The State of AI: Models, Moats, and the Consumer Renaissance

1. 0:00 – 1:20

   Intro

   1. JKJen Kha0:00

      To help me break down all things, uh, around this incredible abundance, so I'm gonna bring up Anish Acharya. And, uh-

   2. AAAnish Acharya0:07

      Good morning, Jen. Hi.

   3. JKJen Kha0:08

      Awesome. Awesome, awesome. Hey, Anish. Uh, Anish and I were at the ch- at GP offsite, uh, earlier this week, and, and he shared with me that he's already running GrokBot, uh, and it's purchased a bunch of jeans for him. So Anish, do you wanna, do you wanna drop what you purchased?

   4. AAAnish Acharya0:23

      True story. True story. Yes, I'm gonna reveal an important secret, um, protected IP, which is that I mostly wear Frame jeans. Frame is a great brand.

   5. JKJen Kha0:31

      Oh.

   6. AAAnish Acharya0:31

      Um, and GrokBot's is an awesome product, actually. I'd say the kinda defining characteristic of GrokBot's is sort of resourcefulness. You know, I went to bed a few nights ago and said, "Hey, uh, buy me a pair of jeans that are inspired by these." I took a photo of my current jeans. Um, I said, "Don't spend more than $500," and get it done. I woke up in the morning and it had researched, found a pair, same fit, different wash, used my credit card, purchased them, and they're on the way. So I think that is gonna be something that we see more and more of. We already have the capabilities, and now a lot of the kinda unlock will come from resourcefulness, um, and also the kind of product architecture delivered in a way that most consumers, um, can understand.

   7. JKJen Kha1:10

      Awesome. Awesome, awesome. Yeah, I told my team that I'm gonna set my bot to finally take care of the pile of things I've been promising my husband that I'm gonna sell-

   8. AAAnish Acharya1:19

      [laughs]

2. 1:20 – 2:49

   Who Wins the AI Model Race in Three Years?

   1. JKJen Kha1:20

      ... for the last two years. That is the project for, for this weekend. Uh, so Anish, we asked the question earlier, which one of today's AI leaders will be the clear winner in three years from now? What's, what's your take?

   2. AAAnish Acharya1:31

      I'm a many winners guy, and I see I'm in a, you know, I'm in good company with many of you. I mean, if you look at what's happened in the last two weeks, um, you know, xAI went from not even being a real contender on the model side to being, you know, one of three. So we extraordinarily went from a two-horse race to a three-horse race. And, you know, even more broadly over the course of the year, we went from Anthropic feeling like they were so dominant they could do no wrong, to OpenAI, who's just had an excellent three months. You know, the new models are exceptional. The new Codex harness and ChatGPT desktop app is very well done, and we're seeing the sort of specialization in different directions of these labs. They're both growing like crazy, you know, despite each other's continued successes, um, xAI and as OpenAI did as well. Um, so I'm, I'm definitely in the many winners camp.

   3. JKJen Kha2:16

      Yeah. It's interesting to see the sentiment also on X, which is not always a perfect, you know-

   4. AAAnish Acharya2:20

      Right

   5. JKJen Kha2:20

      ... uh, weather vane for the future, but, but oftentimes a early indicator of at least where developer sentiment is. And there's been a lot of pushback from Claude, it seems like recently [chuckles] on people, uh, in terms of token usage, et cetera. And so, uh, you know, developers tend to be fair weather fans on these things. They will go where the latest and greatest and very best model is. And in particularly the last six to eight weeks, I think we're gonna see some very interesting, um, traction in terms of the flow of activity. But obviously Anthropic's going public, uh, later this year.

3. 2:49 – 7:51

   What's Next in the Frontier of Intelligence

   1. JKJen Kha2:49

      Um, and also, you know, there's a lot of, lot of keen interest on this. So with that, uh, that actually brings us straight into the topic of discussion today. So where, where and what is next in the next frontier of intelligence?

   2. AAAnish Acharya3:02

      Amazing. Thank you, Jen. So let me tee this up for everybody, um, and please hop in if you've got questions. So let's first cover the kinda macro and what's happening, um, at a market level. Then we're gonna hop into the application layer broadly and sort of talk through why applications are the productization of the intelligence primitive. And then finally, let's talk about consumer, you know, with the launch of GrokBot and a few other products, it's actually been a very fun, um, couple of weeks in consumer. Okay. Uh, hopefully our, our dear friend Leopold doesn't mind me poking a little fun at him here with situational awareness. Please, next. Um, all right. Look, I think that the kinda case for this being a bubble, um, is over sort of discussed or at least fully discussed. I think actually the out of distribution topic that's less discussed is what if we're insufficiently optimistic? And if you look at some of the underlying indicators, what they point to is essentially infinite demand and highly constrained supply. You know, things like B200, which is a non sort of cutting edge GPU prices going up on a per hour basis. That is very, uh, strange. Normally, we see these things be highly deflationary, and it sort of points to very constricted supply and essentially infinite demand. So we're thinking and talking a lot about what's the kind of informed case for optimism here, given some of these second order indicators. The SaaS bubble was a very, or the SaaS sort of, uh, you know, whipsaw was, uh, an, an interesting peek into market psychology. You know, back in February when we saw this, you know, thirty to forty percent drawdown on a bunch of SaaS names, we said that the market has oversold software. Lo and behold, here we are, and many of those names are back up forty percent, so I'm not quite sure what we collectively accomplished. But I'll tell you what we said then, which is still true today, which is for the enterprise, software spend is eight to twelve percent. It's just not a huge proportion of spend, so the upside to vibe code your own payroll or CRM is not particularly high. The downside is essentially unlimited. You know, obviously there's all kinds of sort of compliance, um, implications of not getting things like payroll right. So most enterprise software today demands a level of precision that just isn't afforded by coding agents. Um, the, the one thing that has happened though is the sort of tide has receded. So for a lot of SaaS companies that had a ton of SBC and, you know, things that distorted their economic performance, I think that's very much, um, visible now, and they're going to have to sort of accelerate or die. So, so less bleak for the SaaS, uh, sort of market than perhaps we all collectively thought for a few months there, but still some sort of existential questions to address. You know, there's been a huge sort of discussion of moats. Are there any moats? There's no more moats. And it's, it's very funny because if you actually study, um, moats, which I think are most famously codified in the book "Seven Powers," it's one of my favorites. The vast majority of moats actually are not affected by abundant low-cost intelligence. You know, when you think about network effects, scale effects, which shows up in distribution, brand effects, which we tend to discount in Silicon Valley, these things are as good as they've ever been. You know, no amount of coding agents is gonna make Nike not Nike. And Instagram, um, the power of Instagram was never the complexity of building the Instagram app. Of course, it was the kinda network behind it. So I actually think the majority of moats are as good as they've ever been, and, and of course, are still critibull-- critical to building compounding value. There are a couple of moats that are exposed. For me, the integration moat is the most obvious one. You know, SAP is, is so famously complex to integrate into and out of that it's a sort of existential risk to even migrate from one version of SAP to the next. Coding agents makes this dramatically better. I think there's a bit of an existential question actually for SIs and GSIs as to what will their value be when they've historically been this sort of point of integration. So I do think this moat is a little bit at risk, but for the other traditional moats, they persist, and they're as important as they've ever been. Yeah, I think this is a really important concept. You know, as you start to think about what are the job functions in the enterprise that are alpha creating, it's typically product, sales, engineering, research. And conversely, what are the job functions in the enterprise that are sort of... Maybe administrative is, is, uh, too, uh, too bleak, but they are supporting other functions, legal, um, HR, finance, et cetera. We really think that the kinda rational architecture and the one that is emerging is that for jobs that, um, have unlimited upside, like sales or product, you always wanna use frontier tokens. And the reason for that is you just don't know what the value of the new product feature or closing a customer account is. It's effectively unbounded, and therefore it's economically rational to pay almost any price for a model that's even one IQ point smarter, you know, your Fable 5 or your Grok, uh, or your, um, uh, GPT-5 six. Conversely, when you talk about something like finance, you know, the best way to close the books is accurately. You can't close it, you know, 10X better than accurately. So as a result, you kinda have this bounded upside problem where it makes sense to use open-weight models with reinforcement learning for the kinda Pareto efficient, um, cost

4. 7:51 – 19:57

   Why Open Source Is the Only Option for Some Startups

   1. AAAnish Acharya7:51

      curve.

   2. JKJen Kha7:51

      Maybe before we, we go off this one, 'cause this is a, a great debate, and, and again, when Kimi dropped a few weeks ago, there was a lot of consternation, uh, about this, this topic, just given the relative cost, which was the focus of, of the topic of discussion. But, you know, um, our, our founder, Jesse Zhang, from Decagon dropped this great post around the fact that in some respects and, and for a lot of companies like Decagon, open source is actually the only option. It's not just cost. It's, it's that they can actually localize it, train, fine-tune it. And so maybe unpack a little bit of that configuration. Talk through the, the nuances there and, and why folks shouldn't be concerned, even though that is the case for startups, that there's a lot in the way of abundance around this topic.

   3. AAAnish Acharya8:32

      Yeah, I mean, one of the big topics that we're seeing or one of the big trends is that there are just... One, there are sort of comparative advantages of different models. So, and the models often have sort of areas of focus that are almost at, um, tension with each other. So you see a certain set of models that have a high degree of, um, neuroticism, like the-- they're sort of autistic models. GLM 5.2 and GLM 5.3 are great examples of this, where they're very literal, and they'll only do exactly what you told them to do and nothing more. Then we're seeing models like a K-3, um, that are just much more sort of open, and they're very presumptuous, and they're creative, and there are roles for both types of models in the organization. And, and often the sort of shapes of those minds, if you will, are at odds with each other. So that is like one reason you actually wanna have multiple models. The reinforcement learning is a really important point. Um, you know, if you actually have a problem that you can specialize the model around with your reasoning traces, you can start to create this compounding advantage in your domain for your customer base, where you're able to kinda shape the intelligence to be better than any general intelligence for your problem. I know Harvey's had some great results with this as well. Now, the trade-off of that kind of reinforcement learning is you lose generality. So if you have the best sort of model that's fine-tuned for solving legal problems, it may not be great at solving sort of theoretical math problems, and that's okay for Harvey's uses or, in the case of Decagon, customer support. So this sort of open-weight specialization property is something that's very unique and one of the reasons our startups are selecting them. This is also a big topic. We've learned so much since January. We should really do this monthly, Jen, or yeah, I mean, honestly-

   4. JKJen Kha10:06

      Yeah

   5. AAAnish Acharya10:06

      ... weekly. There's just so much changing. So in January, February, there was a lot of discussion, and it's, it's, it's very idiosyncratic and interesting. You know, Anthropic Claude released what is called a legal plugin. You know, plugins are just collections of skill files. You can think of it as a zip of skill files. Skill files are just prompts. They're just long prompts. And there was this huge panic, and all of a sudden, Thomson Reuters and a bunch of other sort of, uh, you know, big legal names traded down dramatically. But those were really just prompts, and there was a lot of discussion about if labs were going to integrate, vertically integrate up into the application layer. Instead, we've seen the very opposite, which is, yes, they are vertically integrating, but they're vertically integrating down into inference and compute. And it's actually logical now, um, in hindsight because the workloads for inference are very homogeneous, so you can build enormous scale in one part of the value chain. Whereas when you think about the application layer, you know, you've got so many idiosyncrasies and unique needs in terms of pricing, packaging, um, sort of productization, how the market wants to buy. So it's actually a much more challenging and OpEx-heavy proposition to move into the application layer versus moving down into the inference layer. And this is the point I alluded to earlier, which is sort of this discussion of model commoditization. You know, if you use the models every day, which I do, I sort of hold myself to a standard of making something either small or big with every model that comes out. You, you start to appreciate the fact that these things are, are not commodities, that they have comparative ad-advantage at a domain level. So a great example is OpenAI with their new, um, GPT models are just so, so good at knowledge work. The harness is also very well set up for knowledge work. You know, if you've used the ChatGPT desktop app, you know what I mean. If you haven't, please install it. It's very, very cool and interesting, and it's the perfect sort of... When I say harness, I kinda mean kind of product container, like a browser. Um, it's the perfect product container to do spreadsheets and slide presentations and written documents, and all of that type of work. If you look at CloudCode, which many of you, I'm sure, have used, it's just so oriented towards software engineering. You know, it's in a terminal UI. Everything from the small design decisions to the areas in which it specializes, like code planning and code testing, is oriented towards the software engineer, and there are many trade-offs both products are making for that sort of respective specialization. So one, you've kinda got this domain-level specialization that's already occurring, and then two, as I mentioned earlier, you've got this sort of... I think of it as the big five sort of personality traits if, if folks have studied that. You know, you can't be both highly open and highly neurotic, um, and you know, sometimes when you have an intelligence you're applying to an accounting problem, you want neuroticism. When you're applying it to a design problem, you want openness. So you actually have a need for both types of minds in the organization, which is why you would select something like a GLM 5.3 versus a Kimi K3. So definitely not commodities in our view. This is an important point. You know, there are many product categories in which model aggregation delivers a greater than sum of parts outcome, and you know, a good metaphor for this is Expedia. You know, it's so much more useful to use Expedia than it is to go to United than to go to Delta than to go to Southwest. You just want a single place where you can benefit from seeing every airline's inventory. Similarly, you know, in coding, we're actually seeing this with Cursor a ton, where you want to do a very frontier model for planning, for example, but then you can use a lesser model for execution, and you really need to have one product harness or sort of product architecture that lets you use multiple models. Creative tools is another great example where you've got, you know, models that specialize in different modalities. So you've got something like an ElevenLabs, which of course is incredible at voice, music as well, and then you've got something like, um, Black Forest, which is doing such an excellent job in kind of video and, and creative direction. And the correct product is to bring all of these together into one shell. And then finally, research and decisions. We see this all the time where, you know, the models are trained with sort of non-overlapping datasets often. So you're able to just get more information by running the same query through many models adversarially and then having a separate model sort of help you converge. This is a place where the application layer really shines because labs, of course, are both incentivized and structurally only able to provide their own in-house models. You as an application sort of aggregator can provide the best of breed. Okay, let's jump into the apps layer. Now, the key point about the application layer is that, you know, intelligence is a primitive, just like buying cloud is a primitive, and what does Salesforce do? It sort of takes the, you know, AWS cloud primitive and turns it into CRM software that delivers an economic outcome for all of their customer segments. The same thing is true of the AI application layer. You know, it's great to have the raw intelligence primitive, but you really need Harvey to turn that into an economic outcome for the legal industry. Similar for somebody like credit unions. It's a really interesting market segment where they're so idiosyncratic in how they want to buy products, how they want the product sort of productized, and, and the shape of the ambition for their market. You know, most credit unions don't wanna, uh, decrease their headcount by half. They wanna double it, right? And they wanna double it while having a s- uh, an economically performant business. So it's just a very, um, specific way that they see the intelligence primitive playing out in their market segment, and the application layer's opportunity is to be the one that kinda delivers that. This is a bit of an advanced concept, but I think an important one. If you look at the kind of way that the evolution of AI use, um, has gone, it's gone from prompting models to putting models in loops. You know, the, the term agent is overused, but agent is just a model in a loop with sort of tools and memory and a few other things. A great example of this is coding. You know, we've all seen this from, um, software companies, which is a bug gets reported, it gets reproduced, a fix gets generated, it gets verified. If it's a low-risk fix, it gets integrated and shipped, and maybe the customer gets an email saying, "Your bug was fixed." If it's a high-risk change, perhaps a human reviews it. But that way, every bug that actually gets reported to the enterprise now gets autonomously fixed through this coding loop. As you start to take that idea and apply it to other parts of the business, things like price optimization, things like procurement, these are very natural sort of business loops that occur that can be fully automated by these models. And then perhaps the most ambitious type of loop is the business loop, which is, hey, you make a change that's very cross-cutting to the business, and the model comes back and says, "Hey, I think we need to open a branch in Tijuana." Now, the model can't do that autonomously, but it can make a change at the sort of surface level of the entire business, which is extraordinary. This is how enterprise automation is gonna occur through AI. And I think for me, coding has just been over and over again an illustration, legal's another great area, of industries, not markets. This is something that Mark says, and he's so right, which is if you look at intelligence as a primitive, let's think now about coding intelligence as a primitive. All of these products are working in their sort of respective areas of the stack. You know, CloudCode does such an excellent job of kind of exposing the raw hardware, so to say, to the developer, all the way up to Replit, which is a great abstraction layer for the average small business owner that's unfamiliar with code. These are variations of sort of pricing, producting, pack- productization, packaging for the coding primitive and intelligence, and all of them are working as a result. So I think a big mental model shift for us is ensuring that we're assessing these as industries, not necessarily simple markets. Okay, and consumer. Consumer's had a really cool couple of weeks. You know, we've been saying for, you know, for three years that this is gonna be consumer's quarter, but I, I think that this might be consumer's quarter. Let's go into it. The things that have actually held back consumer, um, so far have been a couple of things. You know, the first is consumers don't love paying for software. We've learned this lesson, um, over and over again. And unfortunately, uh, unlike the sort of magic of software in the past, um, AI software has marginal costs of distribution and engagement, and the marginal cost can sometimes be very high. You know, I built a, an app I use to help me browse my X timeline, and it costs $250 to onboard a new user. So if I'm a startup founder looking at that, looking at a kinda $250, even with a $0 TAC onboarding cost, it's very hard to make a mass market free product work. That is changing now because of open-weight models, dramatically cheaper and more performant. You know, the second is we've never had an AI native distribution channel. There's no app store for AI, so this actual product cycle for a consumer looks more like Web 2.0, where you have to kinda build the channel alongside the product, and less like mobile, where you actually have this central point of distribution for the entire ecosystem. Then the final point I think is an important one. You know, command line is we're sort of in the, the DOS era of AI, and for this technology and its capabilities to sort of fully be embraced by consumers, we're gonna need the Windows, so to say. So we think there's just a ton of work to be done around product and design craft to ensure that consumers know how to consume, um, all these magical new capabilities. Two things are working. Um, so coding agents i- are, are extraordinary, I know have been discussed. I think it's, it's interesting to think about how they work for consumers. You know, if you think of this concept of the digitally native entrepreneur, if you're not a programmer, the way that's historically shown up is you're a YouTube creator. And there was a whole moral panic that we had, you know, 10 years ago about the kids wanna be YouTube creators, not astronauts. But I would interpret that instead as the kids actually who grew up on the internet wanna build businesses on the internet, and the only way to do it, again, is being a creator. Now with coding agents, you can build a software product that generates $100,000 of revenue a year, a million dollars of revenue a year. Now, these are not venture backable businesses, but it's a sort of mom and pop SaaS opportunity, which is emerging and I think very, very cool for the country. Personal agents, we had this collective moment of excitement around OpenClua, um, in January, and it was an extraordinary sort of composition of primitives, but it never really crossed over into consumer. You know, it was sort of a developer-oriented thing, more of the homebrew computing club kind of energy. Um, we're starting to see with the emergence of GrokBot and ChatGPT Work personal agents being turned into software that consumers can use.

   6. JKJen Kha19:54

      Anish, actually, um, do you mind just pausing on this before

5. 19:57 – 22:16

   Redefining Consumer: When the Plumber Uses GrokBot

   1. JKJen Kha19:57

      we go to the Town demo, 'cause, uh, i- you know, you, you were a founder, um, building in the last era of the, the consumer app experience. And when I even think about it, I was like, "Gosh, how do you even define consumer today?" Because, you know, the plumber that utilizes now GrokBot to completely turn around their business end to end, like is that consumer or is that enterprise? 'Cause like it's very like, it's almost like a, like PLG-led-

   2. AAAnish Acharya20:21

      Yeah

   3. JKJen Kha20:22

      ... movement, but, but it's coming from as a consumer that then's crossover into enterprise and, and particularly like the last era of consumer application is more towards entertainment as-

   4. AAAnish Acharya20:31

      Yeah

   5. JKJen Kha20:31

      ... the way to, to monetize. And so maybe unpack some of that and, and particularly where you've been spending time as a part of that.

   6. AAAnish Acharya20:37

      I mean, our simple rule is if you cannot justify acquiring the customer through sales, which usually means a 15K ACV, you have to acquire them through marketing. We think of them as a consumer, which is most small business owners. So I think that the plumber is definitely the consumer in our sort of investing mind. Entertainment is huge, and there's going to be a bunch of AI native entertainment companies. You know, I would argue Character was kind of an entertainment company. There's been a huge trend around short-form drama, mostly in Asia, and that's starting to come over here. Many of those are generative or sort of generative assisted. So I, I, look, I think entertainment is gonna be massive. Most people wanna spend time, not save time, and the consumer is not that interested in productivity. So that's definitely going to happen, um, and probably worth a separate deep dive. Okay, and I think Town, for folks who have used it, it's, it's just such a magical experience, and you know, this is like the, the number one sort of, um, piece of advice I give to everybody, friends, family, uh, folks in the industry is like, "Please just use the products," 'cause it's so easy to build intuition when you see how they change day to day. And Town is an investment, um, our partner Alex Rampell made. It's a really extraordinary, uh, productivity product, and you sort of see how the compounding, um, improvement of the product through memory advantages it over time. So the first day you use a product, it doesn't know you that well. It's sort of like an employee, a new hire who's just getting up to speed. By day 30, it's able to make excellent assumptions on your behalf 'cause it just has soaked in 30 days of sort of context, memory, and skills. And this is a pattern that we're seeing more and more, this sort of compounding value being delivered to the end customer, showing up as retention in the business and sort of showing up as pricing power on a per customer basis.

6. 22:16 – 24:31

   Town Demo: Personal Agents & Managing Chaos

   1. JKJen Kha22:16

      Yeah, this is a great one because, uh, folks can utilize Town for their personal use case. Uh, and it's a free, you know, trial. They give you, I think, something like 40, uh, credits to start or something around there. Um, and so you can kind of see it once you plug into your personal email, how productive it actually is. Um, on the professional front, I'm always inbox zero. On the personal front, my inbox is like 20,000. Uh, David George is, is probably cringing on the inside here just 'cause that is unacceptable. However, uh-

   2. AAAnish Acharya22:45

      Totally

   3. JKJen Kha22:45

      ... you know, personal life, things are complex, so if you email me on my personal, I will never respond to you. However, I plug Town into it, and like I don't even check it anymore. If there's something important, Town will surface it to me, and also it does all the scrubbing of like subscriptions and all the things that it can optimize, and it's starting to now self-improve upon itself. So like it'll send you emails where it says like, "Hey, this routine is costing this much. Like here's how you could actually save your credits by this." So it's sort of this, this unlock into what starts on the productivity side. And to your point, maybe people won't pay for that personally, but once it starts to get locked in and then expand in terms of the remit, you're like, "Okay, I'll pay the whatever X bucks," you know, just because it helps to manage my life, and I can put it on autopilot.

   4. AAAnish Acharya23:26

      Yeah. It's such a great point, Jen. And like my mental model for this is just an experienced employee, a tenured employee versus a new hire. You know, the new hire may be brilliant and may even cost less than tenured employee, but we all know the value of a tenured employee. They're just able to make great assumptions on behalf of the organization and you. And you know, may- this is a little philosophical, but I think this is where it all goes. Just as we talked about kind of coding loops and business loops for the enterprise, we think there's a set of loops that are informally defined that really, um, sort of lay out a consumer's life. Think of, um, family, friendships, money, health. These are all areas where you have sort of changing information, decisions, agency, execution, and then the loop continues. So we're starting to see some of these sort of loops emerge around self-improvement, kind of health and finance are the two areas that OpenAI is focused on. We've seen a bunch of startups working on shopping, but we think that, like, the kinda way that this ends up playing out is a dramatic quality-of-life improvement for the consumer and, um, and that really follows the shape of past product cycles, where 80% of the surplus is delivered to the vast market.

7. 24:31 – 25:26

   One Dominant Personal Agent or Many Talking to Each Other?

   1. JKJen Kha24:31

      Do you think, Anish, that all of these... Uh, sorry, maybe just going back to the, to the last slide, there's a question here. You know, when you think about these personal agent examples, whether it be Town or Ethos, et cetera, all point to, you know, kind of one assistant having context, but it seems like there's many different options. Do you think it'll end up being, you know, sort of one dominant platform for this personal aspect of your life as, as time management? Um, or will it be like an operating system where you have many kind of talking to each other and kind of configuring on the back end?

   2. AAAnish Acharya25:00

      Uh, uh, the comparative advantage point kinda comes to mind. You know, I think the, the characteristics you want from your CFA are different from the one that you want from your sort of party planner.

   3. JKJen Kha25:09

      Yeah.

   4. AAAnish Acharya25:09

      Um, and then just the surface area is so broad that I think that, yes, there's overlapping bits of context. And I think GrokBots has done a nice job of kind of illustrating this in product, where you have many bots that are pointed in slightly different directions that all coordinate to deliver a globally optimum, uh, optimal outcome.

8. 25:26 – 29:16

   Apps vs Model Companies: Who Captures the Value?

   1. JKJen Kha25:26

      There's a few questions. I'm gonna go back to topics you've covered earlier. So if the application layer captures economic outcomes, how do you think about the competition from the model companies, um, and what will they allow value accretion to happen downstream and, and, you know, are companies at the app layer able to compete with the frontier labs going after that particular market?

   2. AAAnish Acharya25:48

      I mean, I, I think so. Again, I think that we're, we're underestimating the kinda complexity of product pricing, packaging, and how the end customer wants to buy. You know, the way that, um, you know, a teenager wants to consume the intelligence primitive is different than the way an, a marketing executive at a credit union wants to actually consume it. Um, and it's very heterogeneous. So to, to me, it just makes less sense for, um, the labs to move up to the apps layer than to move down to inference. So, uh, you know, uh, uh, yeah, and the kind of permission point's an interesting one. I think if we lived in a world of 2023 when it was one model to rule them all, it wouldn't even matter if you had permission because the labs would just take 100% of your gross margin over time. But now, because you've got many options at all points in the Pareto frontier, you know, the labs have a harder time actually doing things like that.

   3. JKJen Kha26:34

      Awesome. There was a question just on traction. So do you fund anything where there's, um, there, there's no revenue at this point, just given how quickly people have been making progress, or is it extremely difficult?

   4. AAAnish Acharya26:46

      Um, w-we try not to. I, I certainly have spent less time, um, on that strategy. Look, I, I think that the basket is majority investments that are showing some signs of working. Certainly from a product velocity perspective, that used to be something we measured pretty carefully. Like, it's disqualifying to not be showing a live product in a pitch at any stage these days because it's so trivial to build stuff. So almost everything we, we're seeing are showing signs of, you know, sort of some sort of breakout. I mean, my model is somewhat simplistic, where I just sort of look at once you have stat sig sales and product, if we extrapolate from there, um, do we kinda like the price that we have to pay to be a part of it and the risks that we're taking implicitly? And that's, I'd say, the majority of the work that we do. Look, for very talented, experienced folks, we do kinda take a small call option, which looks like a pre-everything round. Um, but that's not the majority of what we do.

   5. JKJen Kha27:35

      Yeah. Yeah. But when you think about the, the, um, kind of competitive landscape on this, uh, consumer has been unloved for so long. Um, are you seeing now this reversion just given it's clear that apps is sort of this next layer of value creation? Like the model sort of layer has been somewhat set, and I say that with a huge asterisk 'cause there might be new algorithmic breakthroughs, you know, kind of, kind of folks coming out from left field as, as we have in the portfolio as well. Um, but do you feel like the shift from the competitive dynamic shifting more towards application?

   6. AAAnish Acharya28:05

      100%. I mean, it's sort of a renaissance for being a consumer builder because you've got this extraordinary primitive that you can work with. By the way, we now have a primitive that can kind of operate in the, you know, emotional, interpersonal domain. You know, you can, like, have a conversation with Claude or OpenAI or K3 and, and feel feelings. And we've had 40 years of technology that really boosted our intellect and productivity, but nothing that kinda spoke to our humanity. So it's a whole different technology surface. It's very wide. I think there are a set of products that labs are just culturally not set up and big tech not set up to go after. You think about launching, you know, a companion product at Google that may disagree with you, that may, um, have sexual innuendo in it. Like, these are things that there's 1,000 committees at Google, um, are designed to prevent. So startups have areas where they're kinda uniquely capable, and then look, finally, the consumer's sort of excited to download new software, excited to pay for it. It's like Christmas 2009 with the iPhone. People wanna tr-try new apps, but unlike the 99 cents days, they're willing to pay 200 a month. So it's sort of a renaissance for consumer builders and, and yeah, I think that ch- things have changed.

   7. JKJen Kha29:09

      Trying to come up with a joke. The, the autists in, in San Francisco are keenly, keenly waiting for this moment for a very, a very long

9. 29:16 – 30:44

   The New Economics of AI Apps: Margins, Compute & Capital as Moat

   1. JKJen Kha29:16

      time. Uh, um, there was a, there was a good, um, a question from Michelle here. You know, how should we think about the new economics of AI apps companies? 'Cause there's a great... there's a debate around the unit economics of, of, um-

   2. AAAnish Acharya29:27

      Yeah

   3. JKJen Kha29:27

      ... apps companies, right? Like, for example, they have may, may have lower gross margins. They're just getting more pressure just 'cause they don't have as much compute access. Um, capital is such a moat in this environment. Um, it's, it's hard to be competitive. So how do you think about the economics of, of underwriting returns in, in, um, in companies today?

   4. AAAnish Acharya29:44

      I mean, David wrote a great post on this. I think that the kind of, the margin topic is a lot more nuanced than it once was. I think it's actually rational in many cases to trade away margin to have wider product surface. I think the very positive part of what's happening in this product cycle is the willingness to pay is extraordinary, and that's why the exercise that we often do with founders is, like on the consumer side, for example, is if $20 was the historic ceiling, what's the $200 a month SKU of your product? And in fact, what's the $2,000 a month SKU? Like, what's the Birkin bag of software? I think we're gonna have this luxury software. We're already seeing willingness to pay for it. So the margin topic is more nuanced, but the willingness to pay and buy is higher than ever. So, you know, it's a little bit of fog of war, but we're, we're thinking about all those topics.

   5. JKJen Kha30:26

      Anish dropping Birkin bag frame jeans. Like, I had no idea you were such a fashionista.

   6. AAAnish Acharya30:31

      [laughs]

   7. JKJen Kha30:31

      This is like, your, your bot is helping you get up to-

   8. AAAnish Acharya30:33

      Thank you

   9. JKJen Kha30:33

      ... to seed here, my friend. Uh, for a guy I only s-

   10. AAAnish Acharya30:36

       You're telling my secret.

   11. JKJen Kha30:37

       [laughs] For-

   12. AAAnish Acharya30:38

       An investor is a good steward of capital, okay? That's all that I am.

   13. JKJen Kha30:41

       [laughs] For, for a guy I only see in quarter zip ups, I'm

10. 30:44 – 32:20

    Who's Actually Building Apps Today? Founder Archetypes

    1. JKJen Kha30:44

       just saying. Uh.

    2. AAAnish Acharya30:45

       [laughs]

    3. JKJen Kha30:46

       Um, okay. May-maybe one question for you on, just on the, on the founders, 'cause I, I, I don't know if you remember this conversation. This was probably five years ago or so, um, where most of the founders you saw, saw more, uh, diversity in their background, in part because the software and technology was way more sophisticated. So you had a lot of program managers spinning out of Google, for example, and starting a company, et cetera. What are the type of founders you see building in apps today? Are they-- Do they tend to lean, you know, more technical, more researcher derivatives? Are they product managers? Like what, what kind of archetype are you seeing at least the early innings of apps come out from the woodwork on?

    4. AAAnish Acharya31:23

       Yeah. Yeah, less MBAs, more researchers. Um, and they, they both have their kind of strengths and weaknesses. I think the business sophistication of the founders we're seeing today is lower, but the kind of technical sophistication is dramatically higher, and the technical sophistication is kind of upstream of all the good things that happened. You know, business sophistication can be kinda taught and observed, but technical sophistication, typically not. So we're definitely seeing a, a much more technical, kinda earlier career founder, but the things they're doing are extraordinary because they don't have any sort of preconceived notions about what's possible. And so much of what holds back senior founders that don't quite get to the other side of this product cycle is, you know, they're not close enough to the technology, and they've got an idea that's rooted in the past of what the ceiling is. And I think the best thing about these young founders is they assume everything is possible. You know, we are at an offsite where Ben was saying that the, the biggest risk in the past were the ideas were too big, and now the biggest risk is that the ideas are too small. But I think that's sort of illustrative of the different founder archetypes.

11. 32:20 – 34:05

    Why Giving Founders Too Much Money Isn't Fatal Anymore

    1. JKJen Kha32:20

       Yep. Yep. And maybe on that similar thread, it used to be that, that if you gave a founder too much money, it would wreck the company.

    2. AAAnish Acharya32:27

       Yeah.

    3. JKJen Kha32:27

       Because the founder almost always has way too many ideas and is a visionary and doesn't have the talent to actually commensurately land with all those ideas. Um, and we're seeing a whole new paradigm on that. M-maybe unpack that idea a little bit more just, uh, 'cause it was such-

    4. AAAnish Acharya32:43

       Yeah

    5. JKJen Kha32:43

       ... a huge theme of that offsite.

    6. AAAnish Acharya32:45

       Yeah, I mean, for sure this was the historic wisdom. You know, why didn't we give every seed company $20 or $50 or $100 million? You know, it wasn't just the kinda risk reward, but rather typically the constraining factor was they just didn't have enough talented people to work across $20 million of product surface at the same time. They really had to focus on one idea at a time, and the capital was a great way to enforce that focus. What we're now seeing is you could make different sort of product and model trade-offs through more or less capital, and there is a case for a company that raises $100 million, uses it productively and in a focused way, and is able to deliver a different value proposition, um, than the very same team would be able to do with 20. So I think that the, again, like the sort of just as we talked about the sort of fog of war around margins, I think this question of what is the optimal seed round size and how much capital can you put to work effectively is a much more nuanced topic. I mean, it- it's sort of this embarrassment of riches, but I'd rather have this problem than the problem we had five years ago, which is, "Hey, my fintech company is indirectly subsidizing their customers through weak underwriting, and we don't know the path home."

    7. JKJen Kha33:46

       Yeah. Yep. Yeah. The Chris, Chris Dixon model, which is you always want the problem of supply, not of demand, right? Right now-

    8. AAAnish Acharya33:53

       [laughs]

    9. JKJen Kha33:53

       ... we have to fix the supply part, right? Uh, but the demand is like, uh, so, so abundantly there that, that, uh, undoubtedly that will, um, the supply part will, will get fixed. Um, maybe I'll close on this one last, uh, question from Mozfa.

12. 34:05 – 35:46

    Go-to-Market for Startups Selling to SMEs

    1. JKJen Kha34:05

       So, um, double-clicking on the SME sector adoption of AI. So unlike large enterprise, the friction of adoption is much less because they require less change management. I agree with many of that, but, uh, s- you, you [laughs] not, not all. Uh, small, small, uh, medium businesses sometimes have, uh, more habit change that you gotta work through. But the question is, how do you see the go-to-market playbook for startups targeting SMEs, and has that changed in the a-age of AI?

    2. AAAnish Acharya34:31

       I mean, a lot of it for existing SMEs, I think it's the same channels with which you historically reach them. I actually think that one of the interesting things about marketing in the age of AI is that all of the sort of existing networks have been so trained on the methodology of building new networks, that they're very careful to ensure no one does it on their network. So Instagram, TikTok, X, it's very hard to build a new sort of distribution channel off the backs of an existing one. So what founders have to do is actually build a product that has the original network effect, which is word of mouth. So we're definitely seeing more of a focus on word of mouth, yet ask the kinda old channels for reaching SMEs are still there. Actually, I think the most interesting segment of the market though is sort of new business formation, which is, by the way, at an all-time high. I think it's the highest it's been outside of a peak sort of moment during COVID. These are people who would've never otherwise been SMEs. It's not the sort of 55-year-old plumber. It's a 25-year-old who previously would've been a YouTube creator and now is building SaaS for their, you know, neighborhood or their city or their high school or whatever else it is.

    3. JKJen Kha35:30

       Yep. Yep, yep. Awesome. Well, thank you so much for this, Anish. It's always great to have you on. Uh, now I know you're a fashionista, and we're gonna be clipping that endlessly, uh, on the socials. Um, but, uh-

    4. AAAnish Acharya35:39

       Great

    5. JKJen Kha35:40

       ... but thank you for that, and if folks have any questions-

    6. AAAnish Acharya35:42

       Thank you

    7. JKJen Kha35:42

       ... you, you know where to find Anish. Um, and, uh, we'll follow up here for some of the questions we weren't able to get to as well.

Episode duration: 36:01