00:16:03 · Sridhar Ramaswamy
[00:16:03] S0: 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 sized deployment from one CSP to another. There's 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 Coco, 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 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 the company. 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
00:17:59 · Aditya Agarwal, Sridhar Ramaswamy
[00:17:59] S1: be nice if all of these skills were actually converted into simple code that you could then have essentially byte code that can run much more efficiently that a compiler could then essentially optimize? [00:18:09] S0: Percent, yeah. 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 fundamental lessons that will continue to be true. The harness is the new browser regardless of what anyone tells you. Being like buried under someone else's harness, some of us might have to do it. It carries a bunch of risk. And so we have 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.
00:18:50 · Aditya Agarwal, Sridhar Ramaswamy
[00:18:50] S1: 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 twelve to twenty four months? Do you think that... Or maybe even the next few years, do you think that there is enough of a concentration of capitalchipsmaybe research talent at the best three or four organizations, again, OpenAI, Anthropic, Google, maybe Meta, such that they will reliably have a six to twelve 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 kind of get a big closed model release and then open source kind of chips away and then you kind of get a big 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. [00:19:45] S0: 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 six years ago at Neeva, five years ago at Neeva. They're still super useful today to the latest B200s. [00:20:22] S1: And I think their price per compute hour has actually gone up relative to same because there's enough kind of demand for it.
00:20:22 · Aditya Agarwal, Sridhar Ramaswamy
[00:20:22] S1: And I think their price per compute hour has actually gone up relative to same because there's enough kind of demand for it. [00:20:27] S0: This will just 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. What fool academic wanted to work on search if you couldn't get your hands on one thousandth of the data that the biggest practical product that was being used by everyone on the planet had. Similarly, it's not exactly like the algorithms that powered social networks, where a big source of open source, even though Meta itself was a big source of open source for lots of other things. And so I think we could easily have gone down into a world in which one of the model providers was the only model provider. I personally think that we are all damn lucky that it's going to be three, five or seven companies.
00:21:56 · Aditya Agarwal
[00:21:56] S1: Yeah. That's a good observation. I mean, obviously, as somebody who was at Facebook pretty early, you know, and I think there were three of us who actually released Facebook's first open source package, but it was all infrastructure. It was basically an RPC layer thing. But you're totally right that the more broad category of essentially how do you do basically content ranking and content like we had there was a lot of research on it in the early 2000s and then everybody stopped because if you can't get access to the newsfeed data, then what are you doing? There's no point, right? Netflix did its Netflix challenge for a while and then they made it closed source and nobody wanted to do anything. This is totally mean, so I think it's actually just really good for us as users that all these open source models exist. And certainly for me, I'm running my personal agents and I think, it was just like a week ago, I don't know, I just out of nowhere, my Hermes agent decided that it was going to take up $600 of Opus 4.8. I'm like $600 a day, that's a little expensive for even a really good assistant. Then it's pretty easy. You swap in kind of like a cheap open source model and you let it run and you see if you're happy with the results and it's like, oh, actually, I can't tell the difference. And it was fantastic. Switching gears into perhaps Snowflake, you know, like, I'm curious, you know, Snowflake, I think technically was the world's ever biggest software IPO. That's correct. That's correct, right? If I remember correctly, that's right. Was an IPO In terms of dollars raised. In terms of dollars raised. Yeah, it was like the world's biggest ever software IPO. And historically, I think you kind of made this point earlier, Snowflake is in the business of writing really complex software and then essentially selling it and deploying it essentially, right? But we also live in an age where software is a lot easier to kind of create, to produce. And so far, we talk a lot about software that we might wipe code. Maybe it's websites. Maybe it's apps. Maybe it's internal tools. Maybe it's dashboards. Maybe now it's like V1 of my startup idea. But we also have people who are now starting to build file systems. We have a couple of people in our portfolio who are like, actually, five of us traditionally a file system is pretty complex. But you know what?
00:24:00 · Aditya Agarwal
We can write a file system because we can kind of do it a lot. So, at which kind of the logical question here is that there's probably somebody out there who is along with the help of kind of these coding agents building a modern kind of analytics kind of like storage and database solution. How do you think about that?