AI Applications

“I do think that like, it's impossible for one layer to capture all valuable memory because the apps own so much important context that the model labs don't have. And they, the model labs in order to get this to work, they'll have to incentivize the apps to, li…”

“More apps getting built is not going to mean more companies getting built. Because those are two different things.”

“My rule of thumb is that in the next three years, we'll probably have half the people in GNA type activities in companies. Things like marketing, things like finance, things like HR, because there's a lot of process management there, and a lot of process manag…”

“SaaS applications will give way to AI applications, the difference being SaaS applications have no opinion, AI applications will have opinions. That's a fundamental rethink we need from a workflow perspective.”

“The old school traditional legacy cloud business was so latency focused because of some of the applications, but this new fleet of AI applications are far less latency sensitive.”

“The most compelling AI applications I'm seeing across commercial or, you know, so private sector or public sector are being built by these green suitors.”

“The AI applications are so incredibly powerful and amazing, but because they can do so much, and it's not yet clear exactly how to monetize them, it is a bit confusing to understand how will they be evaluated, you know.”

“The more stuff that gets integrated and the more company specific like kind of rules built around the model stuff you have, the stickier it's gonna be.”

“I need to build some kind of system of record for you, some kind of vertical operating system for you so that you can't just go switch out for the cheaper player.”

“Right now, currently, I would say that the complexity of the application itself is very high and is not subsumed in the foundation model. And so I think these things are not as straightforward as they appear and the benchmarks can be misleading.”

“MCP is a protocol between the AI application and like servers, right? So the model is actually technically not involved in MCP.”

“And I think you will see evolutions as we're getting better of like how we use these models and the infrastructure around it gets a bit more mature. And you suddenly can assume that most model like AI applications will have some form of like Sandboxing for exe…”

“But I do think most of the value will accrue to the apps. So it's kind of, and I think that's common sense. It's just, I just don't know which apps.”

“I do believe that to be very successful, and I've seen many examples of that, you need to incorporate a lot of the knowledge, specifics, the tribal knowledge or institutional knowledge into, let's say, the product or the model to be able to be highly effective…”

“My big idea for 2026 is the death of the prompt box as the primary user interface for AI applications. The next wave of apps will require way less prompting. They'll observe what you're doing and intervene proactively with actions for you to review.”

“And there are a few interesting use cases, but you know, fast forward two years and the way that you get a mortgage is going to be talking to your AI application.”

“In the past three months we have almost put in almost, like, 50 plus applications through this process, and 12 of them have already gone into production.”

“Salesforce was rebuild Siebel in the cloud, and it was fun, and we made a lot of money, but you're right, somewhat boring. What is really exciting now is you're not just talking about rebuild Salesforce exactly the same, but with a slightly modern UI. You are …”

“Today, the actual amount of money that we already know people and companies are willing to spend is the combined salaries of all the people they're currently paying to do the job. And that number is like a thousand X bigger. You pay 20 dollars a month to solve…”

“These apps that need to be trained with forward deployed engineers and work. I don't think very many of them are less than 30 or 50,000 dollars a year. And a lot of them actually try to kind of have a price point that's approaching a hundred K like 60 K a year…”

“You can actually build an application just a layer on top of a legacy systems where it can able to give you immense value, but you will be replaced because a new tool comes. So it's like there is no stickiness to it, right? So you are actually as good as a wor…”

“There applications that maybe aren't wildly popular, but they're so popular amongst a passionate group of segments, these companies can scale to like hundreds of millions of revenue.”

“That even if you froze the model capabilities, there's trillions and trillions of dollars of applications just waiting to be discovered here.”

“The bar for what product market fit and what an effective wedge looks like has gone up in AI because you've just seen explosive growth for companies when they do have them.”

“What that tells me is the model layers being increasingly commoditized. And that there's not going to be a lot of intrinsic value in, in, in that intelligence layer, that operating layer. You're going to have to build applications that guys like Bill Gurley an…”

“And like, we're like pretty clear past the sort of rapper phase. Like, I don't think there are any rappers anymore. Like, Building good products with AI is really hard, and the founders doing it now have, like, really good kind of intuition for how to do it.”

“What I think is going to be very durable with AI is, like, everything that I've seen is, like, building the applications is, like I mean, I don't want to say it's, like, trivial, obviously everything's hard, but what's much harder than building AI applications…”

“Applications are just like the Obvious place for me, both because of my background, but it's also what I think is like the best place for venture in 20, 25 to be investing because you're basically investing behind the fundamental belief that like AI is going t…”

“Everyone is now creating new data systems, creating new AI applications, AI devices, but inherently, if you're adding these in, you're also creating more fragmentation because We now have to use this AI application that only works with this device or only will…”

“Before I would say, 18 months ago, it's like, oh, well, who has the best model? But then coming forward, it's like really who has the best workflow and who has the best product. And we humans are just demanding. We want the best. And so when that, when someone…”

“And so, you know, I think there's a question of how it will play out. Will it just be a simple subscription or is there gonna be something more You know, freemium that comes along. I would bet on the latter. I would bet on the latter.”

“And what I think a lot of firms are running into right now is if you're only looking at that last step, if you're only looking at the system's performance as a whole, it can be very difficult to understand when, where, and why behaviors are shifting within thi…”

“I kind of think agent is just a word for AI applications, right? Anything that uses AI kind of can be an agent.”

“I would upload a hard PDF into AI applications, and I would show them that their application is hallucinating. It would be a very clear, like, here's the document that I sent in, and your model is actually making up information that wasn't there. And for any p…”

“I think there's this usual, like, VCEO, we should invest in, like, the tools, you know, the picks and shovels instead of the application layer. But I think it's, this is, like, the first time where, like, the infrastructure is lagging the applications.”

“Half half the challenge with a lot of the AI applications is you don't have the models accredited. You don't have the models deployed in a way that they can actually be used.”

“Anytime that I try to help someone build an AI application, they always get stuck on how to move beyond a demo product. And they get stuck like how to systematically improve things and measure it.”

“People don't know how to do anything other than Plan A, which is to go out and try to collect as much real world data as possible and take months because we're going to employ, like, human annotator teams to do this. Or Plan B, which is I'm going to ask an LLM…”

“If you're building applications that rely on a browser and it's more part of a greater like AI app experience, you probably need something that's more like infrastructure, not a consumer app.”

“There is no difference between taking an AI application to market versus taking the applications to market that used to exist in the past as well.”

“So one thing that I look at is you can look at this by industry and you can look at it by task, which is how expensive is the token? And I, I've come up with like different ways to call this, but basically if you look at something like that share purchase agre…”

“The apps that ended up mattering on the internet literally didn't exist. Before the internet. And I think that's what people are losing sight of.”

“Because if it's really quick to build, you're not building a data proprietary moat around what you're building. It's gonna be really hard to sustain the pricing that you see today, and that's gonna become a really, really challenging business proposition.”

“The next phase and what gets me really excited is there's richness in the data of the users that is being created as they use your product. The more someone puts in an input that's unique as to how they should look at a certain certain asset class or a certain…”

“For applications it's really, really important to understand the use cases well, so they can design like the set of metrics and then you can walk backward from that and map it to like the model metrics.”

“Is that wedge product alone defensible? I would argue no, right? Today, it feels like magic to the providers that they're working with, but I think that capability is going to become commoditized over time. They may have an advantage because they've trained a …”

“I think the bulk of the value though, so there will be valuable companies there, but the bulk of the value will be in the application side.”

“The focus on hardware and models and not enough focus on the applications of AI. I think many of the defining companies in AI will be delivering consumer and business solutions that happen to be powered by AI, not just the models themselves.”

“And so the reality is ClickHouse has been used for in AI applications or to power AI applications as a feature store for years.”

“Most of the apps, the overwhelming majority of the apps that we're seeing in AI today are toy apps that are run as proofs of concept and demos and run in a sandbox. It is not production code.”

“And the reason is because you could build the best experience in the world that runs on localhost, but if all of a sudden you actually try to launch it as an app and the thing takes 35 and 40 seconds to generate something, it's DOA.”

“Part of it is, there's such a gold rush with it that relatively few of the companies are doing kind of good apps on this stuff yet, because they're just kind of working on the pure AI general purpose.”

“So in that sense, AI infrastructure may be well suited for companies that can invest billions of dollars ahead of time in building big data centers. The applications on the other hand, I think that sort of rapid feedback loop and kind of creativity is really w…”

“If you don't engage in these kind of things that look like toys, like, this wave will pass you by. That I'm a hundred percent convinced.”

“I think in the arc of time if it is a shift like the internet was in many ways the applications, Which might not look like an applications, right, in the way that we think of them, but might look more like, you know, this control center idea will have profound…”

“Like, I think there can be a lot of high fructose corn syrup of adoption. Like, there's an amazing magical moment in a wow, right? When you go experience ChatGPT for the first time, or some of the imagery models for the first time. The rate of adoption of, lik…”

“There's whole categories of companies that are going to be cat, you know, Category defining independent successful big public multi-billion dollar companies that are just applications of large language models in a specific domain for a specific user with a gre…”

“The best applications just haven't even been invented yet that You know, we, you know, we're still at, like invention of electricity kind of moment, or invention of the computer, where we don't really you know, we don't really know what the coolest things are …”

“And I'm a big believer that that last mile of integration can be where a lot of the value accrues.”

“It's like letting you provide a great quality of service, but in the AI realm, it's just so mandatory. Like your V 0.1 already needs that critical feedback loop.”