transcribe

Inside the Tech Powering 500K+ Audio Drama Creators | Pocket FM on Lightwork

Lightspeed Venture Partners · 41m · transcribed Sep 2026
More from Lightspeed Venture Partners Business
𝕏 Share ▶ YouTube 📥 PDF 🤖 .md

Section Insights

# 0:00

Introduction to Pocket FM and AI in Entertainment

What is Pocket FM and how does it relate to AI in entertainment?

Pocket FM is likened to 'Netflix for audio only dramas,' offering serialized audio fiction designed for mobile consumption. The platform allows anyone to create high-quality entertainment from home, reflecting a shift in content creation facilitated by AI.

  • Pocket FM enables creators to produce serialized audio dramas easily.
  • The platform has over 550,000 creators and has achieved significant revenue.
  • AI is transforming the entertainment landscape by lowering the cost of content creation.
# 8:12

The Role of Human Creativity in AI Storytelling

Will AI make storytelling scarce?

While AI can assist in structuring stories and character development, the unique creative ideas that make stories stand out are still generated by humans. The bottleneck in storytelling remains the originality of ideas.

  • AI can enhance storytelling but cannot replace human creativity.
  • Unique ideas are essential for successful storytelling.
  • The democratization of content creation is leading to diverse storytelling.
# 16:25

Retention Challenges in Micro Drama Platforms

Why do some micro drama platforms struggle with user retention?

Many micro drama platforms face high churn rates because they offer similar content, leading to a lack of differentiation. In contrast, Pocket FM maintains high retention by prioritizing unique and high-quality content.

  • Content differentiation is crucial for user retention.
  • Pocket FM's high-quality standards lead to better engagement.
  • The micro drama market suffers from derivative content.
# 24:38

Challenges in AI-Driven Content Creation

What are the challenges in using AI for creative writing?

Developing AI for creative writing involves significant challenges, including managing narrative memory and character coherence across episodes. Custom models are necessary to address these complexities effectively.

  • AI struggles with maintaining narrative continuity in long-form content.
  • Custom models are essential for effective AI storytelling.
  • Operationalizing AI in creative writing presents unique challenges.
# 32:51

Impact of Pocket FM on Creators' Lives

How does Pocket FM impact individual creators?

Pocket FM has empowered creators, including those facing personal challenges, to share their stories and achieve financial success. This reflects a broader trend of democratizing storytelling and allowing diverse voices to be heard.

  • Pocket FM enables creators to turn their stories into viable careers.
  • The platform fosters inclusivity by allowing anyone to share their creativity.
  • Success stories from creators highlight the transformative power of storytelling.

Transcript

0:00 The era of AI entertainment, it's going to be insane. >> Eight years ago, Roan Nyak bet that people would binge fiction with their ears. In serialized audio dramas, an episode at a time, first in India, then in the United States. >> I believe the endgame is going to be billion creators producing billion stories for billion people. >> He co-ounded Pocket FM in 2018 Acro Credit Swiss and PGM and now runs its parent company, Pocket Entertainment. Anyone, literally anyone in the world can create studio quality entertainment at their home.

0:32 >> The company has recently crossed $500 million in ARR with more than 550,000 creators publishing. Rohan Nyak, welcome to Lightwork. To kick things off, I think what would be really helpful framing is it feels like we're in this moment where there's just so much discussion around AI and productivity, AI and coding, but largely not as much as I would have expected on AI and entertainment. and we sort of have an expert on that in this room.

0:56 So have a ton of questions for you even just on what happens to the space when the cost of creation goes to zero, how you all been able to scale that level of personalization. But before we get into all of that, I think for viewers who are just less familiar with Pocket FM, what is it? How do you use it? Tell us a little more. >> Pocket FM is is is think of it like Netflix for audio only dramas. So in audio dramas are these bingeable serialized audio fiction stories. you know you have these 6 to 8 minute short episodes. These episodes have really strong hooks ends on a cliffhanger.

1:33 These are designed for mobile entertainment. This is not designed for television where you you have a setup and it becomes boring. It's designed for you know quick payoff. There's just so many cliffhers because you have to design these shows for low attention spans today because I mean if if you if you watch television content I mean at least I nowadays keep forwarding because it just keeps >> sometimes I watch a whole shows in like 20 you know 20 Tik Tok bites and stuff.

1:58 >> Exactly. So when you try to build an audio entertainment category for mobile we we really had to re-engineer the episodes in the sense that really fastpaced engaging ends on cliffhanger. So that's what audio dramas are and audio dramas if you if you think about it is the only 247 entertainment medium available to any any any user today. So you can listen to it anytime anywhere. It's so so much so that our average engagement time per user is 155 minutes.

2:29 >> 155 Oh wow. >> Per per user. Users just users listen to it every day per consumer. >> Yeah. That's crazy. That's like you're constantly listening. >> Yeah. I mean, if you really think about it, right, you have you have commute, you have household chores, and you're at work and then before sleeping. So, the interesting thing is 38% of our engagement time is at work. >> Oh, really? >> Yeah. >> Like people have their headphones in and they're listening to micro dramas while getting like work done.

2:55 >> Yeah. They're listening to audio dramas and even before sleeping is a very interesting one. A lot of lot of users tell me that you before they go to bed, they don't want to look at screen and it sometimes just, you know, makes them not go to bed. I mean if you if you're trying to just relax and ease into watching a screen just you know doesn't feel natural to me. >> Yeah. Actually, one of my hot takes in the space has been I I have been sort of like vehemently opposed to this idea that we're going to all watch and, you know, exist in the world through these Apple Vision Pros when in reality, I think as a consumer, to your point, for me, I find myself translating articles, you know, into audio format or listening to way more audio just because I'm trying to move away from being stuck to a screen. And then this feels like a a transition. I guess that maybe funnels into my next question. just broadly in the entertainment space there's just been a lot of evolution right we've gone from linear to TV to streaming to creator economy where do you see the swinging in that evolution >> if you think about the three big eras we s started with linear TV and in linear TV there were there were a few networks and they decide what is put on and you just have to watch it and then came along streaming or on demand where now there was there was this explosion of content in on demand streaming And it democratized consumption. Now you have so many options. You can watch what you like. And that democratized consumption in a really big way. And then you had social media entertainment which democratized distribution in the sense that you know if there's any creator who has a great text, image, video, they can get millions of they can reach reach out to millions of users and and and and you know make a living out of it. I believe the the era of AI entertainment, it's going to be insane.

4:41 It's going to be bigger than all these three veils combined. And that's and the reason behind that I believe is because for the first time anyone literally anyone in the world can create studio quality entertainment at their home. They don't need a budget. They don't need studios. They don't need a big team. And it's democratizing highquality entertainment. I believe anyone can now make an audio drama, a movie, a TV show. What I mean, I would love to have my own movie, right? I mean, and but if I think about making my own movie, I'm like, it just takes millions of dollars. You know, you you need a big setup and that's what's going to change. So, think about what's going to happen. You know, if you really think about these really highquality stories or legendary IBS, there have been so few of them. Human creativity is limitless.

5:31 But why do we see such you know very very such few shows or few IPs which are amazing like Harry Potter, Lord of the Rings, Game of Thrones but there should be like 100,000 of these but I feel as dem as as highquality content has not been had not been democratized. We just didn't see these creative geniuses all around the world get access to this. Now NAR is going to I I I feel they're in the era of unlimited highquality content and content which will then eventually become hyperpersonalized. So it's going to be really interesting what the next few years are going to be in AI entertainment.

6:08 >> So let me see if I understand this. So you're saying that because the cost of creation is kind of gone to near zero, right? that you can now make so many more things that maybe for the average creator this creates all this opportunity to be more creative and create more IP like a Harry Potter or a Lord of the Rings but you know maybe it starts on pocket. >> Exactly. If you if you think about you know it maybe 10 years back it used to be so hard to to to to record a high quality video. you had to have these really expensive expensive equipments to record a high quality video. But with smartphones, it the cost to record a video just you know crash and that gave rise of to these big platforms of like Instagram, YouTube, Tik Tok. So that's exactly what's going to happen here. If the cost to create highquality content as it's crashing down, the second order impact of that is anyone can do anything. So that's what that's what exactly happening and and and yeah I mean I believe that and that's my maybe my naive personal dream. I'm a big fan of anime and you know and a huge fan of anime and I believe animes are the best stories in the world and and what's your favorite series? You have a recommendation >> One Piece.

7:22 >> Okay. allies >> and u and and what what I'm trying to say is that I I I'm I'm very sure that one of these legendary IBS will originate on pocket just because we have more than 500,000 creators today there going to be millions of creators tomorrow and this with this level of I mean if you think about harnessing the creativity of of the of human civilization of course we'll get you know we'll get more legendary IPS >> yeah yeah it's so interesting that you lay this out because obviously the you know I even think about I don't know if it's extreme to go all the way back here but like the ways that writing and the printing press unlock just so much more intelligence for us as a society and I wonder if there's ways that this unlocks collective you know creativity in many ways and you know as you outlined I've seen it even in the platform shift from you know social media platforms but specifically to Tik Tok in the way that and short form media and that it sort of created to your point a whole population of people who have now access to building audiences and I think this is sort of like another build on top of that. I guess I'm curious do you view or have thoughts on any second order effects on of this? What becomes you know the important thing for humans today when you know AI can do a lot of that output and the the execution part does it does storytelling be scar become really scarce? How do you view all of that? I believe I think ideas are still scarce in the sense that I mean I I I believe that humans will will be the soul of the show >> and AI will execute it for you in the sense that AI will you know structure story arcs AI will help you design characters progression plot progression and even write the episode for you but but when it comes to that unique insight like creative spark work which really makes the show an outlier or really makes the show a blockbuster or a hit is still an idea. It would be an as easy an idea as someone who gets a scar on his forehead and and and you know and and his and his mother's love saves him from you know from a from an evil sort of force. it could that unique idea is what matters and I still believe that's there that's still the bottleneck for so many IPs but again the so billions of humans you would find these creative sparks somewhere in the world and we're seeing that we're seeing such crazy storytelling entrepreneurs producing content on the platform you could be anyone you could be you know we have seen really interesting examples of former you know delivery partners correctional officers teachers who are creative of geniuses >> but life pushed them in a different direction. But then the when they realize this this is something which which they can do. They can create create their own stories, create their own worlds. They're just shining now.

10:16 They they've left whatever they used to do and doing this full time. >> Wow. >> So interesting. You're finding all these people that have those stories inside them, but they just maybe didn't know how to get them out. How do you work with the creators on your platform to maybe support them and show them the kind of love that platforms have done historically? Everyone from the way Netflix works with their creators to way the way Instagram and Facebook work with theirs. We have really elaborate creator programs and even creator educational sort of materials because storytelling is an art and and you know there just so many different components to make a great story. How do you plan story arcs?

10:56 How do you design characters? Of course, a lot of that has become easier now with AI where you can actually I mean I I I usually tell people that are the fiction writing tools that we have built is the best editor for you. Imagine if every every writer on the planet had a great editor who can brainstorm with you, critique your work, give you really good valuable feedbacks. So, of course, we have a lot of these create great creator programs and we're also launching now creator events where, you know, we'll soon host something in in Los Angeles in LA and I'm really excited for that.

11:31 We'll just call all of our creators into one place then design it in a way that people get to learn from each other. but very interesting thing that I I feel that that that you know we saw on pocket is that most of our successful creators are power listeners on the platform. So they would listen to a show for months for years study these shows and then at some point they can make their own. Yeah. At some point you're like yeah you can do this you know.

11:59 >> Oh yeah. I mean now you're like getting me out. I'm like, "Oh, should I play around with this and and be a you not just like a consumer but a creator on this platform?" But tangly, what does that look like? How do I start becoming a creator on Pocket F? >> Yeah, I mean, if you want to write, you it's on the app. You just click I mean there's a there's a right section on the app. You go in, you have an idea. We are a fiction tool called Sherpa. So you put in the idea whatever that vague idea could be and Sharpa will brainstorm with you trying to trying to just probe you with questions. Hey how you thinking about this? Are are you thinking about the world and once it has some level of plan it will design a story arc a spine of the story. It'll break down that story arc into episodes and and yeah you can just click a button and the show will go live on the platform. That's interesting because I think a lot of people tend to have questions about the the balance between consumption and creation on platforms. Historically, it sounds like, you know, you have 1% of people who were actually creating, 99 or 9% remixing, and then the rest 90% consuming. Are you seeing that dynamic?

13:06 It sounds like you're actually sort of ushering a very different sort of new paradigm where people are both creating edited scenery. I think in the last 12 months we have seen a crazy inflection in the number of hours being produced on the platform. Just for context, we used to have around 25,000 hours being uploaded every month. so every year and now it's 2.6 million hours of AI content. The creation has exploded and yeah, you're right. Exactly. I mean a lot of people want to try out. Everyone has stories if you think about it. Sure.

13:39 I mean, if you have to make a YouTube channel or Tik Tok channel, you still have to think a lot. You like, "Hey, what do I do? What's my niche?" >> Stories of all of us have them. I mean, I would like to believe I have a few good ones and I've written 100 shows on the platform, but none of them I've done well. That's a different story, but >> I love that you're still writing them.

13:56 Yeah, >> I love that. I love it. I mean, I just feel that, you know, it's it's great to be in in a situation where you can just create your own world. but that dynamic is changing. I believe what was going to happen is they're going to be billion billion creators creating stories and that's going to lead to hyperpersonalization of stories. I'll explain. If you think about most streaming platforms today, you have a few shows for a lot of people like I like space fiction. I don't see any movie being launched in space fiction.

14:26 It's been a year maybe. The last one was Project Kalem Mary which I really loved. >> Love Mary. Yeah, >> that's best. One of the best ones, right? I mean I I read the book as well >> but if you think about the number of tropes in entertainment and a number of themes there just so many but entertainment unlike social media is doesn't have that huge volume of content like if I like a if I like a tennis video on Instagram I get flooded with so many different tennis videos because there's a lot of supply of tennis videos but if you if I like space fiction there's no supply of space fiction content then what what I believe will happen and why I'm saying there'll be a billion creators because there's so many different niches and it historically producing content for these niches didn't make sense because there's ROI you have to you know you have to spend a lot of money but let's say someone likes vampire cross space fiction it's a niche but maybe maybe a million people would like it or maybe 100,000 people would like it so that's what we'll see more and more that these really crazy concepts will start coming coming around >> there are a lot of those fun crossovers on the on the app. What are some of your favorite of the dramas that have been made that maybe surprised you and are are ones that you wouldn't have expected but wound up being really big hits?

15:44 >> The biggest one is there's a show called My Vampire System. >> Yes. It's a vampire show and vampire cross science fiction in a way and it's it's it I mean if you if you if someone would have like you know in Hollywood looked at it they like dude I mean like right they wouldn't have green lit that >> they would never have green lit that but users should vote for what they like users should tell tell us or anyone what what a what a hit is and that show has now made over $90 million that's community validation that's incredible >> on that though I think historically at least in the market. I've heard a lot of push back against biker dramas because of you know retention not necessarily being strongest often struggling with monetization or or users sort of vanishing and it sounds like you're not seeing that. Why do you think that is and what sort of retention numbers are you seeing? I think one thing I've realized in micro dramas and I'm talking about the video micro drama at this moment that the reason there's high churn and the retention rates are lower for a lot of the other plat lot of lot of microdrama platforms is because most platforms probably all of them have the same shows they have billionaire stories racks of riches they have some re is sort of all the same trope >> all the same so what really happens very interesting data point so let's say if platform A launches is one micro drama.

17:05 Platform B would change 40% of it and launch within a week. Platform C wouldn't change another 40 another whatever. So they're all derivatives of the same show. So there's no there's no value. I mean it's the same content that if if 20 platforms have the same shows, why would you have retention? This didn't happen in the streaming era, right? You have multiple streaming platforms who come up with really unique differentiated IP and all of them like all of the all of these platforms are different from each other. But in micro drama all of the all of them are optimizing for social media ads performance marketing. They're marketing I call them marketing arbitrage platforms because they because they market on meta. Meta makes a lot of money and then you know users come in and they pay for a show and then churn.

17:53 But in in Pocket FM or audio drama, what we realized was our bar for content is very high. So So when you think when I when I talk about a hit on Pocket FM, an average hit or a blogbuster on the platform has 40% retential till 500 episodes. >> 500. Wow. >> Wow. So on an it's a lot. I mean it is so engaging. It's so great that 40% retain and 500 episodes is a lot. >> Yeah.

18:21 >> It's a lot of content. So these shows are really high quality. >> Why you can't even get people to stay for 40% for like 1 minute. >> Yeah, >> that's that's the trick. You know the first minute first minute is so important and and after that if the if the users fall in love with the characters, it becomes a part of their daily life. It's all character love. Character affinity is the single most underrated thing in entertainment. We just make someone fall in love with the character, what they stand for, their principles, they'll be around. Wow. So, so yeah, I mean that well that's one of the reasons the quality of content that we have on the platform is really great.

18:54 >> I think one of the critiques that I've often heard with anything AI and entertainment is just broadly idea that everyone's just going to create slop or everything's going to be a recycling of everything. Are you seeing that or or is it are you seeing something different play out with popular? >> There's going to be content which is recycled. I mean creator if it's a UGC or user generated content platform you're going to see a lot of similar shows as well but we have it it's more about the design of your algorithms. So algorithm is designed in a way to promote differentiated content or unique content. Eventually the shows that do really well on pocket FM are the ones which are very unique because users are smart. They're not going to pay for a show which which they've listened to 100 times. Yeah. Yeah. So they want they're looking for these unique interesting shows that can intrigue them. So as a result even with AI like the shows that do really well on on our platform are the outliers outliers like my empire system which you would not have thought about ever. So so yeah there is a lot of AI slob. I mean not I would not call it AI slob but there's a lot of content which is which creators produce and of course not all of them are are great. I think what's really interesting is while you are enabling the creation tools as well, you're also actually very focused on the the delivery experience at Mechanism for consumers, which I actually don't think a lot of platforms focus on. To your point, as a consumer of a lot of online content, my frustration has been you like one video about, you know, one show and then all of a sudden you're only fed and and so it sounds like there's also actually like equally a importance on the user experience on the consumption end that makes this a much more powerful ecosystem.

20:34 >> It has to be user focused. I mean everything it's a flywheel and creators and consumers I mean if if consumers see a lot of like lowquality content like going to churn which just breaks the flywheel of more consumers bring in more creators more creators more hits more consumers and so on. We we have worked a lot on the algorithms that drive discovery and personalization of audio dramas on the platform. The the most the the hardest part now has been with so many shows being produced on the platform. How do you personalize and recommend the right show? How do you how do you find the blockbuster out of these 2.6 million hours of content? So this this is where we we have been doing a lot of work. So the first step if if you create if any of us creates a show there's an there's an LLM as a judge.

21:22 Now, >> so because of our fiction writing stack, we are really good at and I mean we we've gotten good good at identifying patterns which make gives us you know enough confidence that the show is not going to work. If there's narrative discoherence, there's no narrative coherence. There is just just lot of different characters being introduced in the first episode without any reason. There's so many different patterns that we have now uncovered from our data that we know that LLM as a judge can can at the outset tell us that hey these shows will not do well. So you you don't you don't you don't distribute these shows to a lot of listeners. So there's a lot of work that is that is going behind how do you find the blockbuster without with limited data with so much supply how do you find blockbusters?

22:08 That's a very hard problem to solve. Is it similar that to way you know I don't know Tik Tok works where you know it's set to a small set of people and if those people seem to enjoy it does it continue or or is it working in a different way? >> It is a continuous propagation. So we show it to 50 listeners 100 listeners if we get good signals from there we we sort of you know show that show to thousand listeners 10,000 listeners 100,000 listeners and so on. But it's a continuous propagation algorithm. But the one thing about Tik Tok which makes it easy is that you can just swipe up.

22:42 >> Mhm. >> You don't like something, you swipe up. But in Pocket FM, you you just can't swipe up. It's not a 15-second audio. So, it's you have to be right more often than wrong. So, we just have to be a little bit more aggressive in making sure our user gets the right show. You've done such a good job on the app of like making that serialized activity that I think is pretty unique to you guys. How do you view maybe competitors in the space? I mean like recently of course Instagram started adding the serialized thing to playlists and there are quite a few micro drama players but who do you really view as like primary competition especially as you guys have exploded in growth?

23:22 >> There are a lot of competitors. I mean pretty much everyone in entertainment is going to be a competitor fighting for attention. >> You're fighting for attention you know if you if you think about it our engagement time is 155 minutes per day. So you're really competing with everyone across the spectrum. So anyone who you know who's an entertainment platform has great stories on their platform, you're competing with them. So Rohan, you've had this incredible growth. How have you managed to keep up from a technical perspective? I'm sure that the user growth has been a strain on your systems and all of the creators must mean even more complex things for your LLMs. How do you keep up with all of it?

24:00 >> It's been crazy. I mean the last 6 months itself with such the I mean such a big jump in the number of hours been produced you know the in the compute costs the just the inference costs have just gone up multiffold and and it's so it's it's sort of like I feel and at at this pace of AI research that I mean I mean AI is at a completely different ballgame right now it just the pace of development is just so insane and what's happening is that we we have to make sure we're the cutting edge of it and not just AI but even AI entertainment. So we've we've of course we've built our own we've postrained our own custom models for for fiction writing which is you know pros writing or planning step of a fiction story. So we have built our own custom models but it's super hard. I mean our infrastructure the the inference costs has been on a on a on a crazy ramp up and we're also trying to figure out how to optimize these costs.

25:01 and that's been a big endeavor but but it's fun. It's it it I think these are good problems to have. >> Yeah. Yeah. Scaling problems but >> balance every startup wants. Yeah. For sure. I think also one of my favorite parts about technology and especially like consumer technology is there's a lot of stuff behind the scenes of a delightful experience like Pocket FM that people often don't visibility into. And the ones that I think of even just as I think about the the idea that you can now create entire worlds. There's issues around how do you technically embed narrative memory? How do you have this sort of like coherence in characters and across so many episodes?

25:37 There's all these things that like I think as someone who's been in the seat of a cruer I'm like wait that's really hard to operationalize have you what has been difficult and have you thought about you know productionizing that >> shar which is a AI fiction tool we we have it think of it as a multi- aent orchestration platform which is built on top of custom post-trained models before that maybe I'll take a step back and talk about why general purpose LLMs are I mean at least when we tried them they're not really good at creative writing >> cuz the three fundamental problems when it comes to general bubbles one is memory narrative memory just just >> if you're writing a show which has 500 episodes thousand episodes they're like 150 characters the relationship between the character >> for sure they absolutely forget the first thing that happens >> yeah progression there's progression the character's personality changes I mean if you think about it a character is growing through different going through different phases of life the even the personality traits are changing. So you need to keep track of that as well. It's not just, you know, it's just not the the plot progression, it's all the the character progression and and so LLMs today when it comes to writing long form stories, they just forget a lot of details and and in fiction these nuances matter. You you're falling in love with the character and it how the character behaves changes. You're like, hey, what happened like what what happened on 5001 episode? So the narrative consistency long-term has been a big problem for LMs. And then this the second big problem has been pros or the you know how these shows are written. My point of view is that LM are really good at coding logical reasoning but this is emotional reasoning. If you're writing if you're having a fight sequence like if you're two people are verbally fighting it has to be emotional.

27:30 >> Yeah. >> It can't be just facts. I can just tell you an example from a micro drama I watch. It was like I'm going to fight you. Yeah, stop fighting. I'm gonna fight you. And it's just like it doesn't feel you don't have that emotional intelligence. Yeah, >> it has to be Yeah, it has to be over the top. It has to be those emotions in those in those situations. So that was this pros problem which LMs are not good at at least at the moment. And the third one which is a big one which is probably the most the hardest of them all is planning a long form story. No, you try to write a crime thriller story with any of these LMs, it'll tell you who did it on the fifth episode.

28:07 >> They struggle with payoff. >> Yeah. I mean, I'm like, what's the point? I mean, if and like if you're writing a romance drama show, you don't want them to meet at at the moment. You want that push and pull. If you make them meet, they their users will churn there. There's no need for them to now continue listening to the story. So these three big problems we had to solve for for the pros part. We have so much data on the platform. We have 5 billion hours of engagement data on the platform. These is minute-by-minute retention data. So we know what a good pros is, what a bad pros is. So we're able to post train our own models for pros. Then even on the planning side, we know what a good plan is, what a bad plan is because you have so so many of them on the platform. So again we were able to postrain a model for just planner and then the third component which is the narrative memory we have you know think of it like this we have a memory structure which stores everything about the story literally everything so there's 0% hallucination so these are the fundamental building blocks of sharpa our friction writing tool and yeah I mean there's still a lot of lot of work to do because I mean you asked about what's still hard I would say it's planning so when you think about planning a 500 episode story or a thousand episode story. You have to plan not just plot progression but character progression and and not just individual character progression. If there are like 100 characters, their stories together, they I mean how they how do they how how does their relationship progress? So there's just so many different components or aspects of a plan. All of these have to work work together to to deliver a great great long term long form story. So planning is still really really hard.

29:53 >> It's interesting because you know I think there are often folks who will say I can just do this with you know a bunch of models and chatb. But based off of what you said there sounds like there's a lot in there that makes it all work. You have visibility into billions of hours of data. I'm actually curious just from a entertainment standpoint. What do you think goes into a good story? >> It's very unpredictable. So of course I can wear my analytics hat and tell you some patterns that we have seen. the the most important most important thing about creating a story which is retaining which which retains a lot of users which engages a lot of users is character affinity. So if you can build characters where users users can fall in love with them they they root for them they they you know they like the principles of the character that matters so much more than just the plot big files. So a lot lot of writers what they do is they focus a little bit too much on the plot on the plot progression but not too much about how do you design these characters. These characters need to be you know have need to be so strong that it retains the users for 500 episodes. So how you design characters and if you can build that very strong character affinity matters a lot. So that's one really important really important aspect of creating a great story. The second one I'm I'm sure you will relate to it is the first few minutes of the show, the first hour of the show. A lot of writers would would plan this great story but then the first 30 seconds minute, first one minute, first 30 minutes >> is not strong >> is not strong and there'll be a massive drop off in the first 30 minutes. So >> whatever you do, first 30 minutes, first 1 hour, the beginning is the most important for any blockbuster show.

31:42 >> The hook, >> the hook, but beyond that, it's it's very interesting. You know, beyond these tactical sort of elements, storytelling is very personal. Different people have will have different takes on how to design a character and how to design the world. It's very unpredictable and which is what makes it fun, I guess. >> Yeah. Do you have any good human stories about some of the creators but on the platform and like you know how maybe it's changed their lives and you know certainly from a you know visibility perspective.

32:14 >> Oh yeah a lot of them. So my favorite one is this former delivery partner and you know life life pushes you in a different direction. you're doing to you're doing whatever you can to make ends meet >> and something happens at book and he's out of book and and his friend tells him hey you you you have been a great storyteller and you tell us so many great stories there's this app called Pocket FM why don't you try it out why why don't you try write a story on Pocket and he goes in writes a show like a mythology cross science fiction and there's just so many stories like that one of the other interesting ones that I I at least you know, you you build businesses, there's revenue, there's all of those metrics, but those are just numbers, but these are the stories that really have a big impact.

33:02 So, for instance, there's this really like I mean a creator I I mean I spoken with my many times. So, he's he's legally blind and he's been again trying to make ends meet and he had this talent that he was a great story. He used to listen to Pocket and he was he's a great storyteller. So he tells his partner and wife that hey I have this great crazy story. Can you help me with it? And that story now is doing so well. That show is doing so well that he's able to support his entire family with that. and and and I'm really proud that it's not just about sharing these stories with with with the listeners with users, but if there's an opportunity that you can create a great show that inspired millions of users and make a living out of it, nothing is better than that, I guess. Yeah, it's it's it reminds me of like the early days when you know you saw the unlock of YouTube that like you could see anybody's life even you they didn't have to be Hollywood but now you've taken that a step further and it's like anybody can extract that creative genius out of them and I do agree I mean you know when you see so many people there are so many creators out there but there is a difference between a really great storyteller and when you equip them with the right tools you know you can unlock amazing stories like these maybe switching gears a little bit to sort of like forward facing obviously see a lot of exciting things. I'm just so impressed by from a product standpoint, from a team standpoint, what you all have built. what's kind of on the road map for for Pocket and what's exciting you?

34:33 >> There's so much going on. we're launching new markets in Pocket FM. AI localization has enabled us to now launch literally 100 markets. We can do I mean it's all like you think of this button. You click this click that button and you're live in like so many different languages. It's just so interesting. So we're launching new markets in pocket FM. We are launching Sherpa which in a very very big way which is a fiction writing tool. the the the the latest upgrade is just really really powerful. Our tokens have gone gone through the roof now and we need to figure that out. And so yeah so we're doing a lot of work on expanding genres expanding to new markets but the most interesting one is also we've now launched a new our new AI microdrama app a video drama app called pocket saga where what we did was we took all of these userenerated stories that we have on the platform in audio.

35:32 So we built a harness which takes these audio scripts, turns that into a screenplay and then turns that into a micro drama, a long form micro drama which is coherent and and oneshots it. And the writers are just so happy because they came into the pocket ecosystem. they produced an audio drama that that audio drama went live in many languages and then now in micro drama. So AI micro the micro drama app is really exciting. It's it's growing.

36:00 It's growing almost like in 3 months it it it's it's it been it has grown to $15 million of ARR just in three months and I feel it will be one of the faster grow fastest growing products we have. >> And are you seeing on a consumer appetite for this kind of you know diversification beyond audio? >> Oh yeah big big time. The demographics of both pocket FM and pocket are different. >> so we're able to expand the TAM of that same IP now into multiple into multiple different audiences and demographics.

36:28 >> Yeah, it makes sense. It's kind of a different use case, right? Cuz if you're listening while you're at work or all these other things, you're not necessarily able to watch your phone. So, it's it definitely feels like it's a it's a little bit of a different audience. >> Yeah, it is. It is. I mean, you know, audio in in audio users are listening 3 hours a day. That's not possible in video. You can't just, you know, scroll 3 hours a day. It's really, really hard.

36:52 >> Do you have to change the format, too? >> Yeah, it's a separate app. So, it's a vertical 2-minute video. So, think of it like if an audio drama episode was 6 to8 minutes, five 500 episodes long. This is vertical 2 minute 2-minute episode, but with the same 500 episodes, though, but 2-minute vertical episodes, >> but you're able to bring the serialized format that exists on Pocket in the way that it doesn't exist with some of the slop videos that we all watch where kind of loses a narrative.

37:19 >> The good good labels, they're slop videos. And of course, try it out. I mean you would see that these are highquality microdramas in terms of the story lines. The story lines are unique. We also have animated micro dramas not just cinematic or realistic ones but if it's a fantasy show we have these really cool animes. Of course I love animes. >> We didn't do just because of that but you know just it's just fantasy I mean fantasy shows really come alive in in animated form as well. So we have both animated and cinematic or realistic micro dramas on the platform and we are seeing really high engagement and retention today.

37:56 >> Yeah, I'm a I'm a big fantasy reader. and also you know I I I think I I can totally see how that that world builds and you know I myself feel very drawn to it. From everyone I've spoken with, I think a very clear, you know, personality trait I've I've learned about you is, you know, one, you're creative, you're curious, but the other thing I heard is that you cannot play without winning and, whether that's like Top Golf or pickle ball or cricket.

38:24 tell us a little bit more about how that translates into building Pocket FM as a founder. >> Yeah, I've always been like this. I I'm a I I like winning any sport, even if I'm that really bad at it. I I I would just, you know, practice for hours and hours to certain I can win somehow and or somehow hack my way to winning in a way maybe like especially in Top Golf. I I I just pay to win. I I don't care about these long shots. I just care about those tweaks and you know just winning and and just winning.

38:54 >> So I I feel the way to think about the way I think about building a business is it's it's a game. >> It's exactly like a game >> and it should be fun. If you're living this and building this, >> something doesn't work out in a game, you fix it, you know, you know, you you don't get too serious about it. If something doesn't work, you can try again. It's a game and you have to be competitive. You have to win. You have to keep reinventing. You have to keep upgrading yourself. You have to keep you know, sort of improving and and and reskilling in a way. So like like how a game is, you have to keep leveling up. So the way I think about business is you keep leveling up and there are infinite levels. That's the beauty about this other game. In any game like you play Candy Crush that has like >> that actually has a lot of levels. But you know there there's just so many different levels and that's what a business is. There's no end. You just keep getting better. Never settle. Just even if you're at whatever level you are level thousand there's always something you know on top of that. And that's how I how I build Pocket FM or a business. I'm always paranoid. Yeah, >> it's not a good thing sometimes, but and my team tells me that that you we $500 million $500 million the way out. I'm like, yeah, there's more work. It's >> just how can how can we get to a billion? And you know, and and I'm not h I mean, I'm happy, but there's just this level of paranoia that I have that how can we do it faster faster than ever before. So, it just I'm just really competitive, I guess.

40:22 >> I love that there. I mean it sounds like it sounds like there's no end stake to the game but if there were just to help us understand the big vision for Pocket FM what does this look like in 5 to 10 years? >> Yeah for for Pocket Entertainment now that's the parent company of Pocket FM and Pocket Saga I believe the endgame is going to be billion creators producing billion stories for billion people that's I feel would be the would be a nice end game. Well, Rohan, congratulations on this very big milestone this week. I know you're all about celebrating the next one, but we want to celebrate this one. And thank you for joining us on Light Work. Thank you for having me.

Summary

Rohan Nyak, co-founder of Pocket FM, discusses the transformative potential of AI in entertainment, particularly in audio dramas. He envisions a future where anyone can create high-quality content from home, leading to an explosion of diverse stories tailored to individual preferences. Pocket FM has already seen significant growth, with over 550,000 creators and $500 million in annual recurring revenue.

- Pocket FM is likened to "Netflix for audio dramas," offering serialized audio fiction designed for mobile consumption.
- The platform has an average user engagement time of 155 minutes per day, with significant listening occurring during work hours.
- Nyak predicts that the era of AI entertainment will democratize content creation, allowing billions of creators to produce unique stories.
- The platform emphasizes character affinity and strong hooks in storytelling to retain listeners over long episodes.
- Pocket FM employs AI tools to assist creators in story planning and character development, ensuring narrative coherence across episodes.
- The company is expanding into new markets and launching a video drama app called Pocket Saga, which adapts audio scripts into micro dramas.
- Nyak highlights the importance of user experience and algorithm design in promoting unique content to prevent churn.
- He shares inspiring stories of creators whose lives have changed through their success on the platform, emphasizing the platform's role in unlocking creativity.

Questions Answered

What is Pocket FM and how does it relate to AI in entertainment?

Pocket FM is likened to 'Netflix for audio only dramas,' offering serialized audio fiction designed for mobile consumption. The platform allows anyone to create high-quality entertainment from home, reflecting a shift in content creation facilitated by AI.

Will AI make storytelling scarce?

While AI can assist in structuring stories and character development, the unique creative ideas that make stories stand out are still generated by humans. The bottleneck in storytelling remains the originality of ideas.

Why do some micro drama platforms struggle with user retention?

Many micro drama platforms face high churn rates because they offer similar content, leading to a lack of differentiation. In contrast, Pocket FM maintains high retention by prioritizing unique and high-quality content.

What are the challenges in using AI for creative writing?

Developing AI for creative writing involves significant challenges, including managing narrative memory and character coherence across episodes. Custom models are necessary to address these complexities effectively.

How does Pocket FM impact individual creators?

Pocket FM has empowered creators, including those facing personal challenges, to share their stories and achieve financial success. This reflects a broader trend of democratizing storytelling and allowing diverse voices to be heard.

© transcribe · For agents Built with care and craft by Gokul Rajaram