a16z Partner: Hollywood Is Secretly Replacing Films With AI

I think folks would be surprised by how many TV shows and movies today already have some element of AI being used. I spend a lot of time with studio execs, and Amazon and Netflix have already both announced programs for fully AI-generated animations. It's 90, 95% of the quality of filming, but it's so much cheaper and faster and easier. And I think the mass market consumer does not care that it was made with AI. But many of the people who were early adopters of AI video were not inherently amazing storytellers. And so they were making things that were novel but were not necessarily the most compelling narratives. That is the major challenge. Now what we're starting to see is professional creatives who have worked in the space, seen the opportunity in this. And so we have people who know how to tell a narrative, which means we're gonna get an explosion of really gripping stories. It sort of just unlocks all of this crazy, weird, interesting ideas that people have in their imagination that they could never make before. You see a lot of folks who have either started kind of full new new AI-native studios or who are just individuals running a brand and scaling their own content, creating narrative stories, turning around things in hours or days. And so most people don't wanna be competing on more of the like lowbrow slop content. But I think in terms of what creators should do, it really depends on like—

Justine, welcome to New Economies. We've been in touch kind of over the years on DMs, but this is the first time we're meeting. And wow, so much is happening across AI, new media, and now the latest trend, AI microdramas. So we have so much to unpack, but big welcome to the show.

Thanks for having me. Yeah, excited to chat about all those things.

So I feel like we're at this incredible inflection point around generative media. And so much has just been happening really in the last 6 months. We've had this explosion, new term, new media, which is kind of just replacing the creator economy. We've had obviously so much change in AI. And like we said at the beginning, this new incredible form of entertainment, AI microdramas. Help us unpack what is the current state of everything happening in generative media? What's your hot take?

Yeah, I would say I've been following the generative media space for a couple years, actually really got into it a couple months before ChatGPT came out when the first Stable Diffusion models came out and you had folks like Levels.io building really cool products on top of them, like the original profile picture, AI profile picture generators. And it's really been remarkable to see the progress we've made since then. Like we went from not very good images to not very good 2-second videos, now to like full coherent 15-second videos that people can generate, let alone, you know, all the advancements that have happened in image and voice and 3D. I think we've— we're at a really interesting point where the model quality has gotten good enough that you can actually make things that are compelling to watch and that carry a storyline versus things that are just like, Interesting for the novelty value of like, whoa, AI can make something that almost looks real. And so as a result, we've had a ton of progress happening both on the model layer, obviously, and those improvements will continue to happen, but also on the product layer. Like when the models are finally good enough for people to use, you kind of have to create the right harnesses and workflows for people to actually leverage them, especially people who You know, aren't full-time prompting all of these models. And so they don't know what model to pick for which use case, how to write a really good prompt, things like that. So to me, it feels like where we are in generative media right now is really the entire ecosystem maturing, both consumers, but also professionals, studios waking up to the fact that these are— models are finally good enough to use, and the product builders realizing that there's a lot of opportunity to create like the interfaces through which we use them.

And the content which is being produced by some of these AI models for AI microdramas, some of the storytelling is so good and the accuracy is incredible. What do you think makes these AI microdramas— and we should probably explain actually what AI microdramas are— what do you actually think makes these AI microdramas so cool to watch?

Yeah, so I describe microdramas as almost like the TikTok-ified version of soap operas. So they're usually like pretty dramatic, Salacious, like interesting pieces of content that are relatively short. They're typically vertical video. They play both on places like TikTok, but also on apps like ReelShort or DramaBox. They really started blowing up first in China, like a few, a few years ago. And now the, the microdrama market in China is larger than their domestic box office. There's everyone there is just, you know, subscribing to these apps, viewing all of these microdramas. And I think they started to take off in the US quite a bit last year. Like when you look at the App Store data, they were some of the fastest growing and the highest monetizing apps in the country, which is pretty rare that you have a couple apps in a new category like that grow that quickly. That's kind of when I started paying attention. And then I think the AI element of it came in when, you know, you look at this sort of content and. The microdramas often, similar to soap operas, have like a relatively limited cast. They have relatively limited sets. Like, it's not a game show or competition show where you're filming all over the world. Like, it's not a big blockbuster movie. And the thing that really gets people interested is the storyline and the suspense and the drama and like what's going to happen next. And so it doesn't need to be a Hollywood-grade production to get people interested. And so I think a lot of folks looked at that and said, this kind of content is sort of a perfect fit for what we can do with AI models today.

Some of the dramas I have seen, they are so gripping. I don't know if you saw this one around Dario from Anthropic falls in love with this girl from China. You may have seen it. I've heard of it. So I met the creator who designed it, and I thought the storytelling is incredible. We'll link it in the show notes. And then I later found out the girl who's actually in the microdrama, it is a true story. And the whole story came to light. And I thought this storyline storytelling is so incredible. But there are 2 sides of these, I think, AI microdramas working right. You have the one side where it's incredible storytelling, you have incredible designs, low cost to make, which is absolutely fine. But then you have the other side where it's well designed, but the storytelling is just not as good as it could be, right? Where do you think this actually like will go next in order for AI microdramas actually to stay around?

I mean, I think the good thing about it is there are a lot of— so I have this I have this thought for a lot of AI video, actually, which is many of the people who were early adopters of AI video were not inherently amazing storytellers. They were early adopters of technology. And so they were making things that were novel, but were not necessarily the most compelling narratives. Like, I put myself in that category 100%. Like, I never trained or learned how to shape a really compelling narrative. Now what we're starting to see is, you know, professional creatives who have worked in the space, who know how to tell really good stories. Maybe they're not like what we would think of as a traditional professional, professional creative, but they've been a creator for a long time and they've learned through honing their craft how to tell a story all the way to the big studios are seeing the opportunity in this. And so we have people who know how to tell a narrative seeing the opportunity in the microdrama space, which to me means we're going to get an explosion of actually really gripping stories, like super interesting content beyond the AI element of it. I actually think we might be at this point already, but there's going to be a lot of microdrama content that you don't even know is AI. And people don't even care. It's— they're just there for, for the story.

I'm happy you brought this up. Because honestly, 6 months ago, I was like, I'm never gonna watch an AI-generated program. I'm never going to watch an AI-generated movie. Now I'm like, actually, if it's gripping and it's interesting, I'll watch it.

Totally. I think the quality has gotten a lot better too. I spend a lot of time with studio execs and sort of, you know, who want to get involved in the investing space or are curious sort of what's happening in generative media from an early-stage perspective. I was just down at AI on the Lot. In Hollywood and like sat through a bunch of panels of how people are actually using this. And I think folks would be surprised by how many TV shows and movies today already have some element of AI being used and how many of the big studios are considering fully generative content. Like Amazon and Netflix have already both announced programs for fully AI-generated animations. And I think the bigger studios are afraid of like PR risk around things like that. But are absolutely exploring very similar things.

Yeah. And when I talk to creators as well, execs, founders, people who are interested in this space, and also just wider consumers, some people are like, isn't this just going to replace Netflix or movies or actors, actresses? And I'm like, no, I just actually think it's going to be a different type of content and entertainment we consume. Do you think that's accurate? And is that where we're going?

I think it's going to be a mix. So I think we're going to get new types of content. Microdramas are a great example of that, where I think in the US, they're really taking off with AI now, making them much cheaper and easier to produce. I also think we're going to see AI used in a lot of traditional production pipelines. And so many of the big studios They're not like, we're going to replace every actor and actress and do an entire movie that is fully generated. But they are like filming people on soundstages and generating the background with AI or generating like the big scary monster or the giraffe or whatever kind of crazy animal with AI instead of either, you know, having to go film that somehow or using CGI. So yeah, I think it's going to be a mix. I think we're going to see new entertainment formats. And then I think we're also going to see folks using AI in traditional entertainment formats like TV and movies to bring down the cost of production.

The opportunity also for creators, and especially individual creators, maybe even with a small team, right, is insane. There are a couple of AI microdramas. There's one in China which was published not too long ago, and it's had like 100 million views in the space of, I think, 2 or 4 weeks. Which is crazy. You had this AI microdrama I mentioned around Dario from Anthropic falling in love with this girl from China. This thing has had like 2 or 3 million views. And if it's that gripping, creators can start monetizing that, right?

I think a lot have, honestly. I mean, on X, on Instagram, on YouTube, you see a lot of folks who have either started kind of full new AI-native studios or who are just individuals running a brand and scaling their own content, creating narrative stories. Folks like SyopAnime or Gossip Goblin or Charlie Curran are putting out insane numbers in terms of who actually views their content, the consistency at which they're releasing content, and how their audiences are growing over time. And when you watch the content, it's immediately clear that One person or small teams could have never created this before AI. It's whether it's animation or anime or whether it's, you know, photorealistic live action. I'm not sure what you call it when it's AI, like would've required massive budgets and months to do. And they're turning around things in, you know, hours or days.

We always have these crazy ideas for apps or websites, but sometimes we just don't know where to go to get going. That's why we have partnered with Hostinger for today's episode. Hostinger are the perfect tool for builders and founders. If you have an idea for a website or app, you can go over to Hostinger's latest Vibe Coding Tool to turn your idea into reality. If you need a domain, and of course that email address to help us look good and amazing in front of our customers, Hostinger has all of that covered for you. And that's why they are giving our fans and audience 10% off. All you have to do is head over to hostinger.com/ New Economies, or enter code NEWECONOMIES to redeem the code. Go and build the thing you've always wanted to do. It'll be worth it. Now back to today's episode. Before we started recording, we got chatting around. We've actually like had a go creating our own AI microdramas, and it's actually pretty good fun. Have you had a go at creating your own, and how have you found the experience?

I try to create all sorts of AI content, largely to test the tools. Like, I try not to delude myself into thinking that I have some incredible story that I need to tell that people will want to watch. But I do like testing various things across different styles, like, you know, anime, 3D animation, photorealistic, just to get a sense of how the model quality is improving and then also how the harnesses are improving. There are things about it, like we were chatting how, you know, the models have gotten so much better, but they're not perfect. And so you can spend a lot of money generating over and over and over again to try to fix a small mistake or to try to get a scene perfect. But at the same time, we've come super far, like with things like character consistency, with, you know, the physics of the real world working in most scenes, even if it's not perfect, with video models being able to follow prompts around camera control and transitions and things like that much better than they used to. So it's much more fun than frustrating. I remember, it must be like 3 years ago now or whatever, I was trying to create a super basic— it was almost like a slideshow animation of Calvin and Hobbes through the years. And oh my gosh, it took days. And the output was terrible. At the time, I was like, this is incredible. I made this with AI. But it was a very frustrating process. And now I would say it's a lot more fun.

Exactly. I think we're still at the early infancy stages, right? I was saying before we came on, I started creating our own microdrama around the Nike story. So my— the idea was, anyone who's listening can take this idea, right? All these incredible companies we talk about on a daily basis, they all have an origin story, but nobody actually knows a lot of the time what the origin story is. How cool would it be to have an AI microdrama series called The Origins or The Tech Origins or something around how all of these incredible companies got created. Anyway, I started playing around with a couple of these AI models and tools around how the Nike stories, uh, you know, journey started, the origin story.

Yeah.

And the storytelling, I think, is pretty good. The design, the aesthetics of the actual show is so good. But maybe like where it's still lagging is the camera lens, the lighting, how quickly like People are walking. Every, every maybe AI microdrama needs some romantic story to it. So maybe the romantic story for the Nike story could be around when the founder meets Michael Jordan for the first time. Michael Jordan falls in love with this incredible shoe. You can make it so good. And honestly, I only spent 2 or 3 hours playing around with it, and I think it was quite good. And that's in 3 hours.

That sounds incredible. I think also it's, it's fun to think about different iterations of microdramas like that. Like, I think historically, when people think microdramas, they think like soap operas or just kind of, you know, crazy, suspenseful, pure romance. But there's an opportunity to tell so many different types of stories like what you're mentioning. There's an opportunity for it to be somewhat educational. Like, there's so many different sorts of content that are unlocked here.

And it's pretty interesting to see the consumer adoption of this. I was actually just in China last week and seeing the adoption of technology and consumer social apps is so interesting to see versus what happens in the West, right?

Yeah.

How do you see the adoption of these AI microdramas actually sticking around? And is it just— are we just going through a phase? Is it just going to die or is it going to last?

I think from a consumer perspective, my instinct at least is that people don't actually care what it's made with. You know, at first, maybe the folks on X or Reddit or wherever, the people who care about technology and want to follow the progress may watch something because it's AI and may follow along because it's AI. The mass market consumer does not care that it was made with AI and isn't going to watch it for that reason. And so the question is, do we hit the quality bar where it makes sense for all microdramas or many microdramas to be created with AI? Because it's, you know, 90, 95% of the quality of filming, but it's so much cheaper and faster and easier. And I think we are quickly approaching that point where some microdramas can be entirely made with AI. Some are made with a combination of, you know, human filming and then some sort of AI editing, adding backgrounds, adjusting things, dubbing, that sort of thing. So yeah, I don't, I don't think this is a trend just because it so massively shifts the, what the supply curve can look like. Because now for $100,000, you can make so many more episodes than you could make before. It just makes a ton of sense for the producers of the microdramas to be using AI because historically, like one of their biggest problems was just having enough content, having the volume, having the diversity of content to be able to get viewers in and and retain them on the platform.

I think it's also going to be interesting to see the type of content— it doesn't even have to be microdramas in this case— that individual creators create versus the micro studios, right? Like we're saying, it is, it is still pretty expensive as an individual creator to create an AI microdrama. Like, we didn't spend too much creating our own, maybe, you know, under $500. But that's still a lot, and it still needs multiple iterations, right? Do you think the cost for individual creators will go down over time where you can maybe knock something out for a couple hundred dollars?

For sure. I think the trend we see with all of these models is that over time, the quality that you can get for a certain price goes up, which is great as more and more new models come out. I think the frontier models, like right now that's SeaDance 2, will always be very expensive. I think there's an incentive, uh, with more and more competition in the space for the model providers to both continue improving their models and also to make their models less expensive over time so more people can use them. I think Google did this with V03 when they released V03 Fast. And we saw yesterday SeaDance is now doing this with SeaDance too fast, which is much cheaper and faster. And so my instinct is we're just going to keep seeing that the costs go down over time. And then, of course, hopefully we'll have better and better kind of frontier models that will likely stay pricey for a little while. Yeah.

And, you know, I feel this has all changed in maybe 6 months. I think if you asked a regular consumer 6 months ago, would you watch an AI-generated movie program, whatever you call that episode, they would probably say no. Now I think they don't mind if it's good quality.

Right. I mean, some of the most viral clips I've seen from the World Cup are AI-generated hilarious things of an entire team forming a circle and moving down the field to score a goal, or the Brazil team building a brick wall in front of their goal so no one can score. It sort of just unlocks All of this stuff, like these crazy, weird, interesting ideas that people have in their imagination that they could never make before, that they can then share with the world. And a bunch of other people can be like, whoa, this is super cool or interesting or weird or entertaining, kind of whatever it's intended to be.

Exactly. We're both in this new media ecosystem, right? Spreading so much around what's happening in the category and the ecosystem. And one trend I keep hearing, I'd love to get your take, is around what is AI slop? And to turn this around, should creators be thinking about, is their content timely or timeless? I know your colleague Eric did a great tweet about this recently where really, you know, the terms came from. Is that how creators should be thinking about creating content to avoid potentially being accused of, oh, this is just all AI generated.

The slop concept is very interesting to me, um, because I think it came out around AI, but my take at least has been there's a ton of human-produced slop. Like, slop is more about the quality of the content and the narrative and like what it does to your brain versus whether it's made with AI or not. Like, I think you could argue that a ton of the like clickbait YouTube, like insane slime challenge, whatever shorts that, that people were posting far before AI are like human slop. And AI slop is similar to me. It's like, I don't know, I guess the way I understand slop is like there's no really underlying— I don't even know if compelling it, because some of them do have like compelling narratives. But they just like trigger this primal part in your brain that like makes you kind of keep watching and be hooked, even though you don't, even though it's not the highest quality story or narrative typically. I think, yeah, like we shouldn't all spend our time watching AI slop, but I think also we shouldn't pretend that slop originated with AI because there was plenty of human-produced slop content before that. I think in terms of what creators should do, It really depends on what sort of creator you want to be. I mean, there's a lot of creators who optimize for kind of near-term impressions and growing their audience and jumping on every trend. You see this in non-AI spaces. You see this with traditional YouTube creators. Some are on every single trend. One week they're doing the 10,000-calorie challenge, and the next week they're doing the pour buckets of water on random people down the street. Hopping around between things that are definitely not timeless, whereas others take a long time researching and filming really in-depth kind of deep dives and videos. I think usually for companies, it makes sense to be more on the timeless side of things, or at least to have sort of thoughtful content that articulates your thesis and who you are and who you want to be. But there are a lot of creators for whom they can grow an audience quickly and monetize really well by producing, I guess, more of the like lowbrow slop content.

Are we just gonna get to a stage where we don't actually know what is AI generated or not? And perhaps for a lot of these social platforms, the likes of Substack, maybe it's the slightly larger platforms, maybe some new ones which, which come into play, should creators start citing if their content was created using any of the tools they use on a daily basis for better transparency?

I think my view on that has been it's super hard to say what is AI or not. Like, I remember when people ask this question, I saw a Reddit post like a year and a half ago where like Steam or one of the game marketplaces was requiring games made with AI to be labeled AI. And one of the developers was like, well, I used like a tab complete model and cursor when I was coding parts of this game. Like, does that mean I have to label it as an AI-generated game? And I think Steam's intention was like, oh, if it's AI-generated assets or images or music or whatever. But like, it's just such a slippery slope of like, oh, you know, What everyone assumes AI means. Like, people will look at that, many people I think will look at that and assume this is fully AI generated. And in many cases, it's like a collaboration between human and some sort of model. I think my instinct is like the vast, vast majority of content will end up being partially AI produced or AI edited in the future. So there's probably no point in even labeling it because that's just going to be the default. That said, I am surprised on platforms, especially like Instagram, how many people don't recognize clearly an AI video. So I know the social platforms are putting more disclosures and things like that over that sort of stuff.

It's becoming much harder for sure on what has been AI generated. I mean, for videos, I think it's hopefully fairly easy to figure out, especially for newsletters, right? I saw this stat the other day someone put together where if you look at the top newsletters globally on all platforms, something like 20% of the top newsletter writers, all of their content, 80% of their content is AI generated. And then it's like you get to that phase where it's, if you're a top newsletter but it's AI generated content, I'm not sure where you stand with that.

I think I have a controversial take on this, which is I don't actually care if the person is choosing to read the content and the person is choosing if they find the content meaningful or not and whether it resonates with them and whether it kind of speaks to something in their life. And so it's sort of in the eye of the beholder. I think folks who are— and we've already seen this, people who have a popular newsletter and have pivoted to just putting out purely Claude generations that are like not really in their voice and like don't make a ton of sense get called out for that, like as they should. But I think there are some people who, you know, I've even been surprised by it, find a lot of meaning and really resonate with certain things, certain AI outputs that people get by, you know, going back and forth with Claude or GPT or whatever model about a certain question and then publishing it. I think some people get really offended when they see AI-written output. I just sort of skim it and decide, is this meaningful to me? Do I want to read it even if it has an AI tinge? Is there still some element of sort of, I can understand what the author wanted to say and why? And if so, I'll still read it.

Yeah, my personal take has also changed fairly recently where if it's a creator who I respect and I'm going to that creator Because they're sharing the latest tech trends, or is the latest startups, or what's happening like in healthcare, for example. I don't really care if I'm still going to get the same insights what the creator is trying to say. Great, I don't mind. Maybe 6 months ago would have been different.

Why do you think your viewpoint changed on that?

Because there's 2 points. One is so many of these creators are using these AI models and tools anyway to help them with their work, which I think is totally okay, right? They're very small teams, and I guess they need as much support as they can. And secondly, it's just where the world's going. As long as the information I think is accurate, and what the creators are saying is still accurate, then I think it's okay.

I agree. I mean, I think the thing that I don't like, to play the devil's advocate, is I don't know if they still do this, but for a while, both Claude and GPT had this terrible habit of being incredibly long-winded and in like a way that like no human would be when explaining something. And so you could clearly tell if something was AI-generated because it would take 3 paragraphs to say like 1 sentence worth of content.

Yeah.

And I thought that is like very annoying. But I don't know, a lot of what I do when I write blog posts and stuff now is like, it's really hard to start from a blank page. And so often what I'll do is I'll create some sort of outline of what I want to say. I'll give it to both Claude and GPT. I'll see what angles they come up with. I'll paste them both in a doc. I'll take parts I like from each of them, and then I'll rewrite a bunch of it myself. And, you know, the end output is probably like 70, 60 to 70% written by me, like 30 to 40% written by AI. But I'm actually like, it's the, the process of writing is no longer as painful because I'm not starting from a completely blank page.

Well, exactly the same. We use all of the models to check for spelling mistakes, which I think is always a good thing. There's always like one or two things which you miss out, right? We ask each of the AI models to rate the newsletter edition out of 10 based on all of our previous content. So if Claude is saying 6.5, Gemini is saying 9, and ChatGPT is saying I think it's something like 8.59, you know Perhaps that you're on the right track for something, right? If it's like all of our designs, we do these beautiful like market maps similar to you. All of those are human generated for now, but we would ask it to rate it. What do you think of this design? How should we make it a bit clearer? Should we move like the boxes around? When you've got 200 companies on a market map, it does get a bit hard.

Yes.

But it's actually a super useful sparring partner. on what great content could look like.

Yeah, I find it super helpful. Like, I've also given it sort of presentations that I'm giving at our like GP meetings or wherever. And it's really sometimes when you get deep in, in writing something or creating a presentation, you have so much context in your own brain that it's hard for you to imagine, does this make sense to an external person looking at this for the first time? And I find the models are really good at reviewing something fresh and saying like, there is no coherent narrative underlying this, or like, you really need to move this part of the presentation before these other parts, because without it, the rest of the parts don't make sense for someone who's not super deep in this space. And I find that to be super valuable, and I think will actually really improve how a lot of people communicate, because that's a really hard skill to be able to do on your own.

The hardest part also for creators like us, and maybe you agree, is there are so many tools out there, deciding which ones to stick with and deciding which ones to try out every week. It's opportunity cost, it's just time. At the end of the day, you just got to decide, I guess, what works best for you.

It's hard because you get on something that works well for you. And then something else comes out the next week that you want to try that looks even better. And then your original product upgrades. And so I don't know, it's difficult. I think we— I try to— I think I'm lucky that since I'm investing in a lot of early-stage stuff, it's helpful for me to use both all of the early-stage products, but then also know sort of what the incumbents, which at this point are like the OpenAI and the—

Yeah.

Anthropics and the Googles are doing and releasing, so I can know how the product experiences actually compare.

We have a lot of founders who listen to the show, and I think a useful question would be to talk around what are the hardest parts of building a generative media type of company from a startup point of view? And yeah, any characteristics you've seen there on All the hundreds of companies you've met?

Yeah, generative media, I think, is a great space, and it's also a really competitive space. So I think to me, the hardest thing right now has been figuring out who you want to serve and how you're going to be differentiated, largely because it's gotten pretty expensive to train foundation models that compete with folks like OpenAI and And Google, not as much Anthropic because they do sort of less on the media side. But there's now a number of startups that have raised a bunch of money across different modalities, as well as the big tech companies that are training models here. And so from a research perspective, it's harder than it used to be to train a new image or video or audio model from scratch and be a competitor in the market, which is not to say folks shouldn't do it, but I would say the bar has raised in terms of what folks expect in terms of differentiation, what your model can do that others can't, that sort of thing. I think a lot of folks are now trying to compete at the app or workflow layer. And at first, the applications for these models were incredibly simple. It was like, you know, a text box prompt input, and then you would get the output. And now there's a bunch of great services that host a ton of different models, and they all offer, you know, text input to output. And so most people don't want to be competing on just offering the lowest price for every model, or, you know, offering the fastest inference or things like that. And so they're really trying to think about, how can I serve a certain user base better than anyone else can? Like, what can I build? Product-wise, how can I create features that really resonate with a particular audience? How can I go deeper into a particular type of creator or professional's workflow so that they have a real reason beyond price to stick with my platform? I would say that is the major challenge that everyone is thinking through right now. And it's worth probably having a point of view on that if you're starting a new company today in the generative media space.

There are so many incredible creators out there who you can sell to, but also figuring out, you know, where the opportunities are because things are changing all the time is the hardest part, right? Where do you think the untapped opportunity is for a founder to build a defining company in this category?

I mean, I think a lot— most of the companies in this space initially targeted like the AI-native creators. You know, folks like us who were early adopters of all of these tools, who were using and sharing and posting about it, who were reachable on platforms like X and Reddit. But I think there's a massive market of other people that are largely untapped, who are less technically sophisticated, who don't know every single tool out there. And in some ways, that's a benefit because if you get them on your tool and you get them to have a sticky workflow, they are like, they're less fickle. They're less likely to turn and go to whatever they see the next day. I think there's a lot of opportunity in consumer that's honestly untapped. Like sort of the social layer of generating and sharing stuff is still almost entirely done on the existing incumbent platforms versus, you know, creating really fun and interesting and compelling new playgrounds where consumers can make and share this sort of stuff. I also think types of content like microdramas are a great example where you need really specific things like, and I, saved ideas of characters with certain voices, saved sets, like a sense of the right timing and pacing for a microdrama in particular. Like that sort of knowledge and those features are going to be different than a general video product. And then I think there are also a lot of opportunities in like vertical industries, things like architecture, things like design. Marketing and ad industry is huge and is just starting to adopt AI. And so I think going into those markets that are not necessarily like the most AI native is an enormous opportunity right now.

It feels like there are so many opportunities founders could go towards. I almost feel, is there going to be like a Roblox equivalent for AI microdramas? Maybe all of these AI tools where you can create ads in like a couple of minutes, I think has been so interesting to see. We mentioned earlier around like presentations, creators have to create slide decks and partnership agreements and stuff. That's kind of all now being done. I still think maybe where there is opportunity is actually just the typical creator stack, like running a creative business and just like all the day-to-day stuff you have to do, like reaching out to podcast guests, potential sponsors, partnership agreements. Like we still have to do this, right? And it feels like maybe there is still some opportunity there, but it's, it's pretty hard.

For sure. I mean, yeah, there's so much work that is done beyond just the pure content creation side of it. I think a lot of the things you're talking about is more sort of the business, the tasks that today humans have to do, whether it's, you know, research and outreach, or whether it's creating and sending invoices or reviewing legal contracts or things like that. And obviously, you know, big companies have a bunch of people who do those things, but For most creators, like they can't afford to have big teams who are taking care of all of that for them. And so every minute they spend on that sort of thing is a minute that they can't spend on creating content, which is generally what they like to do. So I don't, one of my takes is that AI agents are going to be particularly powerful for individual creators and small businesses just because they're so much more constrained time-wise and resource-wise. And it's going to be a massive step function improvement for them when they can spend much more of their time on their creation, their business, whatever they like to do, and less on all of this administrative and logistical type stuff.

Have you tried any cool personal AI agents that you're really excited about? And do you have one?

This is going to sound like I'm shilling for an investment, but I was actually a user of this months before we invested. There's this new product called Town. It connects to your email and your calendar, and what I really like about it is it ingests all of this information about you, and then it proactively suggests things it can do for you. So most AI products, you're going to it and saying like, hey, can you set up this recurring workflow and describing exactly how it's done. Town looks at everything you've done and creates the description of what it needs to do and then asks if you can do it. One, one great example is I often get warm intros from other investors or other folks to companies. And one day, Town was just like, hey, I notice every time you get one of these, you, like, send an email back thanking the referrer, you move them to BCC, and then you loop in your EA, Lauren, to schedule a meeting. How about I just do that for you every time one of these warm intros you've agreed to comes in? Or like, I've noticed Friday nights at 11 PM, you are forwarding in receipts that you get to your— to Navon, your guys' expense provider? How about every time you get an email that's an invoice or receipt, I just auto-forward it in for you and then sort it so it's not clogging your inbox? I think things like that, I'm really excited for them to reach more people.

They just announced a fundraise, right?

Yes.

This is almost what I feel we've been waiting for, for the last 2 years.

Yeah.

And it feels like consumer tech intersecting AI. Consumer tech has always been, there's no opportunity, the returns have been super bad, but it feels like it's back. Where are you most excited about seeing consumer go next?

I always say it's super hard to predict consumer. It sort of can only be observed, just because it's like consumer behavior is, it's very difficult to say, like, you know, even if you look at things like a TV show or music or a movie or a game, Like, what is gonna be the next major hit? I think you can see early signs though, just like, you know, you can see early signs of like an indie songwriter, you know, rising up in TikTok and then eventually they become a big star. You can see signs from early adopters of these sorts of products of what the future might look like once everyone can get ahold of it. Like, if something is a magical enough experience for the early adopters, it will likely continue to grow and spread. I mean, that's, that's why we invested in Town. Like, Me and a number of other people on the team used it and we're just like, I think we see similar things for like texting, being able to text AI agents, for example. Like a lot of people do not like having tons of apps on their phone and having to open an app and, you know, wait for it to connect and then send something to AI and then wait for the response to come back versus, you know, you're just on your phone and you're lost or you don't know when your next meeting is or. You know, you get some important email you missed and you can text your AI assistant or it can text you with products like Poke or Ollie, which is doing this for more kind of family and household management. And I think we saw the first wave of a lot of that kind of stuff with like OpenClaw, honestly, and the Hermes agent, where really technical people were setting up super interesting things and finding a lot of value from it. But The vast majority of users are not going to like get a Mac Mini or install the like amphetamine thing on their Mac to keep it running all of the time and, and go through the whole process of installing some open source project to set up their assistant and connecting the API key. Like there's still so much friction around that sort of thing that we have not yet reached the point where it's come to everyday consumers in a cool, well-packaged product. So I'm super excited to see that wave kind of reach everyday people. Yeah.

And we have to remember a lot of these typical consumers, they're not that technically savvy, right? Outside of San Francisco. I mean, they probably don't know how to set up their OpenClaw, right? So when that becomes much easier and massively widely adopted, that's potentially the breaking point.

Yeah. I mean, I think we saw something similar with generative media, honestly, where it's like before it required these really specialized skills to be able to make a good image or video or design or something like that. And AI has really democratized that. And I think we're going to see something similar for code. Like a lot of what a lot of these agents are doing is they're writing some sort of code or some sort of ongoing workflow for people and running it for you, which is something that the vast majority of people, like me included, cannot do well on their own.

It's such an exciting time to be building. If you were to go and build a company tomorrow, what category would you build in? And I'm sure you've thought about that, or you've got some ideas you would maybe noodle on at the weekends.

I never had the bug of, I have to go start a company. I think in general, the good advice is probably start a company in a space that you know a lot about and you're really passionate about. So for me, it would probably be either something in generative media or consumer agents. I just, I think there's so much potential to build the, to bring the power of these models to consumers in consumer-friendly interfaces. Like we have not at all reached sort of saturation in terms of the number of products that people will use. We're just so early, like we don't even have kind of friendly, accessible product experiences, let alone like good brands or like great marketing for a lot of these AI products. And so, I just think that's like a massively underserved space right now. I think there's a ton of really exciting things happening in B2B and on the infra side and sort of in research. I'm not a researcher, or like I've never been like a kind of core B2B software person. So, those are not my areas of expertise. So, I would probably lean more on the consumer side of things. Yeah.

There are also just so many incredible founders out there who people can draw inspiration from. Any founders you really admire right now who are just building incredible things?

There's so many really fantastic founders right now. I think the ElevenLabs founders, I will give a shout out to just because we did the Series A and a bunch of rounds since. So, I've had the opportunity to work with them a bunch over time. I think with all these AI products right now, you sort of have to find like an initial wedge where you can get attention and an initial user base. And for ElevenLabs, that was with a lot of random audio use cases. Like, we talked to them a few weeks after they launched, and it was like people making like voiceovers for games or like people dubbing audiobooks in like obscure languages or sort of Interesting long-tail use cases where you're like, I wonder what this is going to be as a big company. Like, I wonder who's going to be using this. And then, I think kind of co-developing like the research, the product, and the go-to-market, which is super hard to do, so that they're all moving forward so that as the models themselves get better, they have these new product interfaces that can sort of capture and harness the value of the models. And then they can go and sell those to customers, which at first was like largely consumers and prosumers, but now is a lot of traditional enterprises and expanding across categories, like starting largely in the creative ecosystem and media, which is still a really big space for them, but then also exploring like where are the other places voice is used and as voice agents became more popular. So to me, they're, they're sort of like, how do you execute very well across all of these different categories at once? Staying focused on the core, kind of keeping the models getting better and better, but also not letting product or go-to-market fall behind so that you can really take advantage of the opportunity in front of you.

I have so much respect for them, Matti and Pietro, 2 Polish founders from this tiny little country in Europe. They just crossed, crossed $500 million, I think, in annual recurring revenue. And they're only a few years old, so full respect to them.

Yeah, no, it's remarkable working with them. They're like, I think we knew that they were very special, but it's also super fun to watch founders mature from like a team of maybe like 10 people when we invested to suddenly like hundreds of people at this company that they've created from nothing. Like, it's, it's when you kind of follow something like that, you, it's, it's just truly remarkable.

You know what's really annoying? I had a very similar idea to ElevenLabs about 5 years ago. There's a tweet somewhere out there where I said, at some point, how cool would it be for every creator in the world to translate their content into any language using their voice? And at the time, I was at another VC fund, and ElevenLabs came along. So hats off to them. I'm really excited to see where they go next.

Me too.

Hey, Justine, this has been amazing. Thank you so much. We've covered so much. I feel like now is the greatest time to build, to be building, but also anyone who's just interested in new media, give it a go and get building some AI microdramas.

Totally. Yeah, I cannot wait to see all of the incredible content that's going to be coming out over the next few months.

See what happens. Thank you so much for coming on the show.

Awesome. Thanks for having me.