Venkat Mocherla

Yeah, I think it's important to understand in all of this, the context of things and where are we sitting today? So if you and I were chatting, I came to the Bay Area in 1999 at the peak of the dot-com era. And the hottest startup at the time was a company called Webvan, right? So if you think about where we are today,

Yeah, I think it's important to understand in all of this, the context of things and where are we sitting today? So if you and I were chatting, I came to the Bay Area in 1999 at the peak of the dot-com era. And the hottest startup at the time was a company called Webvan, right? So if you think about where we are today,

is we're really in the early 90s era of thinking about everything was a dot-com company and I kind of equate it to sort of AI. The fundamental, if you ask yourself, have we been transformed because of the internet? The answer is absolutely yes. Now, I think what you're trying to get to is where does the value sit over time?

is we're really in the early 90s era of thinking about everything was a dot-com company and I kind of equate it to sort of AI. The fundamental, if you ask yourself, have we been transformed because of the internet? The answer is absolutely yes. Now, I think what you're trying to get to is where does the value sit over time?

And I think just because Webvan didn't work out, Instacart did, and there's business model changes, there's technology changes, there's cost structure changes that have happened over time for Instacart to be a success and for Webvan to be not a success.

And I think just because Webvan didn't work out, Instacart did, and there's business model changes, there's technology changes, there's cost structure changes that have happened over time for Instacart to be a success and for Webvan to be not a success.

And so when I think about, for example, I think you asked about DeepSeek and et cetera, there's been a race, Scott, on the sort of funding of these frontier models for LLMs. And so you've seen OpenAI, Anthropic, and the open source side, there's been a company out of France. There's billions of dollars that have gone into this space. And the reason why DeepSeek is so top of mind right now

And so when I think about, for example, I think you asked about DeepSeek and et cetera, there's been a race, Scott, on the sort of funding of these frontier models for LLMs. And so you've seen OpenAI, Anthropic, and the open source side, there's been a company out of France. There's billions of dollars that have gone into this space. And the reason why DeepSeek is so top of mind right now

Obviously, there's a sort of geopolitical arms race that's happening. Some have equated that this is the race towards the atom bomb, right? So there's a geopolitical race of who will develop the most in AI. I genuinely believe in American dynamism, as a lot of people have said, and I think there's been a phenomenal amount of development done here.

Obviously, there's a sort of geopolitical arms race that's happening. Some have equated that this is the race towards the atom bomb, right? So there's a geopolitical race of who will develop the most in AI. I genuinely believe in American dynamism, as a lot of people have said, and I think there's been a phenomenal amount of development done here.

The reason why DeepSeek is so top of mind for people is twofold. One is Again, the news is just coming out, and I think everyone should do their own research. But the cost structure, because of the sanctions on NVIDIA, the limitations that it put on, the cost structure is fundamentally quite different than what OpenAI or others can provide.

The reason why DeepSeek is so top of mind for people is twofold. One is Again, the news is just coming out, and I think everyone should do their own research. But the cost structure, because of the sanctions on NVIDIA, the limitations that it put on, the cost structure is fundamentally quite different than what OpenAI or others can provide.

So there's a phenomenal difference in what cost of compute and the cost of accessing these frontier models. The second piece of this um, you know, obviously is there's a, there's now a sense because you can, you can, if you go to the app store, deep seek is the number one app, um, now over open AI. And there's an open question of how much data are we sending?

So there's a phenomenal difference in what cost of compute and the cost of accessing these frontier models. The second piece of this um, you know, obviously is there's a, there's now a sense because you can, you can, if you go to the app store, deep seek is the number one app, um, now over open AI. And there's an open question of how much data are we sending?

What's the sort of implications of that? There's actually, there's a lot more implications on supervised learning. There's a lot of implications on, you know, what does deep seek mean for the entire ecosystem for startups? But I sort of go back, Scott, again, my one advice for healthcare startups is, Focus on the last mile.

What's the sort of implications of that? There's actually, there's a lot more implications on supervised learning. There's a lot of implications on, you know, what does deep seek mean for the entire ecosystem for startups? But I sort of go back, Scott, again, my one advice for healthcare startups is, Focus on the last mile.

So if you think about compute data algorithms and sort of what is the application layer on the last side, if compute becomes cheaper for everyone, that's great for startups like Midstream and others. Because if you can think about the workflow, the challenges for the front end user, and deliver on a killer app for that person to make that person's life easier, it doesn't change one thing.

So if you think about compute data algorithms and sort of what is the application layer on the last side, if compute becomes cheaper for everyone, that's great for startups like Midstream and others. Because if you can think about the workflow, the challenges for the front end user, and deliver on a killer app for that person to make that person's life easier, it doesn't change one thing.

And so if you're a venture investor in one of these frontier models, I think there's a lot of implications you're asking, you're talking to yourself. But if you're fundamentally an AI startup in healthcare, this is net effect is there's an arms race happening and you need to focus on the last mile problem.

And so if you're a venture investor in one of these frontier models, I think there's a lot of implications you're asking, you're talking to yourself. But if you're fundamentally an AI startup in healthcare, this is net effect is there's an arms race happening and you need to focus on the last mile problem.