This week’s podcast is about how generative AI is disrupting content creation. This is a major tech shift and a lot of big incumbents (such as Netflix and Disney) are going to be impacted.
You can listen to this podcast here, which has the slides and graphics mentioned. Also available at iTunes and Google Podcasts.
Here is the link to the China Tech Tour.
Generative AI companies to try:
- OpenAI / ChatGPT
- OpenAI / DALL-E
5 Winners in Generative AI
- Specific, niche content creators, usually based on people. Such as sports content, game shows, and reality TV. Also, some creators with do very but will follow a power law.
- Content production tied to community or services
Audience builder platforms
Creator tools providers with standardization Network Effects, such as Adobe.
2 Likely Losers in Generative AI
- Content publishers. These intermediaries (such as book publishers) don’t have much of a role unless they unless they control demand or distribution.
- Stand alone content creators. Content can be created in house and purchased. They will struggle with coming wave of high-quality, free content. That’s Disney, Netflix, most tv studios. Basically, any business that has been relying on scale in content creation. Economies of scale in content production disappearing – especially in animation. Barriers to entry are going away. Companies like Disney and Netflix needs to become audience builder platforms ASAP. They need to expand from mass market to micro markets.
- The Winners and Losers in ChatGPT (Tech Strategy – Daily Article)
- Why I Don’t Like Netflix, Singapore Press and Most Digital Content Businesses (136)
- Why Netflix and Amazon Prime Don’t Have Long-Term Power. (2 of 2) (US-Asia Tech Strategy – Daily Article)
From the Concept Library, concepts for this article are:
- Generative AI: Cheap and Fast Content Creation
From the Company Library, companies for this article are:
- Amazon Video
- OpenAI / ChatGPT
- OpenAI / DALL-E
Photo by freestocks on Unsplash
Welcome, welcome everybody. My name is Jeff Towson and this is the Tech Strategy Podcast where we analyze the best digital businesses of the US, China and Asia. And the topic for today, why chat GPT and generative AI are a mortal threat to Disney, Netflix and most Hollywood studios. Now I hope this one is gonna be valuable to you because I think I’m kind of ahead of the curve on this one. I’ve been spending a lot of time looking at these generative AI. programs. Chat GPT is the most famous one, but there’s quite a few others now. And what that kind of means for a lot of incumbents, what it means for sort of digital strategy, business models based on content. And it’s a big deal. I mean, it is absolutely sweeping. The more I get into this, the more I think this is going to change a whole lot of businesses very, very quickly. And fortunately, I have been sort of studying this particular question. how new digital tools impact existing business models. I have been sort of studying this particular question for about eight years, so it’s right in my strike zone. Anyway, so I hope this is gonna be helpful today, but that’ll be the topic at looking how this is sort of evolving and how it’s gonna hit these businesses. And I think the impact is big. So that’ll be the topic for today. For those of you who are subscribers, I’ve sent you a quick, not a quick, preliminary thoughts on this about a week ago. It was a business called The Winners and Losers in Chat GPT. This is going to be sort of more on that. But you can kind of tell that my thinking is moving along, because that was sort of off the top of my head stuff. So it’s going to base on that. But I’ll also send you a more in-depth article in the next couple days. I think I’ve got this one now. So that’s on the way in a couple of days. If you’re not a subscriber, feel free to go over to jefftausen.com. You can sign up there for a free 30 day trial. And we are getting more active on the sort of online chat group looking at valuations. So if you’re not, if you’re a subscriber and you’re not in the group online, Line the Messenger app, send me a note if you wanna be part of that. Or also WhatsApp, but that one’s really, you know, most people have been online, but we’re moving people over there and we’re starting to take apart Amazon and these companies in terms of valuation. So send me a note if you wanna get involved in that. easy to reach. Let’s see other housekeeping stuff. Again we’re doing a China tech tour that’s April most likely. Five days Beijing, Shanghai, Hangzhou visiting some tech companies doing a lot of sort of in-depth digital stuff you know and it’s gonna be a lot of fun. If you’re interested in that for you individually or for your company go over to TowsonGroup.com you’ll see the link there I’ll put the link in the show notes. Send me a note. can work on that. That’s gonna be awesome. China’s open. These are fantastic companies. Anyways, that’ll be, if you’re interested in that, send me a note. Okay. And let’s see, standard disclaimer, nothing in this podcast or my writing website is investment advice. The numbers and information from me and any guests may be incorrect. The views and opinions expressed may no longer be relevant or accurate. Overall, investing is risky. This is not investment, legal or tax advice. Do your own research. And with that, let’s get into the content. Now, as always, we sort of start with the key concepts, the key takeaways. There’s really only one, which I think we can just put under the topic of generative AI. And, you know, in terms of concepts, if you go to the concept library on the webpage, you will see some topics related to AI, cheap and fast prediction has been sort of my standard one, computer vision, that’s under the catalog as well. If you go over to the concept library, you can find a couple companies in this area, OpenAI, which created ChatGPT, and Dali is on there now. But I think we need a new category, and I think this new category concept is generative AI. This is just a big, big deal. And ChatGPT is the mothership. No question, this thing is incredibly powerful as a large language model. But we’re seeing other versions of Generative AI emerging very quickly in the last, really in the last 12 months. I mean, it’s happening really fast. So we’ll go sort of go through several of those and I’ll give you companies that you should look at, which you probably haven’t heard of, beyond ChatGPD. So let’s put this all under the concept for today of Generative AI. And that will be in the concept library on the web page. Okay. So. The first thing I think to think about here is generative AI. What we’re seeing is, as I’ve said many, many times, AI is cheap and fast prediction. That’s been always what it is. You give the algorithm two pieces of data, and then it fills in the third. So the business use cases we’ve seen for AI are like, predict where my inventory levels should be for this product. over the next three to four days based on what I think the demand will be. And it will make a prediction and then we see if it’s right and it iterates. It’s prediction. And AI based, which is what a lot of people do. This is what a lot of human professionals do is we’re in the prediction business. We’re doing market prediction. We’re doing market analysis. We’re trying to design what’s gonna happen. And well, AI does that really cheap and fast. Initial business use cases where things in like logistics, a lot of demand projection. a lot of marketing. We predict that if you change the marketing ad from A to B, you’ll get more. Okay, that’s all prediction. When we start to apply cheap and fast prediction to content creation, generative AI, the simple use case of that would be I give you a sentence, but I leave out the fourth word. I want to go to blank store. AI predicts the missing word is the. I want to go to the store. That would be content prediction. Okay, well AI, I mean, Chad GBT has just taken this to the next level where they are predicting not just a missing word, they’re predicting the next sentence. I give you two sentences. I went to the store to buy the groceries. At the store, I knew I needed milk. I ended up blank. And then it predicts the next sentence like buying a carton of milk and going home. Now there’s no necessarily right, here’s one of the reasons a lot of content prediction is so effective, because there’s not necessarily a right answer for a lot of it. If it’s creative, it could be anything that makes you sort of say, oh that’s a funny next sentence, oh that’s a funny poem. If it’s more about having a right or wrong answer, here’s the next line to the code. And if you don’t get it right, the code doesn’t function. So For those of you who are subscribers, I’m gonna send you a, in the next email coming, I’m gonna send you a breakdown of different types of content creation based on does it have a right or wrong answer or is it more something creative? Hey, this is a funny photo. There’s no right or wrong answer to that. Okay, but terms in a contract, there’s often a right or wrong answer. So you can start to break down content creation types and you can see that certain AI is better at some than others. And I’ll give you the short answer right now. If it’s about creative content where there isn’t necessarily a right or wrong answer, AI is a lot more powerful and it’s moving quicker. Hey, this is a funny photo. There’s a lot of ways a photo can be funny. Actually I take that back. Here’s an entertaining photo. Humor is actually pretty hard to get right. A nice pleasant photograph or painting, there’s a lot of right answers to that that people might like. The right line of cone is a little more specific. So we start with this idea that like we’re looking at generative AI tools, which is prediction in the creation of content. Okay. What does content mean? Okay, well, text, words is a type of content. So that’s chat GPT, right? They predict the next word. Now it turns out that’s kind of a lot of things. And they’ve gone far beyond fill in the blank. They’ve gone far beyond fill in, give me the next sentence of this. I mean, it can write paragraphs, it can write stories. It is full on understanding the entire, like let’s say I give the AI a sentence. I am going to the bank. And I ask it to say, predict the next sentence. Well, the answer to that depends on context. Am I talking about the bank of the river? Or am I talking about a bank downtown to borrow money? So it has to start to look at context around the subject before it can give you, I am going to the bank and I’m going to go fishing, or I’m gonna take a loan. So you start to get more into context, it starts to get a lot more complicated. ChatGPT is very, very good at this. Startlingly good. Okay, so if it’s gonna generate content that is text, What does that mean? Well, there’s a lot of subtypes. It could be write me a story. It could write me a poem. Write me a monologue by Julius Caesar talking about why he doesn’t like to go to the shopping mall. It can write that. Write me a wedding vow. Write me a marketing pitch for my new sneakers that are for 18 to 23-year-old women. who are stylish but like, you know, you can, and it can do that for a rap song. It’s pretty good at writing rap. It can write an interview. Write me an interview about the subject of chat GPT between Steve Jobs and Jeff Bezos. It can do that. Write me an article. And then you start to get, so that’s all, a lot of that is creative. You can be inaccurate if it writes about Chappie TV and it’s wrong. That would be wrong, but there’s no wrong answer to a poem. As you move closer and closer to right and wrong answers, it would be like, write me a contract. Write me a contract for a rental lease in the city of Santa Barbara, California. It can write you a contract that is in line with all the California codes. and requirements and is compliant with all any specific Santa Barbara regulations for rentals. It can do that. It’s really impressive. As you start to move into more technical things like write me Python code that creates a digital clock that runs on my screen. Okay, that’s far more technical. And if it’s wrong, it doesn’t work. Turns out it’s pretty… pretty good at writing code too, but there’s a lot of debugging you have. So there’s a spectrum here of content types related to text. And that’s where chat GPT lives. Okay, that’s one content type. We go to another content type, images and visuals. A cartoon, make me a cartoon character who is a funny rabbit that, you know. with big ears and it will generate 50 different types of cartoon rabbits. This is, or let me give you a better example, generate for me 10 different types of cartoon mice that children would like and it generates 10 to 15. That’s Mickey Mouse. That’s what Walt Disney started his company by doing. Generate me a photo of a downtown street and I could give it that way or I could say I’m going to upload a photo based on the current photo. and it can look at the uploaded photo, it can search the web, and it can generate. Make me a logo. Draw me a picture of really cool hats for Sweden, and it can start to give you a bunch of hats. Well, that starts to look like product design. Create a character for me who’s a spokesperson. Create an avatar of myself, and it can do that in various styles. So… When you start to generate these photos, these visuals, these images, these cartoons, create a comic book super villain based on Thailand, which I can send you what they’ve done. It’s pretty awesome, by the way. This starts to look a lot like design. Now the two companies like you should absolutely pay attention to, and I’ll put these in the show notes, is Dali, which is OpenAI, that’s ChatGPT. founded by Elon Musk, which people don’t know, co-founded. The other one that people don’t know as much about, which is, I think, spectacular, is Mid Journey. You can go on and sign up for Mid Journey right now. It’ll send you to a Discord, and you can start having it generate images. It is absolutely unbelievable how good it is. I think more impressive than Dali. Anyways, I’ll put that in the show notes. You should try that. It’s okay, so there’s text, there’s images and visuals. We can go the next step. What about videos, moving images? Well, you can say, here’s a visual that I just created in Dali of a cartoon mouse, and I can put in some text. I want this mouse to walk and trip and fall down on the street, and it can start to generate videos that do that, so you can start to make cartoons. You can start to make an avatar. I can say create an avatar of myself, I’ll create it mid-journey. Make a picture of me based on my photograph that I look like a super villain, that’s a photo, picture. Now let’s put it into a video AI program. Now I want this video to start to talk and say what I want it to talk. That’s an avatar, you can do that today, it takes 10 minutes. So you can create, you can put in the text for what it’s gonna say and all that. There’s a company called Synthesia, I’ll put the link in. They basically, you give it text and it can create an avatar that says the text. So I can give it a script to read and my little supervillain Jeff will read the text. And it’s getting better and better. That’s video, moving images. Sound. Sound is another type of content. There’s a company called Aiva, A-I-V-A, which creates music tracks. You want to create background music for a video that’s really scary and moody. it can generate that immediately. Voices, I can give it text, it will say the text in a voice. And I can say, make it sound like Morgan Freeman, it can do it. Sound effects, lightsabers, tie fighters, it can create all of that. Another company is Well Said Labs, which basically takes text and turn it into audio. Anyways, these are all existing today, and I’ll put those companies in the notes. You should absolutely look at them and play with them. I’ve been spending a lot of time. It’s unbelievable how good it is. Okay, what does that mean? We take these tools and the beauty of technology is it’s combinatorial. You can add things together and they all build on each other, like building blocks. So you can start to put all of this content together. And I’ve sort of given you little examples of that. You type in some text and you say, create a photo of a sweater. for millennial women in the winter in Sweden. That would be the text prompt I might give to something like Dali or, I don’t, it will then create the photo. I will then show the photo and I will say, create a product description based on this photo, it will create the product description. I now have a product and a product description I could put on an e-commerce site immediately, assuming I could manufacture the sweater or whatever. So suddenly, I’m an e-commerce company, I’m a designer. You can do picture to picture. I can take a photo, let’s say a painting of Napoleon from the 1800s, upload it, and I would say, please make a picture of Joe Rogan in the style of this Napoleon. It will turn Napoleon into Joe Rogan. So you can start to basically create portraits. If you wanna be a painter or a portrait artist, and you’ve always wanted to do this in your life. You can do this today. Now, you don’t need to go to school. You don’t need to train. You could be a painter or a portrait artist today. You could be a product designer today. The other thing you can do is you can do like, speech to text. Let’s say I do an interview with someone on this podcast. I upload the audio file and I say, please translate that entire audio file to text, which is pretty easy to do. Now, that’s pretty common. I then take that text that has been generated and I slap it all into ChatGPT and I say, based on this transcript, write me a five-point summary of the interview between Jeff and Martin Reeves. ChatGPT will then write a short article with the five main points from that two-hour interview. I can then say, please create a cartoon photo of both Jeff and Martin, which it can do. Now I say turn this into a video where Jeff and Martin are saying have a 30 second interview of the avatars talking that highlight the five main points we’ve just generated. I can generate a five minute video with cartoons of myself, whatever, saying the five main points from the, that all takes 20 minutes, all of that. That’s insane. Like that’s what we’re looking at. It’s like. And those are just simple combinations and you can kind of see where it’s going. Where it’s like anybody is going to be able to create a cartoon now. Anybody is going to be able to create articles. Anyone is going to be able to design products. Anyone can be a painter. Anyone can write poems. Anyone can write rap lyrics. If you want to create your own rap CD, you can generate the music in the style of a famous artist. Create the album cover. you can do all that in an hour now. It’s not gonna be awesome today, but it’s pretty good. And this is only a year into this. So basically high quality creative content is gonna be super cheap, it’s gonna happen really fast and pretty much anyone can do it. It’s gonna democratize all of this. And then you can start to do refinements, and you can start to do combinations, and you can do better. So the argument is like, you don’t need to know how to paint anymore. You just have to be really good at writing prompts for a painting AI. So the skill turns out to be, how good are you at writing the prompts that tell the AI what to do? And then when the output comes back, how good are you at debugging that, or refining it, or feeding it back into the AI? and say, this was good, but I want you to make it like this a little bit more. So that’s kind of what we’re doing. And I’ve been playing with this a lot and it works. And we’re only a year into this. Imagine where this is going to be in two years. Anyways, that’s kind of a way to think about generative AI within content. And you look at the various types of content, and then you can immediately start. combining them. Oh, there’s a fun company called Lexica, L-E-X-I-C-A. Lexica helps you do prompts. So you put in a photo, let’s say someone has a really awesome piece of AI art, like a super villain comic book that they’ve generated, and you see it online. You can take the picture, you give it to Lexica, and they will tell you the prompt that was used to create that. So it’s like you’re reverse engineering the prompts of high quality content. And what you can do is if you see a photo you really love, you say, oh, that’s amazing, but I don’t want it to be a super villain. I want it to be of someone else. Well, you just put it in, get the prompt, and then you change a couple words in the prompt to what you want. And then it creates a new one that looks pretty the same, but it’s more what you wanted. So you can start to take other people’s work, reverse engineer it, fiddle with it, create your own. Anyways, there’s a whole lot of this going on. It’s fantastic. All right, that’s sort of half of what I wanted to talk about today. The other half is, okay, what does this mean for business, which is my area. So for this, I’m gonna basically give you four points, and then I’m gonna tell you who I think the winners are and who the losers are. And, you know, spoiler alert, I think Netflix and Disney and Amazon Video are in deep, deep trouble. like serious disruption. Okay, first point, as said, this is a reiteration. AI makes prediction and content creation fast, cheap, and pretty much usable by anybody. Now, if you’re writing code, no, it’s highly technical, you need a coder, but for a lot of these create, anyone can make a cartoon. Right, it doesn’t take a special skill. So for a lot of it, it’s totally democratized. Okay, that’s point number one. Point number two. When you think about AI content creation and AI prediction in general, you can put it into two buckets. There’s autopilot and there’s copilot. Autopilot means, look, you don’t need a human being involved at all. Like you don’t, it just runs on its own. The car drives itself. The plane can fly itself. You can have an AI generating a thousand different types of shoes per day. You know, the photo, the design, the… product description, posting it online on its own, testing it, and then at the end of the day, it gives you a little summary. By the way, we tested 15,000 different product designs today. Here’s the three we think are the most valuable. That’s autopilot. Can kind of do it on its own. A lot of stuff when you get to creativity, you’re gonna be copilot, right? It’s like, okay, here’s a picture. You’re an artist, you’re a painter. Let’s start with the AI. It will generate 20 images based on prompts you’ve given it. I would like to show a picture of a mother and a child in the style of Van Gogh on a sunny day, but I would like it to show sort of stress and worry. You could give a prompt like that. It would generate a bunch of, you take what it gives you and you start to refine them and combine them with other things. So it’s sort of like you’re in the business of prompt engineering. Now there’s a, I think it was Vitalik who said this. He said like, you know, the Ethereum founder, he said like, in the future, we will tell the AI what to do. The AI will do it and then we will debug it and fix it. So it’s like three steps. So this one, you write the prompt for the AI, it generates lots of painting and stuff. So it does the building for you. Then we take it and we start to combine it and mix it and debug it if it’s code. So we work on the front, but you could start doing paintings. Okay, that’s more of a copilot scenario. And for code, that’s definitely the way it is because it’s not like you’re gonna tell the AI to write code for how to do the ERP system for a business and then just plug it in. No, you gotta debug it and do a lot of human stuff on the backend. Okay, but autopilot versus copilot is a good way to think about it. Copilot needs a person. And then the question, which is sort of the point of today, is what kind of person do you need as a copilot? Now if you’re just making paintings and cartoons, anybody can be the copilot. If it’s highly technical work like legal work, the copilot probably has to be a lawyer, right? Unless it’s something simple like a rental agreement in Santa Barbara. Okay, professional. If it’s a programmer, if it’s a doctor, okay. You can sort of have a spectrum of if this is a copilot scenario, can anyone do it? No. The reason I think Disney and Netflix are in trouble is pretty much anybody can be creative. There’s a lot of people sitting out there doing regular jobs, working in retail, who are highly creative. They just don’t have the skills to create cartoons or comic books or TV shows or movies. But if you give them the tools, they can do it. So the creative spaces, yeah, there’s a world of people who can create a lot of what Hollywood does. Okay, so that’s sort of point number two for today. Autopilot versus copilot. Now three and four is sort of my area. Question three, how does this impact a platform business model? Question number four, point number four, how does this impact a linear business model that is mostly in the business of content creation? That’s Netflix, that’s Amazon Video, that’s Disney, that’s pretty much everybody in Hollywood. Okay. I’m going to put a link in the notes. I did an article a couple weeks ago called The Winners and Losers in Chat GPT. And that was sort of preliminary thinking, but I think it’s pretty right. And I basically said, look, the biggest losers are going to be professional writers because talking about content. Journalists, opinion piece writers, authors, copy editors, professional editors, poets, marketing and business writers. That’s all going to be democratized. gonna come way down in price. And the biggest winners would be the platforms, Facebook, Twitter, probably Amazon Publishing, because as there is an explosion of people and digital agents, humans and digital agents, that can create content, it makes anyone in the content business facing a lot more competition, that’s bad. But it makes those companies that controlled demand like Facebook and Twitter, it makes their platform much stronger. You know, it’s a lot harder to, you know, this idea that look, there’s a lot of quality content out there, but there’s too much. I need a matching engine to tell me what to look at. So if you’re that intermediary, like Facebook, you’re in a much stronger position as the supply of content explodes. So it’s gonna make them stronger. Those were my two points, but that was mostly about text. Okay, now that we’re talking about video, cartoons, music, we’re talking about all types of content, not just text. you can sort of make the same conclusions. The biggest winners are gonna be audience builder platforms, YouTube, TikTok, GitHub. Because the sea of people who are making good videos on TikTok is gonna explode. The sea of people that are making cartoons, short movies, television shows, high quality stuff, music, is all going to explode. The whole content creator side of these platforms is gonna go way up. That’s going to be fantastic if you’re one of the two or three platform business models that control demand and are very good at matching. Two other types of platform business models to think about. Marketplace platforms. You know, you’re selling goods, you’re selling services. Now, in a marketplace platform, Taobao, Amazon, we’ve always assumed that the merchant, the supply side of the platform, is a business or a person. It’s someone selling shoes, it’s someone selling paint. There’s this idea that it doesn’t have to be a person, a co-pilot using a lot of software and digital tools, including AI, it could be an autopilot. What if we have completely automated digital agents that create and sell goods on something like Amazon without any people? So the product side, the merchant side, we could see completely digitally created suppliers. The distinction I would say is human agent versus digital agent. So I think the marketplace platforms, especially in simple product creation, might change. And then the other type of platform I talk about is learning platforms, like Google. It’s going to impact that. I’m still trying to figure that out. So three types of platforms, they’re all going to get impacted. The audience builders, YouTube, TikTok, GitHub, they’re all going to do tremendously well with this. So we can put them in the winners category. And then we get to the linear business models. And this is where things are going to get bad. Now, I wrote an article, I mean, I did a podcast earlier this year, podcast 136, which was called basically why I don’t like most digital content businesses. And I basically said, look, I don’t like the content business. I think it’s a bad business most of the time. But there’s only a couple types that I do like. I said there were four basically. I kind of said, you know, I like… anything that’s a platform business model where you’re not creating the content, you’re sharing the content like YouTube. That’s and you get network effects and you have very low operating costs. I like that model when you’re not necessarily creating it too much. I like content creation where you create valuable intellectual property that then has value and the example I gave with Disney, you know, they create Iron Man and they own that property forever. And then I like content businesses that are paired with service businesses or with communities. So McKinsey does a lot of content creation. It’s very, they write a lot of good stuff about business, but they pair that with their consulting business and they make their money there and that’s a good model. That’s pretty much what I do. If you compare your content creation with a service business or you can compare it with a community. which is fantastic. You could argue that that’s what most religions kind of are from the business model side. Someone wrote a book at some point, there’s your content, and then they built a community around that. People go to the churches of the mosque. Not to insult people are religious, but there is a mechanism there that’s very effective. So you can say, look, there’s a couple business models if you’re in the content creation business that are good. When I look at generative AI, what this looks like to me is the power in distribution is gone. You know, you can stream anything anywhere. There’s going to be a sea of content coming. It’s going to be very high quality. The strongest position you can be in is if you are the intermediary between the sea of content and what people want. If you have the matching algorithms and you’re very good at that, which is a company like TikTok, that’s what they do. Anyone else who’s just purely in the content business, things are going to get pretty rough. So here’s, I’ll give you five winners and two companies I think are in deep trouble, types of companies. Okay, the winners. As I said, platform business models. So audience builders, mention those. Learning platforms, mention those. Content, where you’re creating content or you’re pairing it with services or community building. I just mentioned that one. That’s one, two, three types of winners. If you are creating creator tools and you’re selling those to people that… then use them. Adobe is an example of that. That’s what they do. They create tools for people who create things. Photoshop, Adobe Premiere, all of that. And then you basically get a standardization network effect, you charge good fees. Adobe has a very good business model. So all of those as generative AI takes off should get better. And then the last one, so that’s one, two, three, four, five, and I’ll list these in the notes. Certain types of niche content will do well. I don’t think anything in generative AI is going to impact the NBA. Sports, ESPN. It’s people doing competition. You can’t really replace that with generative AI. So sports content looks pretty good. Game shows looks pretty good. Reality TV looks pretty good. certain types of famous content creators will do very, very well. But that’s going to follow a power law, which means 1% of content creators are going to do well and 99% are going to be slugging it out because everyone’s creating good content. So if you’re a content creator, there’s going to be a brutal power law. First mover is going to matter there. So those five types of winners. Who’s in deep, deep trouble? content publishers. If you are not creating the content, but you are sort of publishing it like book publishers, this is what they did. They didn’t write the books. They would just sort of publish them and they would go in the bookstores or they’d upload them to Amazon because they didn’t have distribution and they didn’t have the customers. They were just the publishing step. Well, they’re all gone. I can write a book and I can upload it to Amazon directly. Book publishers have been cut in half in the last 10 years. Any sort of content publisher is probably dead. Standalone content creators. If you’re creating content in-house or if you’re contracting, buying it and then doing it, that’s Netflix. And that’s Disney. And that’s Amazon Video. And that’s most of what goes on in Hollywood. These smaller production companies, these smaller TV studios, Paramount. CBS, NBC, and it, because if you think about how these businesses, how did they do well? Did they control the customers? No. They’ve never been strong on the demand side. Yes, I like Netflix, but if I don’t like the shows, I’ll just switch over to Disney or I’ll just watch YouTube. They’ve never had that much of a hold on the demand side. Their power has always been on the supply side. And they had two big weapons. They were relying on scale. That we have a hundred, we can finance 20 movies per year. Each movie costs 50 to 200 million dollars. We have a budget in content creation. That’s our studio, right? That would be a Hollywood studio. And we spend the money, we hire the directors, we hire the producers, we hire the scripts, we film the thing, we spend two years, we develop it, and we create our movies and we put them out. So a lot of their power was we have a scale advantage, which is what Netflix has been doing. They’ve been trying to beat everyone by saying, you know, we are so big that we can outspend you on content. And it turns out when you outspend people on content, you create a lot of garbage content. But it was still a scale effect in content production. Well, what does generative AI do? It just made content production incredibly cheap, incredibly fast, and anybody can do it, pretty much. Anyone can create an animated movie now, anybody. Animated movies, old friend of mine is sort of big in this field. They were in Disney animation and… Now they work over at Netflix and animation. Those movies take two to three years to make. Animated movies take a tremendous amount of time, especially a company like Disney. Disney’s engine forever has been we create animated movies that have music and characters that children love. That’s how we got Mickey Mouse, Snow White, Cinderella, Belle. Aladdin. That’s the engine. It’s the music plus the animated movies plus children love them. And then those characters fuel their merchandise. It fuels their spin-off content. It fuels their theme parks. That’s been the engine of Disney. Well that core engine has just been democratized. Anyone can make an animated movie now. Maybe not this year. Give it another year or two. We can already see the tools. So their idea that we have all these skills in animation that you don’t have, well, everyone’s going to have those skills. Well, we have money, and we have supply side power in our scale. Well, that doesn’t matter either. So their big weapon has just been wiped out. The barrier to entry and their competitive advantage based on economies of scale and content production has just been, that’s just a body blow. So what do you do if you’re? power has traditionally been in content production and content production has just been digitally disrupted. I think the only move, what can you do? Can you try and move on to the demand side and lock people in? Customer capture, switching costs, not really. People don’t like that in content. They don’t like being locked in to their streaming service. They’ll stay there a little bit, but they like to move around. No. probably the only move they can do is they have to start to become a market. They have to become platform. They have to say we do some content production because that’s our legacy, but our core business is now a YouTube. We’re a matching thing. As the sea of content explodes, we are the intersection. So they have to become platform business models. I don’t know why Netflix hasn’t done this. Anyways. And then you start shifting away from… We are a mass market company that creates Frozen because all the little girls of the world love Frozen. That’s a mass market approach. We have to become a platform and do thousands of micro brands. We have to have the iconic little brands for every little sub-niche in the long tail. The only way we can do that is with a platform. So that would be my advice to Netflix and those is you need to become. your traditional content creation, which you’re very good at, and you need to become a platform as soon as humanly possible. Because this other business is gonna get a lot more difficult. You’re gonna start to see random people coming out of nowhere, five people who are all maybe art and design students who are creating unbelievably good animated movies coming out of nowhere. That’s gonna happen, just like we’ve seen it happen in other areas. Anyway, so those are kind of… how I put this together. The winners, I’ll summarize it and then I’ll finish up for today. Some niche content is going to do well. It’s not going to be impacted by this. Sports is a good example. And some creators of content are going to be incredibly powerful. We’re going to see like Joe Rogan type things where random people sitting in their basement now become the biggest content creators in the world in various sub niches like rap lyrics or superhero movies or whatever. Content plus community or content plus services, that’s a good business model. Audience builder platforms, learning platforms, both good. And anyone creating creator tools that get a lot of standardization network effects, like Adobe. That’s five business models, look good. The two that I worry most about, anyone who’s in the content publishing business, like book publishers, and anyone who is a standalone content creator. whether they’re doing that in-house or a combination of we create our own movies in-house and we also do a lot of purchasing. They are not going to be able to compete with the sea of free content that’s going to be coming that is very high quality. That’s going to be a brutal game and their old power based on economies of scale and content production is basically being disrupted. The same way… Blockbuster’s big source of strength for a long time was we have all these stores. And then one day, those stores weren’t valuable anymore. In fact, they were a liability because they cost so much. So these in-house content production teams at places like Netflix, that’s a cost structure that’s gonna put them at a disadvantage to a small team that’s just using AI tools and has no cost structure. So we’re gonna say the same thing. Anyways, that’s kinda my current thinking on this. I’m gonna send out probably… a couple daily articles to subscribers in the next week or two, sort of detailing this out a lot more. But that’s what I’ve been working on the last couple weeks. Okay, I think that’s kind of a lot of me talking, a lot of theory for today. I am late this week. I’m sorry about that. I’m about a week behind in terms of podcasts. I was bouncing around Malaysia and Singapore and it was just… It wasn’t working very well. So I sometimes have to be in the right mood. So I’m one week behind. I’ll put out another podcast in the next couple days to catch up. But yeah, it was a pretty great, I spent a lot of time in Kuala Lumpur. I was writing a book. I think I’m pretty much done with the book. I sort of just burned it out in about five to six days, sitting in cafes in Kuala Lumpur, which is really a, it’s really a nice place. I haven’t spent a lot of time there in the past. It’s not like a spectacular tourist city where you go and go, oh my God, this is amazing, but it’s really a pleasant, pleasant place to be. Like I was really, I had a great time. The best description I’ve heard of Kuala Lumpur is like, it’s like a city that happens to be in a jungle. Like it’s very much like green and there’s mountains and it’s really a pleasant place to wander around and great cafes, amazing food. Indian food, Chinese food, Malaysian food, great. I had a good time and I just sort of camped out in a cafe and typed like crazy for five to six days. So yeah, it was great. Anyways, okay, that’s it for me this week. The main takeaway, I always like to repeat myself, the main takeaway today, the key concept, generative AI. That’s kind of a big, big deal, and I really encourage you to go play on a couple of these sites. It’s not, you gotta log in, there’s a little bit, it’s not that hard. It’s really unbelievable. I mean, I’m just stunned. So anyways, that’s it for me. I hope everyone is doing well, and I will talk to you probably in a couple days. Bye.
I write, speak and consult about how to win (and not lose) in digital strategy and transformation.
I am the founder of TechMoat Consulting, a boutique consulting firm that helps retailers, brands, and technology companies exploit digital change to grow faster, innovate better and build digital moats. Get in touch here.
My book series Moats and Marathons is one-of-a-kind framework for building and measuring competitive advantages in digital businesses.
Note: This content (articles, podcasts, website info) is not investment advice. The information and opinions from me and any guests may be incorrect. The numbers and information may be wrong. The views expressed may no longer be relevant or accurate. Investing is risky. Do your own research.