unstructured

rag. So we've got all these unstructured vast amounts of data out there. Y

>> they're getting loaded into platforms. They're getting metadata synthetically created by AIS coming through and just processing them sweep by sweep.

>> How well is that working? And what is actually key to making it work? We've

been through eras of chunking strategy is really important or I've done episodes on graph databases and entity recognition and figuring out various ways to traverse the entity graph that

and obviously it's it's going to be quite distinct for each enterprise you with all the different entities that they're going to have that nobody else has. What is really driving

has. What is really driving results in that rag paradigm today?

>> Yeah, at Snowflake we've been actually very fortunate. We acquired um a company

very fortunate. We acquired um a company that that that I came with uh called Neva which was a uh webscale search engine and and that

>> all right awesome. So we brought that technology into snowflake uh to to build out our search and rag solutions and uh there the basically what determines

quality is um uh the the quality of the embedding model that you're using. Uh,

of course there is uh, you know, more and more sophisticated uh, you know, chunking strategies of of uh, you know, what you're indexing and then there is kind of other layers like the uh, the hybrid search and the re-ranker that you

build on top of it and so forth.

Increasingly a a core part of it is also to be able to understand uh, you know complex documents. You know PDFs are

complex documents. You know PDFs are messy. You have images, you have tables,

messy. You have images, you have tables, uh, you have you know multiple columns in a page and so forth. So being able to handle all of this, extract information really accurately, figuring out which model uh embedding model to use, you

know, whether you should use a multimodal embedding model or a text embedding model and so forth. So all of those are incredibly important, but increasingly, you know, we're getting to a point where you can, you know,

automate many of these things and then reduce the complexity so that a lot of the um a lot of what we used to, you know, require practitioners to do can be

relatively automated at this point. and

and now you're getting to a point where more interesting opportunities get unlocked. So for a company like

unlocked. So for a company like Snowflake for instance, being able to do what we're calling um analytical um agentic document analytics is something

that is that is possible to do. So what

I mean by that is let's say that you have kind of thousands of PDFs and there is information in it. Let's say you have uh you know quarterly results over the

last 10 years. So being able to say what's the average uh uh you know revenue over the last 10 years and if that is in multiple different documents being able to extract all of that and

then do analytics on it is now possible.

So overall I think rag is both uh getting increasingly you know higher in quality uh and also simpler to build and and increasingly more and more powerful

uh to handle some of the new agentic use cases. Would it be a fair

cases. Would it be a fair distillation of what you've said there that you're trending more toward more powerful models? Like a project

that I've been involved with recently is built around understanding often scanned on like a physical scanner forms

that are associated with the sale of a car from either a dealer to a person or person to person. These things of course have to get filed at the state and reviewed and they're super messy and whatever. So working a little bit with a

whatever. So working a little bit with a company that's using AI to automate that. In that context, I've really seen

that. In that context, I've really seen a pretty substantial simplification where 18 months ago it was like you might need your specialist embedding

model here and your kind of table extractor model there and all this kind of deep specialization often not super large models but like really dialed in on these use cases. And now I would say

today Claude 45 Opus or Gemini 3 mostly just solve the problem off the shelf in terms of understanding those documents at a

higher cost certainly inference wise but definitely a lot lower cost in terms of AI engineering time. Am I right to to say you're seeing the same trend less specialized models?

>> There are different use cases. If you

are going to process, you know, in some cases hundreds of millions of documents, you're not going to use cloud to do that. Instead, you want to use a

that. Instead, you want to use a specific embedding model to to embed certain aspects. You want to extract the

certain aspects. You want to extract the information so that you could reuse it later and so forth. But if you were talking about one or two documents, of course, these large language models can handle them really right now. So I still

do believe there are different use cases and then those use cases call for different different tactics, different models >> in in terms of why you wouldn't send millions of documents through cloud. Is

it just about inference cost or is there some other >> it's it's cost and throughput?

>> Yeah, exactly. So you know how long would it take for you to process that many documents is a is a challenge. So

we we you know at Snowflake we have a u document extraction model that that we've uh built and fine-tuned. It is in multiple orders of magnitude smaller than these large language models. That

means it's substantially cheaper and much faster to go and process. Uh if the task is specific I'm going to go extract information and and and extract these specific fields. it is faster and

specific fields. it is faster and cheaper to do versus kind of using these you know very large models which are super capable but again will be limited in terms of how fast they can do this

and of course the cost is another issue.

So that's interesting we're sketching out a little bit of a paro frontier so to speak here where we have >> on the like simplest

but most expensive at inference time potentially also like rate limit issues.

We have our clouds and other kind of frontier models. You're in the middle

frontier models. You're in the middle with a snowflake specialist model that is much smaller, does just what it does, but is still like something that is

advertised off over a whole bunch of enterprise customers that you have.

Is there what are you seeing in terms of the other end of that spectrum? the is

there still value in an individual enterprise trying to create its own super specialized model for some of these tasks or is that does that curve

stop at the snowflake scale model?

>> Yeah, it's a good question. So, first of all, we partner very closely with uh all the large language model uh labs out there uh and and they have incredible uh

capable models that that we use every day. Um

day. Um there are some cases where our customers would want something uh very specific and and this is this this is a case when

a customer has large amounts of data and uh the use case is something that the model has not seen before and then they have uh you know strict either throughput requirements or cost

requirements. Those are the cases where

requirements. Those are the cases where a a custom model that is usually you know based on some of the other large language models out there makes sense.

So we work with these customers to build custom models for them. Uh but in most cases a you know well tuned rag solution text to SQL solution with the data that

they already have with a large language model that's that's the uh you know frontier model is is is is usually you know the go-to scenario.

>> I I'm halfway through doing an AMA episode. One of the questions I got was

episode. One of the questions I got was is fine-tuning really dead? What do you think? So, it sounds like you're saying

think? So, it sounds like you're saying it's not quite dead, but it seems like it's on the decline in your >> I wouldn't say it's on the decline. I I

think it it is really well suited for certain types of things. And then maybe the best example is actually what uh you know, cursor recently did, right? You

know, at their scale, it does make a lot of sense for them to have a custom model that is doing their um you know, autocomplete for instance. Um so so being able to figure out in which

situations you need a custom model versus not is something that uh that is evolving. You know starting with the

evolving. You know starting with the large language models u makes a lot of sense. And then over time as you have

sense. And then over time as you have more and more data and if you have specific needs if you have uh um either specific needs because of data or

because of cost or throughput that's when you know specialized models come into picture.

You mentioned these partnerships you have with the Frontier companies. Before

getting to that, would you like to shout out or highlight any particular open source models that are your go-tos? We

hear a lot about obviously the Chinese ecosystem is like continuing to open source a lot more than the American ecosystem at this point. I don't know if you guys feel like comfortable using

Chinese models in your stack. I get very different answers on that when I ask that question. But what are the models

that question. But what are the models that you guys go to today when you're like, "Okay, we're gonna explore some new custom direction either for all of our customers or even just for one

customer." What are the handful of

customer." What are the handful of models that you go to as starting points to begin that journey? Yeah. So for

Snowflake, we have a platform where we offer a series of models and our customers choose which model they'd like to use. And then there are certain

to use. And then there are certain products where the model is just part of the product and not necessarily a a specific choice. For the models that we

specific choice. For the models that we offer there's of course all the frontier models, open AAI models, Anthropic, Gemini as well as models from Meta, Mistral and others. Some of these models

are open source, others are proprietary.

We also have deepseek as a model that we provide for customers. And then in certain arrangements where customers are looking to build custom models, some of

them are open to use using model weights from these models from China, others aren't, but it really depends on the customer.

>> Does that break down along like industry lines or is it more of just an idiocrat idiosyncratic like gut feel on the part of the customer as to what they're comfortable with?

>> It I think it's the latter actually. It

doesn't it's not necessarily a industry specific thing like we have customers who are in technology for instance who who will not who say yes uh sometimes or

who who will absolutely not touch uh you know some of these models for other customers.

>> Do you have a sense of how much they are leaving on the table?

Are they leaving much on the table by cutting off the Chinese model option?

like models like Quen are incredibly powerful and and then then you know if if they'd like to work start with models like that and then fine-tune it you can you can get very capable models but you

also have other alternatives um so it really depends on you know the internal policies of uh of these customers to decide which route to go I'd say it's

such a competitive space that uh I don't think there is you know one model that dominates um dominates it all whether that is in proprietary world or open source world. So there are a lot of

source world. So there are a lot of choices out there.

>> Gotcha. Okay.

>> Hey, we'll continue our interview in a moment after a word from our sponsors.

>> If you're listening to this podcast, you're probably thinking seriously about where AI is headed and maybe about how you can actually contribute to making it go well. I want to tell you about an

go well. I want to tell you about an opportunity that could become a pivot point in your career and a springboard for you to make a positive difference. a

program that I've been so impressed by that I've supported it with a personal donation. I'm talking about MATS, a

donation. I'm talking about MATS, a 12-week research program that connects talented researchers with top mentors working on AI alignment, interpretability security and

governance. These are researchers at

governance. These are researchers at Enthropic, Google Deep Mind, OpenAI, the AI Security Institute, Redwood Research, Meter, the AI Futures Project, Apollo

Research, Govai, Rand, and other leading organizations. The track record here is

organizations. The track record here is remarkable. Matts has accelerated over

remarkable. Matts has accelerated over 450 researchers with 80% of alumni now working in AI safety and security. 10%

have co-founded AI safety initiatives including Apollo Research whose co-founder and CEO made the 2025 Time 100 AI list. Matt's fellows have co-authored over 120 publications with

more than 7,000 citations and helped develop major research agendas like activation engineering, developmental interpretability, and evaluating situational awareness. The program is

situational awareness. The program is fully funded. A $15,000 stipen, $12,000

fully funded. A $15,000 stipen, $12,000 compute budget, housing, catered meals, travel, and office space in Berkeley or London. Everything you need to focus

London. Everything you need to focus entirely on research for 3 months with the chance to extend up to a year.

Applications open December 16th and close January 18th. If reducing risks from advanced AI is something you care about, you should apply. For more

information, check out mattsprogram.org/tcr.

mattsprogram.org/tcr.

That's matsprogram.org/tcr

or see the link in our show notes. The

worst thing about automation is how often it breaks. You build a structured workflow, carefully map every field from step to step, and it works in testing, but when real data hits or something

unexpected happens, the whole thing fails. What started as a timesaver is

fails. What started as a timesaver is now a fire you have to put out. Tasklet

is different. It's an AI agent that runs 24/7. Just describe what you want in

24/7. Just describe what you want in plain English. Send a daily briefing,

plain English. Send a daily briefing, triage support emails, or update your CRM. And whatever it is, Tasklid figures

CRM. And whatever it is, Tasklid figures out how to make it happen. Tasklid

connects to more than 3,000 business tools out of the box, plus any API or MCP server. It can even use a computer

MCP server. It can even use a computer to handle anything that can't be done programmatically.

Unlike Chat GPT, Tasklet actually does the work for you. And unlike traditional automation software, it just works. No

flowcharts, no tedious setup, no knowledge silos where only one person understands how it works. Listen to my full interview with Tasklet founder and CEO Andrew Lee. Try Tasklet for free at

tasklet.ai and use code cogrev to get 50% off your first month of any paid plan. That's code cogrevllet.ai.

plan. That's code cogrevllet.ai.

So on these partnerships, we've got announcements recently of partnerships with Anthropic and also with Google for the Gemini models. I believe there's also one although I think it was not so recently announced with OpenAI. Uh I

didn't see I didn't catch anything with respect to XAI and Grock providing all the latest and greatest stuff to customers is at the heart of that strategy. But tell me more about kind of

strategy. But tell me more about kind of some of the nuances of the partnership.

Is there a XAI relationship? If not, why not? And does it have anything to do

not? And does it have anything to do with them putting women in bikinis all over the place? And then I definitely want to get into the how are we bringing these models to data because that sort of is a bit of a narrative violation

relative to what you typically hear is we can't use that because we'd have to send the data to them and we're not comfortable with that. So I'm very interested in unpacking like how you are reversing that and bringing the models to the data on the snowflake platform.

>> Yeah, absolutely. Actually let me start there because that's incredibly important for us. So when we started the journey 2 and a half years ago or so, we heard loud and clear that our customers do not want to move their data out of

the snowflake security boundary.

Instead, AI needs to come next to data.

And that gives them a lot of advantages.

You can just respect all of the security that that you've established. You

respect a lot of the governance on the data so that you're not replicating this data. The attack vectors shrink in terms

data. The attack vectors shrink in terms of securing all of this information. So

what we have done is thanks to our relationships we've built we are bringing we're we're bringing essentially inference to run inside the snowflake security boundary so that that's accomplished through these

partnerships through the connections as well as a lot of the legal guarantees around the data. So essentially these models become subprocessors. There is no there is no state that's saved in any of

these models. So that that's super

these models. So that that's super helpful for our customers who who are very sensitive, many of them in regulated industries. So when they're

regulated industries. So when they're using any of these models, they know that the data still stays inside the snowflake security boundary.

>> So that does that mean then that the model weights have to come inside that boundary and like how is that happening?

Because obviously the IP >> Yeah, the IP still belongs to Yes, absolutely. And and the IP is is owned

absolutely. And and the IP is is owned by the model providers. uh the

inferences run uh by you know the cloud providers in in in their stack. Uh the

difference is uh you know we have series of guarantees to ensure um to ensure data residency to ensure uh there is there is no state that that's left. So

all of those are through uh you know through the relationships and the deals that we're we're doing with these model providers as well as the cloud providers.

So the cloud providers are key in this because they are certainly able to provide the inference and they're also providing the underlying like physical infrastructure that snowflake is built

on top of and so it's because both of those things are true that we can draw the right dotted line >> security boundary >> both of these things is there more that

I should understand about this because one thing that has of course I don't know what I don't know but it seems like the more you move weights around to different clouds and stuff, the more risk you as a frontier model developer

here. I'm thinking open aanthropic

here. I'm thinking open aanthropic Google obviously runs their own clouds to a very large extent although everything's showing up everywhere. We

one of the fascinating things about this whole moment has been how many alliances or at least partnerships we've seen between big tech companies that previously were very much at odds with each other. So with all the models

each other. So with all the models showing up everywhere, I'm like, how has it been that none of these have really leaked? seems like there's so many

leaked? seems like there's so many people that work at these platform companies that if access isn't like really well figured out. I don't know.

It just seems like something would leak at some point. But we haven't really seen that. We haven't seen the weights

seen that. We haven't seen the weights of a Frontier model leak at all as far as I know. And then people will speculate. Maybe some state actor might

speculate. Maybe some state actor might have stolen them and not put not told anybody about it. But we haven't seen fundamental breakdowns. So how should we

fundamental breakdowns. So how should we understand how that is happening?

to seemingly such a high degree. What

role does trusted execution environments play? What role do other kind of

play? What role do other kind of measures play? Like how are we my

measures play? Like how are we my general view working heristic is like every everybody's hacked, everybody's pone like nothing is secure and yet at the same time we seem to not be having catastrophic leaks. So how can you help

catastrophic leaks. So how can you help a simple person like me understand how we're achieving that?

Yeah, I mean u as you called out these are very sensitive important IPs that that belong to the model providers and then secured by the cloud providers.

They have a very strict series of requirements and setup to ensure that access is limited because they are the ones that are running the inference and

setting up the environment. They've set

it up in a way that is airtight. I don't

have a lot to say beyond it. I think the they absolutely take security very seriously. We work with them. We

seriously. We work with them. We

understand how important it is. You've

talked about all the different risks that that are out there that they need to protect against. So this is something that both cloud providers and model providers as far as I can see are taking very seriously. And as we work closely

very seriously. And as we work closely with them for Snowflake, of course, security is at the heart of what we do.

So we set up our own environment in a way that is that has all of the security considerations in mind. I'll just say this is an incredibly important area and there's a lot there's definitely focus in this area across across all the

parties.

>> How much would you say of this kind of security has gone to the level of provable guarantees or cryptographically

secured as opposed to more roles and access controls and things where there's still more fundamentally like human element. I just did an episode not long

element. I just did an episode not long ago with a couple experts in formal methods, including a guy who's a VP at Amazon who's pioneered a lot of their

use of formal methods to derive a lot of these security guarantees. But I'm not clear on how much of this is resting on that kind of like we have proven that

this is secure versus how much is like we have a process that we we feel good about and we want you to trust. So this

is not my area of expertise. So I don't have a lot of depth but purely from talking to both the cloud providers and the model providers. When you start looking at what are all of the attack

vectors and whether there is what is possible it doesn't seem like this is a human factor is an issue. The way the systems are set up is inherently very

secure. That that said in in security of

secure. That that said in in security of course you can never say hey this is completely airtight and it can never be penetrated. But uh yeah as I said

penetrated. But uh yeah as I said security is taken very seriously and I don't think it's a human factor necessarily. The way the systems are set

necessarily. The way the systems are set up is such that the execution environment doesn't have access to you cannot do a lot with it other than just run inference through it. So but by by

design access to the weights are limited.

>> Yeah. Okay. Cool. Thank you. I'm always

trying to get a little bit better read on that particular corner of the world and it's not one that is as freely and openly as some of us curious minds might like. Going back to the models though

like. Going back to the models though themselves, what's your read right now on this is another thing where I think people have very different intuitions.

Are the models going to be commoditized or are they going to be sufficiently differentiated as to maintain pricing power as we continue to go into the

future? I guess there's that. There's

future? I guess there's that. There's

like how do you help customers decide which model to use for a given case? Do

you have an eval platform built in or do you help them do eval? How do you help them think about being like keeping agile so they can switch?

Obviously, new models are coming out all the time. So there could be something

the time. So there could be something better, faster or cheaper that you can upgrade to, but you have to know with some confidence that you're going to be upgrading with good reason. There's a

whole ball of wax there. Take your time in melting it. Right.

>> I Yeah, super interesting. I mean I uh as you mentioned the differences between these models uh is not large you know each model keeps getting better and and

then there is you know great healthy competition out there between the model providers which again kind of benefits uh companies like ours our customers and

and so forth. for us because uh we're providing the choice to customers. Uh

the second part of the question is also really important which is how do we help our customers choose which model is the is is the best fit for their needs. Um

there there's a couple of considerations like one is um for many customers they do not want to leave again because of data residency requirements uh if they

are for instance uh an Amazon you know shop and today open AAI is available through Azure or through directly open AI like that becomes a consideration so

you know some customers are okay with their data leaving uh that Amazon cloud uh boundary others aren't so that that's one uh you know this decision point. The

second one is of course uh from a quality perspective uh you know many customers will go run evals side by side to decide which uh which model is best best suited for their needs. You know

what's interesting is you know some of these models are cheaper faster but when you add reasoning on top of it uh the equation changes right so you know certain models are very good at certain

things again just to call out you know cloud is incredibly good at coding and continues to be a great model for that.

So we help our customers uh in assessing in which models to use for their needs.

>> Does that extend to like how much of that is a service a consultancy type relationship and how much is productized at this point or do you know productization?

>> Yeah, it is more productized than a service on on our product. You can

easily choose models. You can easily do side-by-side comparisons. You can run

side-by-side comparisons. You can run evals and for many customers actually it's not necessarily for the first reason and not all the models are available in their environment anyway because as a company

they've decided that only these models are approved or only this environment is approved for me.

How often do you see people switching?

This is even in my so I started a company which I used to be the CEO of.

I'm no longer and we're only 40 people.

We're doing like $10 million a year in revenue. We're not an enterprise. We can

revenue. We're not an enterprise. We can

fly a little faster and looser than and we're also not in a regulated industry.

We basically do video content creation for local and increasingly like mid-size businesses.

>> And I feel like we should be changing models more often than we do. Honestly,

I feel like the leaprog effect is it's just happening so often. And if I were to grade our own performance, I would be like, eh, b like we're definitely better than most, but I wish we were even a

little more on top of eking out the the latest and greatest performance from the latest and greatest models, but it's hard. It is hard to resolve sometimes.

hard. It is hard to resolve sometimes.

You've got even just human interrator disagreement, which is tough to overcome. And how does that play out at

overcome. And how does that play out at the larger scale? Do you see people like, oh, we got a new cloud when cloud 4 6 hits? It's like how many people move to it in a week, in a month, in a quarter.

>> Yeah, I think it really depends on the use case. In most cases, we don't really

use case. In most cases, we don't really see a lot of uh switching happening. uh

because you know the prompts get optimized for a certain model and uh you get high quality because you've optimized it for a certain model and it's not as easy uh without further

optimization to switch and because the deltas between the models aren't much and uh they keep improving u you know on a regular basis the need for switching is also not that much um you know as as

you were describing this I was actually thinking about you know Google versus Bing you know at some point Bing you know got to a good enough quality uh but you know there was the habit of continuing to use uh you know Google the

familiarity of the interface and so forth that switching wasn't as necessary I don't think we're there yet necessarily for models there's still um a lot happening a lot of innovation happening and also for certain use cases

like if this is a one-off I'm going to go run something and then do sideby-side comparison like then you go pick the model that that works best for you but if you've already been investing in um

in an application [clears throat] and you have thousands of lines of um you systems and prompts that you've built, then there's a cost to switching in in

in which case the gains have to be large enough to justify that cost.

>> So, I'm no Ben Thompson, but it seems like from your perspective, you would want to commoditize your compliments and would want to do everything you could to reduce those switching costs, right?

Like one of the virtues of being on the Snowflake platform would be you've got all the things but not only ideally you have all the things but also you can the interface is presumably a lot more unified than it would be if you were

going directly to the model providers and I imagine there's a bunch of different things you could do over time.

You've got like things like DSP out there that you can say sure this is my one prompt with this model but maybe if I throw it into CS PI I can auto evolve my prompt to to be more optimized for some other model what have you. Is this

like a goal? Do you want to would you think of it as like a success metric that you like would help people be very fluid in switching from model to model as you go into the future?

>> I mean that's not how we think about it.

For us it it really does boil down to how do we bring the most value for customers quickly. Um so it choice is an

customers quickly. Um so it choice is an important uh factor there. So we'd like to offer a choice. Uh and and customers make uh model choices for a variety of reasons. As I said, you know, some of

reasons. As I said, you know, some of them have only approved this certain model. They have their own AI governance

model. They have their own AI governance boards where they decide which uh which model to use and so forth. Um but but for us, we start with the data at the core. So ultimately, anything that you

core. So ultimately, anything that you do is as good as the data that you provide to it. So a lot of the optimizations for us are can we do a phenomenal job at the uh retrieval layer

and then can we make sure that all of these models uh you know are optimized to the fullest extent so that any customer that's choosing one or the other for for the variety of reasons I

called um get the best uh quality data agent if you will uh that they're building with us.

>> Okay, that's really interesting. What is

what do you think that implies for the competitive dynamics between model providers? Like

one takeaway you might have from that is whoever has the best model at any given time like wins. Of course there are these other constraints but leaving those aside for the moment. If I'm a customer that has no binding constraints

and I can pick whatever frontier model I want. It seems like whoever has the best

want. It seems like whoever has the best model at any given moment in time wins that business and then actually stands to keep that business. Even if that business that might not be that whole enterprises business, but that

particular use case you're it's you're saying is stickier than you might think.

Switching costs are higher than they intuitively seem. And so

intuitively seem. And so having the best performance at the time it's initially evaluated is actually like pretty important. So I think the model quality is incredibly important

but increasingly we're moving up the stack so that the product also becomes incredibly important. So when you look

incredibly important. So when you look at from a consumer perspective chat GPT as a product starts having its own kind of stickiness because you start using it and you get accustomed to using it.

Similarly on on coding side cloud code has its own again benefits. You you'll

start writing your instructions your prompts to to optimize for that workflow. So I think we're just moving

workflow. So I think we're just moving up the stack. The model quality is absolutely central. But as the quality

absolutely central. But as the quality kind of keeps up across model providers, the next level of differentiation happens at the application layer.

>> So what that's a perfect transition to talk about agents and what you guys are doing with agents. The way I structure my own thinking about agents is on a spectrum from on the one end your cla

code style choose your own adventure. I

just give you the goal essentially and you the agent break it down and search around, GP around and figure out how to get there.

>> And then on the other extreme is like potentially a totally linear structured workflow where we're going to run a series of prompts one after another.

>> The cloud code is undeniably awesome interface.

>> But I often feel like people are a little bit too drawn to that and I sometimes say that's a don't try this at home sort of project. like by all means go use claude code but don't think like

at your business you should be spinning up a cloud code choose your own adventure thing probably for most cases I advise people like even still today more structured is probably going to get

you more of what you want faster in a way that everybody feels good about at the end of the project what you know what distribution are you seeing across that spectrum >> yeah I think it really depends on the

persona who are using these tools so I'm huge fans is a huge fan of clot code and the coding assistance make a big difference unlock great capabilities and

clearly very helpful for AI developers builders if you are a business user who's just asking questions like what what was my the usage of this product over the last week as a product manager

for instance I want a structured way to do this I want an agent that is already optimized for that use case that has access to the underlying data I do not want a cloud code interface for this I want something that I know will be high

quality and that's optimized. So that

that's kind of how I think about it.

Really depends on the persona that you're building for.

>> So for the talk to data product surfaces that you guys expose, how where would you say you you tend to fall on that spectrum? Is it a you're going to use these tools in this order

or is the model kind of choosing which tools to use at any given time?

>> So we have a product that we built for business users. So this is Snowflake

business users. So this is Snowflake intelligence where you can build a series of assistants. For instance, for the whole company, we built a sales assistant and we've deployed to 5,000

sellers. That product is

sellers. That product is is a think about it as like a chat interface on top of all of the company's data so that you can ask questions like

what are my upcoming renewals, how is my book of business doing and so forth and you can get answers for it. So for that clearly you want a highly optimized set

of agents for those set of use cases and and these are business users using it and they they need to trust the answers that they're getting. Then we have a set of products that we're building for for

data engineers and for analysts to to build data pipelines to analyze data that is more coding assistant if you will. So we have our own coding agent

will. So we have our own coding agent that's integrated in that platform where they're just either analyzing data or writing code. Um so that is that of

writing code. Um so that is that of course is a lot more flexible and it's also not tuned for very specific set of use cases. How do you think about the

use cases. How do you think about the question of like one big agent that might be long running versus

the other kind of big pattern is your sort of initial agent that then routes tasks to sub agents. Back when

OpenAI came out with their agents SDK, they had this notion of the handoff as a really central idea. I was never quite clear on were they doing that because

they thought that was the best way to maximize performance or was it more a nod to we think at these enterprises that are going to use this thing there's going to be different

teams responsible for certain areas and we want to be able to modularize the work for human reasons as opposed to like for AI performance reasons with Claude on the other hand and we have a sponsor of the podcast called Tasklet

which is a maximalist when it comes to just let Claude cook basically is their philosophy. Give it everything it needs.

philosophy. Give it everything it needs.

Let it make all the choices. Let it run for as long as it can run. Give it

feedback. But like it's one long one agent that kind of does it all in in one long session. Of course, I'm sure you

long session. Of course, I'm sure you could say there's different use cases deserve different paradigms, but what do you see working the most in practice today?

>> Yeah, even the cloud codec case, you have uh you have skills that uh that are being developed, right? So you you're you're still modularizing uh the different kinds of things you want cloud to do and then giving instructions for

cloud to do go do these things. I think

the way you called out is is is is what I'm seeing which is you know in in in especially large enterprises you have different teams building different agents. You also have different agent

agents. You also have different agent platforms uh that are being used. Uh so

for instance, if I'm using, you know, Salesforce to manage all of my CRM, maybe I'm going to go build uh my sales related uh experiences with an agent

there, but I still want that agent to talk to this other agent I'm building for something else. So being able to do that uh agent handoff and coordination

is uh emerging. I wouldn't say this is necessarily super top of mind for for everyone. I think still customers are

everyone. I think still customers are focused on let me get this one agent right and and and working well before I start thinking about multiple agents uh kind of coordinating with one another.

Uh but that's starting to become increasingly important uh for customers.

One of the biggest considerations is um you know they do not want to be locked in to a certain platform. So they still want to be able to you know make sure that the uh you know again open standards are supported so that agents

can talk to one another, agents can use the tools that you built for one agent uh by another agent and so forth. So you

know MCP A2A these are important protocols uh that our customers expect to be supported.

>> I was just going to ask about A2A. Are

you seeing traction with that?

>> Uh still early we don't yet support it.

Uh we're starting to hear our our customers don't necessarily ask for A2A specifically. They do ask for ensuring

specifically. They do ask for ensuring that uh you know some kind of agent to agent communication is is possible to do.

>> Is there any standard or protocol or platform that is bridging those the Salesforce continent and the the various other continents of agents >> to today we either see a bit of a hack

where these different solutions are used as tools through MCP. So still the orchestrator uses them as tools and then manages them from a agent handoff perspective other than I haven't really seen anything else.

>> Yeah, that's interesting. And one of the things I find very funny about this whole thing is that just it seems to me like a fundamental property of intelligence is that you can everything one this is a

an overstatement I don't mean it literally but one one of my refrains is like everything is isomeorphic to everything else meaning you can always squish and rearrange and play hide the

intelligence and you can have a smart MCP that's actually an agent and how you actually classify these things seems to be

much more of a choice and much less a requirement imposed on us by nature because the just the nature of intelligence itself is so flexible, funible subdividable whatever.

>> Exactly. No, couldn't agree more.

>> One of the big things, huge theme, right, of the communications that I've seen from Snowflake in preparing for this is the importance of trust. So I'd

love to hear your thoughts on what are the levels of what are the dimensions and what are the levels that we have to hit in order for an enterprise to trust

an AI process.

>> Yeah, just to uh reiterate like for us trust is incredibly important. You know if I were to call out you know two important tenants. One is super e ease of use.

tenants. One is super e ease of use.

how easy it is to build out these solutions and and to use them and and of course trust is at the core of everything and trust it spans multiple different dimensions right you know

trusting from a security perspective then from a governance perspective then you have the quality layer on top of it and then there is uh kind of evaluations and then monitoring and so forth so it's

a full stack so the way we think about this is by running AI next to data a lot of the core governance that is put on the data is by design respected in our

system. So what that means is let's say

system. So what that means is let's say you have you know sensitive data that's only visible by the HR team. If you go build an agent uh you know the person who asks the question can only get the

answer that they're eligible to see and nothing else. This is super obvious and

nothing else. This is super obvious and important, but because we have uh kind of these types of very granular access controls from the ground up as part of the core data platform, building agents

that respect that uh becomes much easier to do. Then you have uh you know

to do. Then you have uh you know governance at various layers. Um and and of course uh next level is evaluations and and the of these uh so a lot of the

trust is in are you able to build highquality u retrieval to for of context to pass to the agent? Uh is the agent orchestrator doing a great job figuring out which tool to use, which

trajectory to use to answer the question. Um so evaluation is a core

question. Um so evaluation is a core part of the platform and and then ongoing monitoring getting feedback and then that that cycle of improving improving the quality from a user

perspective. The way that that trust

perspective. The way that that trust manifests is um when a user asks a question uh you know we have uh UI elements that says hey you know this

question uh has an answer that was verified by uh by an owner. So again

bringing that uh trust element into the user experience is another tenant to um uh to our philosophy.

>> So did I catch correctly at the data governance level that the shorthand rule is like the agent can only access the

same data that the user can access. And

so in theory that could mean like multiple users could come to the same agent and have different experiences because the agent has different data access based on the user that's using it

at the time.

>> Exactly. And and this this is exactly what our customers are asking for and then that that's relatively easy to build uh on our platform. So for

instance the example I gave with our uh own kind of sales assistant. If a

salesperson comes in and says, you know, what is my book of business? Summarize

it. You should get an answer that is only, you know, your list of customers assigned to you versus another salesperson.

As an HR person, if if I ask uh or as a manager, for instance, using an HR bot, if I say, you know, what's the uh I don't know salary of this person, I can

only I should only be able to see the salary of the person that I have access to seeing versus somebody else. and

underlying it is the same agent uh and it's the uh access controls uh that govern what I'm able to see.

>> Yeah. Interesting. On the sort of performance reliability side, my experience has often been and sometimes it's for good reason.

Certainly in like the self-driving car realm, it there's a certain logic to saying we don't just want these things to be like roughly human level. We want

them to be like clearly a step up before we're going to adopt them societywide.

Good news. It seems like we're getting there. What do you What do people

there. What do you What do people have in mind as the intuitive standard of performance? Is it like they want

of performance? Is it like they want these agents to be perfect? Is it that they want them to be like at the level of the human that used to do the job? Is

it some is there some like heristic in the middle that you think people often land on?

>> Yeah. Yeah. Super interesting concept, right? The more natural the interface is

right? The more natural the interface is like the more humanlike intelligence we expect intuitively. If I'm talking to

expect intuitively. If I'm talking to the agent versus typing, I think talking has much higher expectations. For

instance, versus I'm just typing. I know

I'm typing to a computer. So the

expectations become a little less less high. I think I I I don't think that

high. I think I I I don't think that adoption of of of this technology requires humanlike intelligence because

even for the specific things that these models and these applications do well that is such high value that we're seeing huge adop as you all know we're

seeing huge adoption of AI already and and it it keeps getting better and it and it keeps getting better at a super rapid phase rapid pace. Yeah, I'm very excited about where the technology is.

>> Before going into your expectations for the year ahead, what are you seeing in terms of guard rails? Obviously, one big pattern that I think is like very natural to you guys is sourcing answers

back to the document or the the sort of authoritative place from which it came.

Beyond that though, we've got this a whole constellation of different patterns, right? in terms of you can filter inputs

right? in terms of you can filter inputs for appropriateness, you can filter outputs, you can log things and post-process logs, you can AWS has a I think a really interesting new service

called automated reasoning checks where you can put a policy in. They convert

with the language model your natural language policy into a set of rules and values and then they do like literal

formal methods to ensure that at runtime like the agent or whatever the system whatever it gave you back that it actually passes those like formal reasoning checks that were derived originally from a natural language

policy. That's like pretty interesting

policy. That's like pretty interesting and pretty cutting edge from what I've seen. What? But I think in most places I

seen. What? But I think in most places I my sense is like the frontier model companies are doing a ton of this stuff.

Anthropic has pushed this to probably farther than anyone when it comes to preventing you from using cloud to do certain things in the bio sphere. But

are people at the enterprise level actually doing much of it or are they just saying this thing seems to work.

We've got an eval set. It passes and we'll go with that. I I think the sophistication is increasing and usually companies start with products that are more internally focused. While it's

important, the bar is a little lower than something that's externally focused.

At Snowflake, we offer products to check for guardrails. They're doing things

for guardrails. They're doing things like checking for hate speech, violence, and other violations. You can detect them and flag them and not have the model respond. But we of course also

model respond. But we of course also benefit from all of the great work that as you called out companies like Anthropic do on their own models as a baseline. Some the other thing that is

baseline. Some the other thing that is also super interesting is as the models keep getting better their adherence to instructions of course keeps getting

better. So some of these also get

better. So some of these also get codified as as instructions to the agent. So, not only do this and that,

agent. So, not only do this and that, but also here's the policy that you need to to comply with. And that that tends to work quite well as well.

Have you seen anything in the interpretability realm being used for practical guardrail monitoring kind of purposes at

in in enterprises so far in we're seeing um evaluations become really important uh right so a lot of uh

you know what companies tend to do is they'll go create their own eval sets uh but also use LLMs as judges across various different dimensions to go score

uh what's happening and then continue to monitor it on an ongoing basis and uh as agents become more and more complex you know it's a it's a pretty new area to

understand is the agent you know taking the right route should I be optimizing it understanding where things go wrong becomes uh really interesting so from a so that that that's what I see not

necessarily in the core consumer experience but in the uh kind of developer experience where you're seeing you know what the model is doing, what the agent is doing through evaluations and monitoring.

>> Yeah. Okay. So, you mentioned being excited about the year ahead. You also

mentioned voice experiences. It seems

like we're at a moment like literally right now where I don't know if people just had extra time over the holiday break or whatever to get into cloud code for the first time for many people, but

it seems like the discourse has really shifted and expectations have really shifted in just the last 30 days. People

have said, "Oh my god, like the coding experience now is like it's not just vibe coding and eventually hitting a wall and giving up. Like you can actually really make this work." And

then the next big thing that people are saying over and over again is the same thing that's happened to coding over the last however many months is coming to a great many domains of knowledge work

over the next year. So do you buy that hype? And what do you think that looks

hype? And what do you think that looks like? Are we all going to be agent

like? Are we all going to be agent managers or are we all going to be like talking verbally to agents while getting lots more exercise than we used to? What

is the 2026 plus vision for success?

>> Yeah, I'd love to see the world where I don't need to do anything and I can just go get more exercise. I I do actually see the opposite. We're able to do a lot more and we end up doing a lot more.

Especially in AI where everyone is sprinting, there is more work and more productivity out there. So what I'm seeing is agents are absolutely getting more and more capable. Coding agents I

think as you called out have passed this threshold where they're a lot more because they're a lot more capable a lot more people are using them. I think it changes how products are developed. It

changes jobs like product management for instance. I've been in product

instance. I've been in product management for 20 years and like the way we build products has to change given the coding the coding assistance. How

you deploy quickly, how you test things quickly is changing because of how capable these coding assistants are. And

that is a combination of different things, right? One is the agent can do a

things, right? One is the agent can do a lot of things, can do coding well, but also the reasoning capability of the agent is is increasing. The tool use becomes incredibly powerful. So you can

apply that to other domains. If you have like now figuring out which tool to use and using it effectively and then reasoning and figuring out the next steps is very like that allows you to

build very capable agents across the board. I don't know this is not a 2026

board. I don't know this is not a 2026 projection or anything but I'm absolutely seeing clearly increasing capabilities with agents and also increasing use of them for production

work.

Do you have a sense of how the progress in AI coding assistance or agents has changed how work is happening

at Snowflake? Are you like instrumenting

at Snowflake? Are you like instrumenting that or measuring? Obviously lines of code would be like too primitive, but features shipped or burndown points per

cycle. Is there a way that you can begin

cycle. Is there a way that you can begin to quantify the impact? We're it is actually there's the impact piece but there's also the um philosophy that is

changing you know how we build products is changing um so that requires uh you know change of behavior you know usually my go-to is I have uh like let's say

there's this feature to be built I'll think about a UI I'll go build this UI and and go make it happen um whereas with a coding assistant if my users are also in living in coding assistance

maybe that's as simple as let me just build a skill for this thing and then quickly test it out. I can just write the skill in a day, put it out in front of my customers uh and then have them use it, give them feedback and only when I know exactly uh you know what the

shape of the product is, I I I can go and kind of solidify it in uh in more of consumer experience. So I think again

consumer experience. So I think again product management is changing, product building philosophy is changing because of these coding assistants.

>> Yeah, the working prototype is the coin of the realm these days for sure.

>> Yeah. Yeah. One of the big predictions that I have heard a lot recently, sometimes with a remarkable level of specificity, including from some anthropic people, is that we should see

the first drop in knowledge worker products offered this year, specifically folks have said Q2 of this year. And

what that means to them is basically a new employee that's and ultimately at heart is an AI, but it will have a very

similar surface to a remote worker on your team. It'll have a name. It'll have

your team. It'll have a name. It'll have

all the same accounts or at least you'll be able to give it all the same accounts that you can give to a human employee which means they'll be on Slack and they'll be accessible via email and they'll be all over the place and can

probably join calls. And the expectation is that this will be good enough in Q2 of this year that people will start to get value from it and this will be a new product category. I guess first do you

product category. I guess first do you buy that that can happen that soon? And

second, how many of your customers do you think will be eager to try something like that out when it drops? I

>> I do see that it's a natural progression. uh you know today uh the

progression. uh you know today uh the agents that are being built are either automating certain processes uh you know

from a productivity perspective or they are more like uh you know co-pilots that I can ask questions to and and get get responses uh versus kind of these autonomous uh

intelligence entities if you will when exactly that will happen I think really depends on how scoped can you get them to I don't think we're at a point there

uh you can just create uh another colleague uh that that can just do anything and everything but

if you can just very easily scope uh the task then I think that uh like that absolutely is possible yeah so I don't know like that's u I do see it happening

you know whether this is Q2 or not um not so sure yet and again >> find out >> yeah but plus uh you know as a data platform uh you company. I I will call

out the importance of ultimately all of these uh capabilities come down to you know for any given company the differentiation is their data uh and

then the access to that data being able to figure out and retrieve the right types of data to be able to answer that question and then increasingly the tools that are given to these agents to take action. I I think that changes

action. I I think that changes industries, right? That changes how we

industries, right? That changes how we think of data. That changes how we think of making AI, making the data AI ready as well as making the tools AI ready. Uh

so that more and more kind of capable agents can be built.

>> Do you see changes to data itself? I

guess one that we've talked about already is just retroactively going back and applying structured unstructured data, creating metadata, so on and so forth. In the wild, one big change that

forth. In the wild, one big change that we're seeing to data is like the web itself is increasingly comprised of AI

generated data. And so that's a weird

generated data. And so that's a weird feedback [clears throat] loop that we accidentally created. Are there any

accidentally created. Are there any other perhaps surprising patterns in data in within enterprises that you're

seeing as a result of AI coming onto the scene? or is it maybe just still too

scene? or is it maybe just still too early for something like that to be?

>> I am seeing two things. One is I'm seeing

two things. One is I'm seeing the access getting a lot easier. So that

democratization of access to data and access to insight is a big shift. The

other thing that I'm seeing is the value that our customers get from data is increasing because you're able to very easily glean those insights by just describing what you want in natural language and then getting it. So the

value you get from data is increasing.

How that just opens up new and new opportunities. You start using the data

opportunities. You start using the data in ways that you haven't thought of before. One interesting study, one of

before. One interesting study, one of our customers as S&P, they analyze the earnings calls to understand when CEOs

are responding to analyst questions in either directly or indirectly or if a question was already answered in opening remarks or not and using that as alpha

to determine which stock to buy. So

stuff like that becomes very easy to build and then new use cases open up.

>> Yeah.

That's an interesting metric. I've seen

a bunch of stuff recently over the last even just the last week. It was really the Venezuela moment where all of a sudden people were like bragging about how they had created these AIs that monitor the prediction market platforms

and were looking for early signal and trying to capitalize on that. That is

going to be a really interesting phenomenon. How about in terms of just

phenomenon. How about in terms of just let me go slightly a different direction then I'll do maybe a little lightning round to close us out.

>> I have this sense that right now we're in this kind of expansionary phase. I'm

no astronomer, but my experience with platforms in the past and definitely experiences with the meta platform, formerly known as Facebook, where it was like they came on the scene, they opened

up a ton of stuff. Everybody could tap into all these data and social connections and there was for a moment there was like an incredible flourishing of a ton of different ideas. And then

after that supernova came the black hole and it was like actually we're going to close all this stuff back down. And a

lot of that value that entrepreneurs created on the edges experimenting with different ideas, the things that really mattered mostly ended up getting sucked back into the platform and there wasn't

nearly as much value created on the margins as it seemed like there was going to be. And I I might be just over indexing on this experience of having

lived through this pattern once before.

But I feel like the AI moment is set up for that to happen to a lot of people again where for example just this week, right, Chad GPT launches medical

version, which is great. I think that's going to be awesome for a lot ton of people. And the fact that I can now just

people. And the fact that I can now just connect my chat GPT into an EMR instead of having to go laboriously copy and paste or find some other third party thing to do that kind of connector work for me like the consumer surplus of that

is going to be amazing. That's like my strongest belief is we will see high consumer surplus >> but for the businesses it seems like it does create a very tricky balance where you're like I want to go do a bunch of

cool stuff but how do I know which of these things will be durable over time?

How do I know I'm not just doing like R&D for the next generation of the mega platforms that are ultimately going to eat my lunch? How do you guys think about like where you want to place your

bets, but what is going to get absorbed into the models versus what you know only you can do over a longer period of time.

>> Yeah, I think we are in a fortunate place because ultimately data is an incredibly important asset for all companies. that is what kind of defines and differentiates them and that

is not getting commoditized anytime soon. So as a data platform we sit in in

soon. So as a data platform we sit in in in that layer in between the application and the model if you will. So the way we think about this is how do we help our customers build very high quality

products that are catered towards their use cases which is all powered by their data and whether any of that can be

subsumed by these other platforms. I am sure the shape of shapes of products will continue to evolve. I

think we're still in the very early innings of huge transformation. But I

also believe intelligence is a I it like once these models are out there and there's enough competition and we're seeing enough competition like the dynamics seem to be playing in such a

way that is all pro customers and consumers versus these mega platforms. So I do believe the competition will keep things in check and the opportunity is so massive that that growing pie will

also create lots of new opportunities.

Obviously, people have to have some place to keep their data. One,

tell me what's wrong with this theory.

If I were to be a skeptic or if I took the perspective of the snowflake bearer for a second, re a recent experience I had was my company's been a customer of intercom for a number of years

>> and [clears throat] I was trying to do just some basic analysis of recent tickets. So, I and they didn't have the dashboard to do

what I wanted to do. So I went to their docs and the docs were 100 pages of docs, right? It's fullfeatured platform

docs, right? It's fullfeatured platform at this point. So a lot of docs. So the

first thing I did was told a web agent, hey, go compile all these docs. And it

literally went page by page, copied them all into a Google doc, like in a browser. I ended up with some 600 pages

browser. I ended up with some 600 pages of text. Then I took that to Gemini and

of text. Then I took that to Gemini and said, "Okay, hey, there's a lot of repetition in here, but can you streamline that down to what I really need to know?" And so it did that fit in the even 600 pages whatever it like fit

in the million tokens. So now I've got my consolidated single view of the docs.

Then I go to a coding agent. I say here what here's what I really want to do is export all my data from intercom. And

that also ended up being just one prompt to work. And then it exported all my

to work. And then it exported all my data from intercom and was able to do the analysis I wanted to do. But then

the sort of Eureka moment was like wow it's never been easier to unplug from intercom if I want to take my data somewhere else. They didn't really

somewhere else. They didn't really anticipate it being that easy when they created all these APIs. So

what what prevents the data platforms of today from running into trouble there in the past presumably like somebody were to say hey I'm not happy with you or I want a better price whatever you had some

leverage of just what are you going to do you going to pull out all your data and I'm not saying that would be easy to say cuz I know you guys handle vastly more data than I have in intercom but it does also strike me that like it's become a lot easier to move things around it's become a lot easier to

understand what it is especially once you've gone and done all this metadata layering on. So what are the modes? What

layering on. So what are the modes? What

are the sticking points? Has it changed or or will it change?

>> It is going the direction that you're calling out which is I think great for again customers and consumers. So today

um Snowflake supports uh you know open file formats for storing data that we support uh iceberg which is a uh you know open file format. What that means is you do not have to have your data

locked in somewhere. you can just put it in a uh you know in a you can keep it uh in a in a managed uh place that's managed by you or by us and then you can

use Snowflake as a um as an engine to go process your data. So we are absolutely uh embracing and and supporting the ability uh for for our customers to uh

to use these open file formats to be to to not necessarily feel like they're locked into one platform. And I think that's that's great. That's great for customers. uh that's great for

customers. uh that's great for innovation. Ultimately, customers will

innovation. Ultimately, customers will end up using the product that is going to give them the uh you know best performance uh best best cost uh for the things that they'd like to do and you

know we're absolutely embracing that.

>> Does that translate to increased pressure on you and your team? Like it

would seem like maybe one way to think about that would be like in the past if somebody wasn't happy for a year maybe they would start to think about a switch where now it might be if they're not happy for a quarter. Does it shrink the

timeline where you have to deliver?

>> I I love it. It's it's great for product teams right ultimately you know we're all driven by you know creating value for customers building great products and then we want to do that as fast as

possible. uh you know and and

possible. uh you know and and competition allows uh it's a great uh incentive in the system to to go you know keep things in check uh and and

then uh you know have you deliver so I don't think things change from uh for my team we we already do feel pressure not necessarily because of competition because of the opportunity the

opportunity is massive and uh the it's also there's never been a great time to be a product manager right so you're you're easily able to build awesome products very very quickly and then

you're sprinting. So it's incredibly

you're sprinting. So it's incredibly satisfying uh to to build these great products and then you also reap the benefits by seeing how these are getting used in the market. So I love the competition. I also love the pace of

competition. I also love the pace of innovation in the industry.

>> Does that lead you to a point of view big picture on this is a classic question and again it's striking to me how very informed and technically sophisticated people have very different

answers. Where does the value how do you

answers. Where does the value how do you think about the breakdown of where value occurs? Obviously, we've got

occurs? Obviously, we've got infrastructure whether that's like chip creators or owners models application layer on top. If you

had to assign those three layers relative value capture from the sort of AI opportunity, how do you think that breaks down?

>> Yeah, I I think may maybe the way I think about it is the uh the middle will erode and uh and and the uh the sides will continue expanding. So so far we've

been seeing a lot of value acrewing to uh you know chip makers uh Nvidia as well as the model providers um uh and uh and then increasingly application

developers who are able to go build kind of very quickly you know unique businesses uh on top of these capabilities. Uh cursor comes to mind as

capabilities. Uh cursor comes to mind as an example. So so I I absolutely do see

an example. So so I I absolutely do see the value continuing to acrew at the um infrastructure layer as well as at the application layer. And uh you know

application layer. And uh you know traditionally there's always been this kind of middle layer that's that's kind of facilitating and and uh connecting those two things

because of the capabilities of uh of these models that middle layer may not be as valuable uh or as important anymore.

>> And that middle layer is the models. Is

that right? Uh no no no I I do believe it's um so so the the middle layer is is um all the companies that are kind of creating custom business logic for

certain applications right so that business logic as you called out for instance if you want to just uh you know build your own extractor you can just vibe code it over a weekend and and go do it versus a company that goes and

builds it for you. So so that the that layer isn't as as important. So um

models are in my opinion will continue to acrue a ton of value.

>> So it yeah okay so to try to play that back to you it sounds like you think all three of the layers that I described will do fine but at the application >> traditional businesses will change.

Yeah, >> you're going more horizontal platform and less you like relatively less excited about vertical because the

fact I mean so many SAS applications like essentially exist to encode business logic or best practices or whatever and we just probably don't need

dedicated teams building out those kinds of things when we can just have agents kind of do it on the fly as as needed.

>> That's what I'm guessing over time.

Yeah.

>> Yeah. Okay, cool. One other big kind of question that I've asked a lot of people a lot of times and I think you're the perfect person to touch on it. So you of course you know that data bricks

acquired this company called Mosaic ML not too long ago maybe two years ago now and what Mosaic was doing I thought was really interesting which was starting with open source models working with

particular customers to do continued pre-training on data sets which I assume were very often internal data sets like the sort

of data sets that might sit in a snowflake. I was really surprised. I

snowflake. I was really surprised. I

spoke to Ali Gatsi, the CEO of Data Bricks at a event not too long ago and he said, "We killed that product." So

they basically turned Mosaic into a in-house research unit, but that product of offering this continued pre-training to try to create a model that like really knows your business inside and

out. Basically, they don't offer it

out. Basically, they don't offer it anymore. I was very surprised by this

anymore. I was very surprised by this because I think man if I had a if I'm GE or if I'm 3M or any number ofundred-year-old millions of employees over the

generations companies that have this incredible history and so much data that's acred that nobody really understands at the company these days.

If I could have a model that could have similar command of that information that doesn't that only exists in my company and just nobody else outside has

ever had access to as the foundation models today generally have world knowledge. I would think that would be

knowledge. I would think that would be insanely valuable for a lot of enterprises and yet we don't seem to be

seeing to my knowledge many instances of whatever 3M GPT or GEGPT fiser GPT like why don't we see that

>> do you have a point of view >> I do um so this is kind of similar to how up until recently when you'd ask a question on chat GPT you'll say hey you

know my uh you information cut off is whatever you know a year ago and I can only answer questions up to that point and then uh web search as a tool came in and now uh you know all of these

platforms would use web search to give you the most up-to-date information so that their world information can be it is more about the intelligence to figure

out when to use the tool to retrieve the information and then make sense of it and then give it back to you versus having been trained pre-trained with all that information up front to to me that

that that is that that that pattern is exactly what's playing out right so in the enterprise world you have a lot of information uh and then your text to SQL

and and rag solutions can bring that information in for the uh for the agent for the platform to reason with uh and then give you uh give you information.

The nice thing about that is it is substantially cheaper. uh the model

substantially cheaper. uh the model keeps getting better as the underlying you know premier premier model keeps uh improving and uh it's also relatively

easily tunable. You can you can update

easily tunable. You can you can update it uh you can change things and so forth. So that means for me majority of

forth. So that means for me majority of businesses uh would continue to you know benefit from this architecture rather than codifying all of that information

in the weights of the model. um they'll

just uh use the information uh and then use tools to retrieve parts of the information that are relevant. The

exception to that is what we discussed earlier which is if there are certain tasks that require either high throughput low cost uh if you have a lot

of data uh and uh in an area that the model has not seen before then it might make sense to go create custom models for those uh specific tasks. So I do

believe there's going to be increasingly uh need for kind of task specific small models in large corporations or you know when you have you have that need uh but

still the majority of the uh use cases will be more retrieval oriented.

So I think that's a great first pass answer. If I

answer. If I think though even just about my own ability to search through my own stuff, right, my own Gmail, my own Google Docs, one of the intuitions I have pretty

strongly is if I were to give you full access to my Gmail and give you full access to my Google Docs, you couldn't search through it nearly as well as I can. And that's despite the fact that

can. And that's despite the fact that you're clearly smarter than me. So I'm

like there seems to be something about the fact that I have had this like free training on this corpus that allows me to even just search through it a lot

better because if nothing else like one of the intuitions is I know when I found what I'm looking for. Right? You might

not know you could search you could do a 100 searches in my Google Drive and never be quite confident you got the absolute best document for whatever the question is. Whereas if it's my Google

question is. Whereas if it's my Google Drive and I created all those documents when I hit that document that is the one that's yes this is the one now I yes I remember this now this is the one. So I

have that sort of confidence that I've got to the answer if nothing else I think which is strikes me as like really hard and I've seen this when I try to give my own just whatever give cloud access to search my drive like it also struggles in that way. It doesn't know

how many times to search or if it's found the right thing or sometimes it's satisfied too easily whatever. So I

still feel like there's something there where you could expect that a model that really had a sort of more in the weights

familiarity with the data could do a better job of navigating it. And then

maybe it just comes down to upgrade cycles are terrible for this kind of thing. And yeah, as you said like you

thing. And yeah, as you said like you want to keep taking advantage of better and better models. This potentially

workshes obviously influenced the discourse recently with kind of focus on continual learning. So maybe you need

continual learning. So maybe you need either a new architecture that's more suited to that, you know, some sort of new training paradigm that would be more suited to it. I guess maybe one way to phrase this is if that were to flip, if

you were if you imagine a world a year from now where it's no longer the case that the best approach is take the pick the best models and like tune leave

them as they are, but tune them through the search as you described. and it

instead becomes one of these things where they actually do have this like deeper familiarity with like all the enterprises data. What do you think

enterprises data. What do you think would what would have changed to flip us from one paradigm to the other?

>> I was trying to kind of think about how humans do this. We'd go into a environment that we don't know and then we'll go do a bunch of searches and then we'll read to create to create some kind

of knowledge and then as you build out that knowledge that there's intuition that comes with it. So you don't need to keep referring back to it and then somehow that turns into intuition.

Right? Right now I think what these models are doing is the first part. I'll

go pick the information as the context windows of these models also keep getting better. I can stuff more and

getting better. I can stuff more and more information in these models and then get an answer. What intuition is not really understood. So I don't really know how to how that changes the

dynamic. What what would change if if a

dynamic. What what would change if if a model is trained with your data? you

clearly need much less data to steer it to a certain direction. You'll have much more consistency in in in the responses.

It still feels I don't think you ever get away from feeding it information, up-to-date information and so forth. But

what I would imagine happen is first of all that model that you want to do certain tasks doesn't have to write a poem in French. So you'll benefit from using the weights more efficiently for

the task that you want to do. And

therefore perhaps again you may not need as large a models so you get benefits from more optimizations to to reduce the cost, increase the speed and so forth.

>> Yeah, I think that many small models paradigm is also one that I'm pretty bullish on for quite a few reasons. one

being just I think it we stand a lot better chance of staying in control of the meta if we have a lot of like narrow AIs doing their jobs as opposed to you know a relatively smaller number of

giant AIs like running things for us.

The pull of that is obviously pretty strong but the I do worry that we're racing into having such general AIs that

can do sort of anything before we've really thought through what the ultimate consequences of that are going to be.

And I the narrowness safety through narrowness and maintaining control through narrowness I think is an underdeveloped paradigm.

>> I fully agree that that's again it's going back to human analogy. That's how

we operate as well. There are certain parts of the brain that that that are specialized to do certain tasks and yeah so I can absolutely see that.

>> This has been amazing. Couple of quick uh closing questions. What are you watching right now in terms of horizon scanning for surprises? Is there

is there a capability threshold that's on your mind or like what because obviously nobody can keep up with the AI news these days, right? So everybody has to pick and choose. What are the areas that you're watching? And maybe another dimension of that is what are the

metrics that you're watching? Are you

looking at ARC AGI scores? Are you

looking at GDP val? Are you looking at the MIRI task length chart? Are there

other what do you trust to give you the highest signal on what is actually important in the latest things that are coming out?

I I actually don't watch the public benchmarks as as closely like we do have internal series of benchmarks that watch very closely in terms of quality and latency for the tasks that we're optimizing for which is of course built

on top of the models that we get from the model providers. Whenever there's a new model that's about to be released, we'll go run our our tests, figure out what's improving, what are the gaps, and then that I watch very closely. In terms

of technology trends, one thing that is maybe unique to Snowflake of course is we have a lot of tabular data and that technology so far has been all about

text to SQL and semantic models. So

watch that space quite quite carefully and there are some new trends happening in that in that space. These tabular

foundation models are interesting. Being

able to quickly build forecasting models and so forth are now possible. So those

are other trends that I watch as well.

Yeah, that's an interesting one. The

there's a few public uh forecasting like benchmarks and competitions. I think

those are really interesting too. At the

point where the models are better able to predict the future than our best super forecasters or even aggregations of super forecasters, that will feel I think like a very meaningful shift in in

what's going on in the world.

>> Any contrarian takes? any anything you think like the anything you want to correct that you think the audience at large might be misled or misconceiving right now?

>> I mean we touched upon the one that that's very top of mind for me right now. I think you know the way we build

now. I think you know the way we build products is is change. I don't think it's contrarian but I don't think it's happening fast enough. Um we're at a point where how we build products needs

to radically change and uh you know that means change of behavior uh because we've been kind of trained to build products one way. So so to me that that's the biggest one you know in a in

a world where these coding agents are such kind of capable platforms. How do you build new products?

And you know in my mind it is all about kind of starting with that first uh and then validating things quickly u before you build the product in the first place.

>> Yeah I think that's I've been doing that with my mom. I made it made their the custom travel app travel planning app for the holidays and it has been yeah it's inverting that process right I made

a version she has it and now I sat down with her this morning over coffee and I'm like what do you want this thing to do that it can't do and she's I want to ask you to do more on this. I'm like,

"Mom, it's honestly so easy at this point. If you can articulate what you're

point. If you can articulate what you're missing, there's a pretty good chance we can get Claude code to just make it and you can have it from one session to the next."

next." >> My last question, then I'll give you the final word. What advice do you have for

final word. What advice do you have for enterprise leaders in general?

Obviously, you guys are much closer to the core. For all the executives and

the core. For all the executives and product owners at the companies that you serve, what do you think they should better appreciate or what can they learn from your experience? There are

[laughter] there are some uh you know enterprises that are still quite careful about adopting AI and and at this point it is

uh it is so powerful that it is uh there's a race. So you know the the faster enterprises adopt AI the more benefit they're going to get and and and

the more intuition that they're going to get that changes the trajectories of of these businesses. So to me it's

these businesses. So to me it's incredibly important to intuitively understand and natively use AI because it is going to change industries and

then underlying that is uh all about many of the hesitations tend to be about uh you know getting the uh data ready for AI from our perspective you know so

so that means investing in in in that core foundation to uh you know essentially get the data AI ready for AI to use. So that means breaking down

to use. So that means breaking down silos, getting the data uh you know accessible locked in for certain use cases like that becomes a core enabler to build on top of.

>> Makes sense. We've covered a lot of ground. I really appreciate your time

ground. I really appreciate your time and jumping through all these topics with me. Anything else that we didn't

with me. Anything else that we didn't touch on that you would want to leave people with?

>> No, maybe the thing to call out is we talked about a lot of great capabilities as as well as trends. One thing that is sometimes not necessarily appreciated is

how easy it is to to use AI and how easy it needs to be to use AI for adoption.

So that that's an area for Snowflake that's super core. So as we build products make making it very easy to deploy highquality AI at scale is something that we strive towards. To me

from a design principle perspective that is key as well.

>> Yeah, couldn't agree more. Beer

Schultzan, vice president of AI at Snowflake. Thank you for being part of

Snowflake. Thank you for being part of the Cognitive Revolution.

>> Thank you, Nathan. Thanks for having me.

>> If you're finding value in the show, we'd appreciate it if you take a moment to share with friends, post online, write a review on Apple Podcasts or Spotify, or just leave us a comment on YouTube. Of course, we always welcome

YouTube. Of course, we always welcome your feedback, guest and topic suggestions, and sponsorship inquiries, either via our website, cognitiverevolution.ai, or by DMing me on your favorite social network. The

Cognitive Revolution is part of the Turpentine Network, a network of podcasts which is now part of A16Z, where experts talk technology, business, economics geopolitics culture and

more. We're produced by AI Podcasting.

more. We're produced by AI Podcasting.

If you're looking for podcast production help for everything from the moment you stop recording to the moment your audience starts listening, check them out and see my endorsement at aipodcast.

And [snorts] thank you to everyone who listens for being part of the cognitive revolution.