No conversation with Larry and Sergey about any topic was ever easy. We argued endlessly about every single thing. We think very much in terms of what is the deployment life cycle? What does value creation mean? What does Snowflake bring to the table that anyone else cannot? I think every company needs to look really hard at what their enduring value is.
You clearly like putting yourselves into a position where you might fail or succeed or succeed. Sorry. Yes. Yes. Yes. Yes.
At SPC, we really focus on the minus one to zero part of the journey. And what that really means is our mission is to basically work with really talented people and to give them some time and space to go figure out, you know, ideas worth tackling, quests worth kind of like, you know, embarking on. One of the reasons why your trajectory is so impressive, Sha, is that you've kind of brought yourself back to minus one in your career a few times. You know, you were already having a successful kind of I would say trajectory in enterprise software when you kind of joined Google, and then you kind of went back to being an IC at Google. And then, you know, we'll get into it, but I'm curious, the question here is how have you picked what to work on? Like you have a good track record of picking good things. So like what are some rubrics that you think about like that are things worth tackling?
Yeah, I have a bit more of a nuanced thinking at least about how and when to drive change. I think change first of all is good. Resetting is good. But you also have to internalize that it's pretty brutal to start over from nothing. This started pretty early. I tried to get into IITs which are one of the elite institutions in India. I failed miserably the first time and much to my dad's disgust decided to spend the next year just studying for the exam. He's like you're stupid. You should just go to college. But that was a worthwhile bet to take. And then similarly I concluded sadly after 10 years that research was not something that I was very good at and I reset that and said okay I'll try and be a software engineer but I had enough of a software engineering background. So I think finding these moments in time to make these changes is important but most of all having a sense for what creates value in this space that you're playing in ideally combined with what are strengths that you bring to the table while you're making this change. That's the balance that I would strive as an example of a change that in retrospect I think was a bad change. I think going from Google running a very large team, very large business to starting a company looking in retrospect wasn't exactly a great change because many of the things that made me an effective exec at Google being able to run a pretty complicated business or a big team, building teams, motivating teams, setting product strategy, most of it didn't count for anything once I started Neva because it's a consumer company. So all about product market fit and that's a very chance occurrence. I felt very helpless when I was at Neva and that was more of an example of a change where there was just too much risk packed into it. While on the other hand the change from Neva to Snowflake quite honestly being an SVP at Snowflake to being the CEO, these were challenges, these were much easier changes to make because I brought in a set of capabilities that were super helpful even though it was still a new position.
So there are two things to take away. I mean there are many things to take away from what you said but here are two that I'll highlight. So the first one is you clearly like putting yourselves into a position where you might fail, right? Where you're kind of like, okay, I didn't get into IIT or succeed or succeed. Sorry. Yes. Yes. Yes. Yes. High upside, potential upside, right? Do you feel like you need to feel that thrill of like potential, you know, like you kind of I think that deep I mean certainly I as a founder, you know, I also like doing stupid adventure sports and I do wonder if deep down there is a desire to kind of like I want to put myself onto that edge of my capability to see what I can achieve.
I love challenges. I think you find out more about yourself. I think failures are actually more interesting. Where I tell people I tell myself like you aren't your failures, you aren't your successes either. And so that aspect of striving that's not for everyone, but I think it's very human and really important.
Before we move on to Neva, I'm curious. You know, you spent a long time at Google through what an incredible kind of run, not just at Google, but kind of working on products that I think most of us probably use on a daily basis, search and ads, right? Both of those being kind of like, you know, discrete products. What do you think the founders in this room could learn from what you learned at Google about today's time and age like building in the age of AI today? Are there things that you took away from your time at Google either in terms of company culture, how to build, what to build, how to prioritize that are particularly relevant today.
I mean all of you I think intuitively understand this, the true superpower of great founders, you know this could be an Elon, definitely Larry and Sergey, is their incredible reductive ability, is that ability to remove all superficial detail and get to the heart of problems and to arrive at it with relentless discourse. No conversation with Larry and Sergey about any topic was ever easy. They argued endlessly. My IITs were the same way. We argued endlessly about every single thing important or unimportant. But out of all of that was that desire to push for like what are the things that truly matter in a particular situation. The other side of Larry, Sergey and Eric who was a CEO for much of my time at Google was they did this while still being really good human beings. I tell people repeatedly that the three of them have never yelled at me even though I've given them an endless number of opportunities. I would just like go there and argue with them about everything and we would just go at it for hours and hours. They took these as intellectual challenges. I've never seen them be disrespectful to anyone. I never seen them scream at anyone. I think to both be good human beings, they're relentless competitors, make no mistake. We put lots of companies out of business. But it was still fair competition and they were really good to the people around them. To me, that combination of relentless drive combined with the humanity that is that's like defining both for those three and for Google as a company. And some of this you only realize and learn after you have left and you're able to look at yourself there or look at a company with an outside in view. I think is a truly remarkable place.
Yeah, I really like that phrasing of having extreme reductive power because you know it's interesting when you listen to interviews with Elon, a lot of the common refrain people have is like oh he just keeps on talking about the laws of physics. But ultimately the reason I think he does that a lot is because he has reduced kind of the problem down to like where is like the actual like what are the things that matter where are the places I can change something versus not. And I think that simplified almost zen-like view at the end of it is a result of a lot of kind of like reductive power over the intervening kind of like gears.
Yeah. I mean for the things that he deals with those are the immovables. I'm sure if you got Elon to do physics research, he'd be arguing about string theory and what's the immovable of physics. It's just it's a way of thinking.
Yeah, it's interesting. Some of the kind of like at least at SPC when we're talking to potential founders probably some of the most determinative kind of characteristics are just what we call internally clarity which is that for a given problem are they able to essentially like they can spend 30 minutes kind of walking you through all of the like the logic and the detail but if you just ask them like ultimately like what are the one or two things that matter for your business in the next 6 months and the next 18 months they have often have very short succinct answers which is like this is just kind of the one or two things that matter. If this happens, then I think we have a shot to kind of take it to the next level. If it doesn't, we're kind of like, you know, we'll have to zig and zag, but that's a pretty bad sign, right? Then they have a lot of clarity and in some ways it's almost not emotionless, but they're very passionate about it, but they also just have clarity like these are the two things. I wake up and I think about it and this is what I'm trying to do. Especially in the early days to use this framing and you already talked about Neva a little bit. You left Google after 15 years to go. What are the one or two things that were going through your head to kind of like that you wanted to tackle with Neva? Because part of where I'm going with this is is it that you had the right idea but perhaps one year too early, two years too early or that's actually a hard way of thinking about the world kind of like to kind of play the counterfactual back.
Yeah. I mean to be brutally honest I think that the first of all the idea was born abstractly that a search ads I mean an ad supported search model had reached sort of its useful limits and there's nothing wrong with that conclusion if you're not bound by ads like obviously you can do a better job. I think the real sin of Neva and my own arrogance in it was a lack of a clear road map on okay how like what must happen for there to be a much better search engine. It's not something that I should have figured out by building. It's too big a thing to build if you can't conceptualize in the first place. It took the arrival of GPT-3 nearly 3 years later for us to realize ah I want to click on a link said no one. And so that's like the brilliant insight for what can AI powered by a search engine do because you've removed the click. You give the person what they are looking for in one shot. I think the lack of that clarity. There's a lot of detail about how and sometimes you can draft onto a wave that comes that you are not quite expecting. That's all the you know it's better to be lucky kind of story. But the fundamental sin of Neva was that lack of clarity for okay new search engine but how. You know arguably the people who built GPT-3.5 and ChatGPT didn't really know what they had right until it kind of came out and we started using it and everyone was like oh this is actually the right way to kind of do everything from like internet search to commerce and so on. So sometimes you also luck but also just serendipity just kind of get the stuff out there and then you see it it takes off takes on a life of its own you know.
Yeah I'm not first of all serendipity absolutely but execution still matters. If you go back and look at early 2023 how many people here remember the we will make Google dance speech.
Oh yeah yeah yeah. This is from Satya, right?
The opportunity was there for the taking. I'm sure most of you don't remember Sydney. But Sydney was way better than ChatGPT. ChatGPT would like spew out trash through much of 2023. Anytime you asked a factual question, if you ask for a high query to do a great job, if you ask for a fact, it'll just like happily make it up. Sydney Bing's chatbot had solved all of this. It had this fatal flaw that every interaction with it would take 2 minutes. They never fixed it. These are the people that are supplying capacity to OpenAI. They didn't see the opportunity. Bing itself makes 10 plus billion dollars. So, they had a lot to lose by pushing on Sydney as the definitive search experience. But these are the people that actually created a much better search, not OpenAI, and they still blew it. So, it's hard building new products. It's really hard.
Before we kind of get into Snowflake a little bit, let's talk I mean I'm curious to get your take. Right now we are is this kind of an interesting time to be an open-source kind of like model right now. We obviously had GLM 52 come out which is getting rave reviews and in general it feels as though that yes GPT 5.5 is exciting. Fable whatever the whatever it is you know if I could use it is sounds pretty cool. There's a line of thinking by the way that says that all model progress is just dependent on underlying hardware and fable and 56 are merely B200 models. Yeah. Yeah. Yeah. I know. So I guess my question coming to you is like what is your current take on the model ecosystem kind of like the leading lab providers like you know obviously Gemini OpenAI sorry Gemini ChatGPT and you know Claude and then also obviously the six or seven kind of big open source models like how do you think about navigating in that world from your current position as you know the CEO of Snowflake.
The Snowflake CEO I mean first of all these have been recurring cycles of the like frontier model labs creating models that are significantly better along some variety of dimensions and then open source models catching up. I think by the way what is remarkable is that there's just a small handful of companies that can truly create frontier model still blows my mind.
Still kind of crazy yeah I'm kind of like what money can't buy stuff.
I still find that just wild. I think definitely what is happening with the power of the frontier labs is that at least with lots of enterprise customers there's a little bit of a worry of like wait all software and you see like Satya had both a tweet and I think an article Wall Street Journal on this where if you look at a world in which kind of all software disappears and these models are the interface between pretty much all of humanity and data underneath. Let's say it's a very different kind of world.
So there's an amount of trepidation about being too reliant on a single model source, right? Which is really kind of an intrinsically a trepidation about centralization.
And so one of Snowflake's strengths has always been that we were the neutral data player when it came to the CSPs. There are tons of customers that will run their primary on AWS and run a disaster instance on Azure and part of our selling prop has been your implementations are easily portable from AWS to Azure in ways by the way all of you know this you can't actually move software that way this whole thing of cloud computing blah blah in practice is just miserably hard to move any modest size deployment from one CSP to another just a whole lot of hooks that are super CSP dependent that most companies routinely fall trapped. And so part of our setup now when it comes to AI is that with both our kind of like our sort of Snowflake orchestrator which is Koko the coding agent for Snowflake as well as what we think will be the work front end called co-work. We are now able to say things like well you know you use it we will make sure that they work just as well on both OpenAI and on Anthropic and that's a selling proposition and the open source models in many ways reinforce that because we have a team that's working on running our harnesses on top of the open source models. We have a very good inference team. We can run those models directly ourselves very different economics for us as Snowflake. So this is a pretty good setup overall.
We can also do things like cost optimizations. One of our engineers came up with a very clever compilation technique for skills which is almost deja vu all over again which is you had code and then you had models and then you had English describing what to do and then you go back to code for the most simple things that you're doing. But it's very cheap. It's like wouldn't it be nice if all of these skills were actually converted into like simple code that you could then kind of like have like essentially byte code that can run much more efficiently that a compiler could then essentially optimize 100%. And so you see this cycle coming back and so in that sense it's actually pretty good news for companies like Snowflake. But it also again I think there are many like fundamental lessons that will continue be true. The harness is the new browser. Regardless of what anyone tells you, being like buried under someone else's harness, you know, some of us might have to do it, it carries a bunch of risk. And so, we've consistently believed this is partly my consumer background. I have my co-founder who also worked in a consumer background. So, we place a lot of value on create products that directly touch the end users. And so, that comes in very handy in moments like this.
If I can push you a little bit here to make a forward-looking kind of prognostication about how do you think that the model ecosystem evolves over the next 12 to 24 months. Do you think that or maybe even the next few years do you think that there is enough of a concentration of capital slash chips slash maybe research talent at the best three or four organizations again you know OpenAI Anthropic Google maybe Meta such that they will reliably have a 6 to 12 month lead or do you think that the gap between open source and closed actually starts to narrow. I don't actually have a point of view on this myself.
I think it comes in cycles. You're right that you know you kind of get a big closed model release and then open source kind of chips away and then you kind of get a big so and we'll see how Fable does obviously I think it's a little hard to say so it's hard to know how big a jump Fable is right now.
You either make a lot of predictions or you make none. So I think it's just hard to tell. I think what's been unique about like both chips and AI models has been the infinite demand.
For all kinds of different chips and for all kinds of different models. I think it's pretty unique to have a lot of jobs to be run on top of A10 chips. Like they're perfectly good. We used them 6 years ago at Neva, 5 years ago at Neva. They're still super useful today to the latest B200s.
Um so, and I think the price per compute hour has actually gone up relative to because there's enough kind of, you know, demand for it. This will just like continue to evolve. But I think what feels hard right now is the ability for frontier models to truly have like a big big big gap. The fact that there are two of them first of all or more potentially plus a bunch of basically open source labs that are able to catch up relatively quickly. I think that's the most interesting phenomenon and honestly again that is good for all of us because it means that there's not concentration. All of us talk a big game about things like open source while quite forgetting that Google effectively put an end to all search research.
