The Exchanges

Q Take over the world. Yeah. Well, you know, nothing's perfect, I guess. Um, you have been a very active investor in the space. Um, Equally, this is not the first time there is this kind of like crazy hype cycle. Uh, what have we learned in terms of what happened last time that may or may not happen this time?

A Yeah, absolutely. I think it's, it's, it's important to think about like both, both what we learned, but also, you know, what might be different this time around. Um, I think what we learned back in, uh, 2017, 20 18, when there was a lot of hype around reinforcement learning, GANs, uh, things like AlphaGo, uh, things like Alexa, was that Applying AI is not a substitute for building great products. The kind of same principles of delivering value to users, of managing kind of their attention, all of those principles hold. You know, AI is just a technique. It's just an underlying technology that you can build, use to build great products, but you still need to build, you know, a compelling product that solves An urgent problem for somebody and has, you know, the opportunity to expand into perhaps like other parts of the workflow, other markets, et cetera. Um, I hope that's a lesson that we learned. I think time will tell. One thing that makes me nervous about, you know, what I see today is that I see a lot of startups that I think recognize that AI is not a silver bullet, but instead of, instead of focusing on delivering concrete value to people and almost delivering like Mundane value to them They,. They, they seem to be more focused on competing for mindshare. It almost feels like a lot of AI startups today kind of think about themselves as being embroiled in this, in this battl…

AI assessment note: “Applying AI is not a substitute for building great products.”

Q So one of the big questions, um, in from, you know, founders and investors is, um, big tech and whether that's an opportunity for startups to truly build, you know, self-standing long-term companies, um, you know, in a world where those incumbents are not lazy incumbents, but actually very much at the forefront. What, what do you think?

A Yeah, it's, it's a great question, and perhaps, like, a perfect illustration of, of, you know, what I was saying earlier about, like, new information, kind of, unveiling new patterns about the markets and, and, uh, the delivery of AI. Um, I think, like, even a month ago, there, there, there seemed to be, kind of, like, consensus that, like, startups had an advantage over incumbents. When it came to, uh, kind of leveraging LLMs and integrating them into their own products. People expected that existing tech companies, including, you know, some of the bang fortune, 500 companies would be slow to release new systems. The quintessential example is, is, you know, Google versus chat GPT. Um, I think the past month has demonstrated that that is not true, that, you know, a lot of existing companies can actually act pretty swiftly to integrate LLMs into their applications. I actually think this is one of the most powerful thing, one of the things that has changed the most in the ML landscape. It is so much easier to prototype LLM driven applications. You no longer need to hire an ML team before you can even determine, before you can even run an AB test or experiment to determine, like, if there, if this is something that users like. Uh, that said, you know, I think there will be an opportunity to kind of redefine workflows around LLMs. Uh, people often, uh, they draw comparisons to mobi…

AI assessment note: “It's not to build better versions of things that exist today”

Q So is that where you're looking then, um, LLM powered applications? Uh, are you also looking at, um, you know, this unfortunately named, uh, emerging category of LLM ops, um, you know, the, the, the technical stack that enables the deployment of LLMs. Is that equally interesting, more interesting, less interesting?

A Yeah. Interesting is, is a ambiguous word. I, I would say I love tools and infra. So, so like if you're asking me like what is more intellectually interesting, like I just love tools and infra and like I will find opportunity there because it's what I love. It's what I'm most passionate about. Um, I don't think we'll see that value is accruing exclusively at the infralayer or the app layer. We haven't seen that in any other paradigm shift. Again, if you think about, you know, mobile, if you think about like social networks, et cetera, there was value at the app layer, there was value at the infralayer, and I don't expect this to be any different. Um, Um, it feels like right now some of the, the value at the app layer can be captured immediately. Whereas the value at the infra layer will expose itself over time because so many of these companies who are, you know, trying to integrate LLMs into their existing products, like they started in the past two to three months. Uh, so, so like they haven't hit that brick wall yet. My, my colleague Sunil calls it the, the, you know, like stubbed your toe problem, uh, before, you know, tools and infra can like really get a breakout adoption. Uh, people need to try to do things on their own. They need to try to do things like without adequate insufficient tools, stub their toe, and then, you know, they'll come back looking for, for better to…

AI assessment note: “I don't think we'll see that value is accruing exclusively at the infralayer or the app layer.”

Q Let's, uh, switch to maybe talking about founders and the kind of founders you work with. You mentioned the example of Mother Duck, like people coming from Google. Do people need to come from Google or one of those, um, shops to work with you guys? How do you think about who's the, the best kind of technical founder that you tend to naturally gravitate towards?

A Yeah, that, that, that's such a tough question. You know, I, I, I've been investing for, uh, seven, eight years now, and it's something that I think about a lot, but frankly, like, I don't see that many patterns, um, related to, you know, successful founders or, or those who don't succeed. Uh, you know, we've had founders who Uh, have run small businesses and, you know, recognize a problem in their environment, you know, realize that that could be potentially scaled to other businesses who've been successful. We've had founders who come from Google. We've had founders who are, you know, commercializing the research that they, they did at MIT. Um, I think what matters is that like the founders Deeply understand and care about the problems that they're solving. Um, no startup is ever easy.

AI assessment note: “what matters is that like the founders Deeply understand and care about the problems”