GenAI in Hollywood: Threat or Opportunity?

My Presentation at the MoffettNathanson Media, Internet and Communications Conference

Doug Shapiro
29 min readMay 29, 2024

In the grand tradition of windowing in media, starting December 2023, all my writing will be posted first on my Substack, The Mediator. This essay was posted on The Mediator one week ago.

Sign up for free to get The Mediator delivered to your inbox “day-and-date”!

On May 14, I spoke at the MoffettNathanson Media, Internet and Communications Conference about the potential effect of GenAI on Hollywood. I spent the first 10 minutes trying to summarize the thesis and a lot of what I’ve written on the topic. Then Michael Nathanson, Robert Fishman and I had a wide-ranging discussion, touching on:

  • Why GenAI applications in video are advancing so quickly.
  • The tension between GenAI being an opportunity and a threat to the studios (and the difference between sustaining and disruptive innovations).
  • The structural hurdles for the studios to adopt GenAI tools.
  • Why Jeffrey Katzenberg’s warning that AI will reduce the costs of making an animated film by 90% might not be too far off.
  • The increasingly-extreme power law distributions of popularity and what it means for established and emerging talent.
  • Why the music labels have been able to maintain primacy in the value chain even as the barriers to creating music have fallen, and whether it’s a hopeful comparison for Hollywood.
  • How Netflix could go from disruptor to disruptee.

The full video is here. Below is the transcript, edited for clarity (and to correct more than a few times I misspoke!).

Robert Fishman

Doug is now going to enlighten us all about the very important topic of AI and how that’s going to impact Hollywood. Doug was at our BCG (Boston Consulting Group) AI conference in November. It was one of the most talked about panels there and so, we thought to bring him back and enlighten everyone here about how the disruption and what’s happened since.

Michael Nathanson

Just to know how prescient Doug is, he told me this and I have to share it, even though he told me in confidence. He was known inside Time Warner as “Doctor Doom,” because he kept scaring the company about threats at Netflix.

Doug Shapiro

Well, Kevin Tsujihara (former CEO of Warner Bros.) used to call me that. I preferred to call myself the CUE, the chief urgency office.

Michael Nathanson

When Doug talks about changes in business models and urgency, I listen to him, because you pretty much laid out the Netflix threat, you and Spencer (Wang, VP Finance, IR and Corporate Development at Netflix). Some listened, some didn’t.

Doug Shapiro

I think that’s a good context for what we’re going to talk about. I was running investor relations at Time Warner in 2008, 2009, and at that time, we used to mostly worry about Netflix’s effect on our DVD business. They’re mailing red envelopes. They’re using the first-sale doctrine and arbitraging that. We thought streaming was kind of a joke.

And then at some point, they did the Starz! deal and they got pay-1 movie rights for Sony and for Disney and 2,500 movies. And we’re like, “Oh boy, this is a problem.” And that was in 2008, 2009.

A lot of what we thought was going to happen — maybe not quite the extent or the speed — but a lot of what we expected would happen, happened over the next ten years. And we did a lot of things to try to stop it, but they didn’t work. And so, I think we’re back there again, which is what we’ll talk about today.

I’m going to spend a few minutes trying to answer this question, whether GenAI is an opportunity or a threat. Then I’ll show a couple demos and then we can talk through it.

The premise, the 1 or 2 sentence summary that I’ve been using over the last year, year and a half, is this idea: The last decade in TV and film was really defined by the disruption of content distribution, and the next decade is going to be defined by the disruption of content creation.

What I mean by that is that the value that resides in any value chain is a function of the moats in the value chain. That’s chapter two of every microeconomics textbook. No barriers to entry, no moats, no profit long term. And historically, in TV there are these two big moats.

There was a moat around distribution because you needed to own the underlying physical infrastructure to distribute content. Very capital intensive. You had to dig trenches and lay fiber and launch satellites, build antennas, whatever. And there was a big moat around content creation because it’s risky and it’s expensive. And what happened is that one moat fell.

The internet came in and unbundled information from infrastructure and the cost to move bits around plummeted. And that paved the way for Netflix, which came in and ran the whole disruption playbook. It started with crappy stuff that people basically ignored, and it got a foothold and it got better and better and better.

And today, Netflix is arguably the most powerful company in Hollywood. We all know what happened as a result of that. And these guys (MoffettNathanson) have all kinds of awesome charts on this. I always use Craig (Moffett)’s chart on pay-TV subs and how it’s going down and how the pace of decline is accelerating. And (Michael and Robert) have these charts on viewing declining.

But this chart is kind of the bottom line. I mean, maybe we could show the stocks, that would be the bottom line. This is a different bottom line. The literal bottom line. And what this is showing is the performance of the video business for the big five media companies from 2018–2023 — although at the beginning of this period they were the big seven media companies.

In any case, if you look at their combined video businesses, traditional TV, streaming and studio combined, revenues are up a little bit, but profit margins are down 1,000 basis points over this time frame. And profits are down 40%.

Right now everybody’s running around with their hair on fire in the industry, there are a lot of negative headlines. When you look at what’s going on — whether it’s questions about Bob Iger’s succession, Paramount being for sale, RSNs going bankrupt, the writers’ and actors’ strikes, box office struggling to get back to where it was pre-COVID — all of those things you could trace back to the last disruption. You could trace them all back to the fact that the barriers to entry fell in distribution. Netflix came in, squeezed out the excess profits, and now there’s less money to go around. Everyone’s running around crazy because there’s less money to go around.

But even as one moat fell, another moat got stronger and deeper. And that was the moat around content creation, because the cost to make stuff actually went up. As the distribution costs fell, costs went up, and you can see here in this chart on the left, the average tentpole film budget doubled over 20 years. The average cost of a premium hour of scripted TV doubled over ten years.

Another way of thinking about this is that distribution was subject to Moore’s Law. That’s Moore’s law on the left and on the right, it’s health care premiums and sushi. And so I think the real question is what happens if Moore’s Law comes for content creation?

You have an industry that’s already reeling from the last disruption. What happens if there’s another disruption coming?

Just to be clear what I mean by this, I’m not making the case that we’re going to have a GenAI-created or -enabled blockbuster movie any time soon. What I’m arguing is that GenAI will democratize high quality production, and in doing that, it will exacerbate a low-end disruption that’s already underway from creator content.

Everyone’s seen this chart on the left. YouTube is now the single largest streaming service in the U.S. to televisions, and it’s share is going up. For years, the argument was that YouTube is not competitive with TV. It’s people slipping on ice and cats playing the piano and whatever. That it’s not the same use case.

But this is exactly the same use case. This is watching on a television. If you think about what’s happening, this is just a redux of what we already saw, a different kind of low-end disruption. And in this case, Netflix interestingly goes from disruptor to disruptee.

If you think about how disruption plays out, typically, again, it comes from the bottom. First, cats playing the piano. Then, as it creeps up the performance curve, the first thing you go after is the least demanding customers. And that’s kids. The most popular kids show in the world is Cocomelon, on YouTube.

Then you keep going up the curve and you get to the next least demanding customers, unscripted TV. The most popular unscripted show in the world? Mr. Beast on YouTube.

And so the question is, do these tools, Pika and Runway and Stable Diffusion and Sora, do they then pour gas on the fire of something that’s already happening? And my argument is yes.

What I want to do is just show you a couple of videos just to illustrate how quickly these things are moving, because I’m sure we’ll talk about it, but it’s moving a lot faster than I thought it was going to when I started writing about this a year and a half ago. I thought this process was going to happen over the next five to 10 years, and now I think it’s probably more like three to five.

This first video is from April 2023, made with Runway.

This next one is from January 2024, which uses Midjourney, Runway, Pika and a variety of other tools.

And, of course, you all have seen videos like this from the initial Sora announcement, from February.

One of the hidden elements of all this is that you still need — if you’re going to make an AI video — you still need some kind of technical fluency. As I mentioned with that Borrowing Time video, you had a guy (Dave Clark) who used a bunch of different tools. He has his own custom workflow, and he still needs to know Premiere Pro or DaVinci Resolve. He’s still editing all this together.

The future is you’re going to have holistic platforms where you can start with an idea and end up with a movie. And a few weeks ago, I got beta access to this thing called LTX Studio — very early, very janky, a lot of problems with it, but I basically took this thing and in two hours I had my own movie, not having any editing capability at all.

Anyway, I think that what Sora really shows is that AI can solve video. What I mean by that is, if you think about something like self-driving cars, we may never have a self-driving car because we may never societally accept that it’s inevitable that a self-driving car is gonna kill people. Video is an illusion to begin with. It’s twenty-four still images per second. When you walk in the theater, you know you’re sitting down to see something that’s not real. So, it’s got a much lower bar.

And what we’re seeing is that scale has tremendous value. And maybe we’ll talk about this as well. But that was really the main takeaway: AI will solve the video problem.

The last thing I’ll say — and then I’ll stop and we can hash this all out — is that I think another question is how quickly could this happen? I drew a parallel between the last disruption and this pending one. And I think in a lot of ways, this could happen much faster.

One of the things to keep in mind is just the math of this, how overwhelming the math may be. Last year, Hollywood put out about 15,000 hours of TV and film, and there were 300 million hours uploaded to YouTube. If .01% of that is considered competitive with Hollywood, that’s twice Hollywood’s annual output.

To me, there’s always this Game of Thrones analogy where we can argue Disney versus Sony versus NBCU versus WBD versus Netflix versus Amazon — it’s like the Lannisters versus the Snows versus the Targaryens, while the Army of the Dead is amassing at the wall. That’s what the industry needs to be worried about — the Army of the Dead. So that’s one thing, the math is overwhelming. This isn’t one disrupter. This is 10 million disruptors, 20 million disruptors.

The other thing is that the last disruption took a very long time because a lot of things had to happen. A lot of things had to go right. You had to build out broadband infrastructure. People had to adopt broadband, people had to adopt streaming. They had to do all these things that are now in place. As I mentioned, YouTube’s already the #1 streaming service in United States to televisions. If there was a great AI-enabled show on YouTube tomorrow, it could be the number one show in the United States tomorrow night.

And then the last thing is there’s an incredibly complex ecosystem in Hollywood. And what I would say is my experience with the studios is that they’re very much handcuffed right now because of that. To put it a different way, when you think about the disruption of distribution, the biggest media companies should have been somewhat agnostic to how they distribute (but they were disrupted anyway). Creation is much more core to the DNA of what they do.

So, I will stop there. To answer the question I posed at the beginning, I think it’s more of a threat than an opportunity. And maybe we can talk it through.

Michael Nathanson

Can I ask you, you mentioned that the past 18 months since you started this, you’re now convinced it’s going to be quicker. Is there any other nuance, takeaways that you’ve had since you’ve seen the next generation, besides the speed of change, anything that nuanced in terms of that’s evolved your thinking?

Doug Shapiro

Well, I think it may be why it’s happening so fast, which I didn’t appreciate at the time, which is there’s really three things happening: scale, open source and composability.

I showed a chart before from this paper called “Scaling Laws for Neural Language Models.” And it makes the point that as the size of the training set goes up and the number of parameters goes up, and the amount of compute goes up, the loss rate of these things continues to fall. I’m sure it hits some wall someplace, but it’s basically saying that almost regardless of the model architecture itself, scale solves problems. I didn’t appreciate that.

The other thing is, it’s amazing how much of this stuff is happening out in the open domain, because, I don’t know if people talk about it, but it’s amazing to me that probably the biggest existential threat to Google is OpenAI, which only exists because Google DeepMind put out a paper called “Attention is All You Need.” And if you read the technical paper and look at the citations, everything they did is just a combination of stuff that’s being released publicly. So, this kind of open source ethos within the AI community is dramatically speeding the pace of things.

And then you have composability. For instance, Sora is a transformer diffusion model. What that means, it’s both a transformer model and a diffusion model. And you can basically use these models like Legos.

So, that was what’s been so surprising to me and what’s driving the speed.

Robert Fishman

So building on your analogy, it sounds like either winter is coming or winter is here. The question as it relates to our media companies and the coverage that we follow, in terms of the opportunity piece of it — how hopeful are you that AI can actually bring down the production costs? Is that the big opportunity for them on the cost side? How else would you characterize some of the opportunities that come out of this?

Doug Shapiro

I think there is an opportunity. But, I would paraphrase something that I think Ben Thompson said here last year, which is that it’s kind of like asking, “Will the internet be good for magazines?” We all go to school and read about the five forces, but there’s one force that matters most, and that’s barriers to entry.

I think there are two problems. One is that the fact that you are going to see that lower barriers to create content is a much larger force than potentially lower costs. The second thing is that I think the way the studios are approaching this is they are somewhat hamstrung.

I wrote this piece last year, called Is AI a Sustaining or a Disruptive Innovation in Hollywood? And what I meant by that is that often technologies are not inherently sustaining or disruptive. It depends on how they’re used. And so Hollywood is approaching this in a very sustaining way, where they’re trying to figure out, how do I plug AI tools into my existing workflows? How do I use AI for localization services? Or how do I use AI for concept art? Or post-production, VFX, etc. And in totality, maybe you’ll get 10–15% savings, but then you’re going to have people who are approaching this with a completely clean piece of paper and have the potential to see much more dramatic savings than that.

Michael Nathanson

And then you touched on it before, but how do you think this ecosystem where you have all these call options on people’s time and value, how the complexity of the system impacts the adoption of AI in Hollywood? You think about all connecting pieces. Walk us through why that deters change.

Doug Shapiro

Well, from what I’ve seen, all of the studios to their credit are taking this very seriously. They’re looking at it, they’re paying a lot of attention to it. But I do think that they are very leery of doing anything that would be perceived to reduce labor demands.

You saw how AI went from being a non-issue to almost the central issue in the writers strike. And so I think they’re really concerned about that perception. And AI is a little bit of a third rail in terms of talent.

But the other thing is that the studios don’t actually make the movies, they attract the talent that makes the movies. They can’t force anyone to use the stuff. They can’t say to Christopher Nolan, “Hey, we love that project, but you just have to use these five tools.” That’s not going to fly. And I think an interesting analogy is what has happened with virtual production.

Its main application in Hollywood is filming on what’s called a volume. It’s a soundstage with a bunch of LED screens that use the LED screens to produce the set, as opposed to going on location. And you’ve seen empirically that it saves a lot of money, but they can’t force anyone to use it.

And what’s happening now is that virtual production is basically dying. Warner Bros. built a volume in Leavesden under Jason Kilar, and then after he left, they dismantled it because they can’t get anybody to use it.

And so I think that’s the problem — it’s twofold. It’s the sensitivity on the topic and the fact that the studio can’t actually require it.

Michael Nathanson

Right. Directors are going to say, “Why am I going to do that? I won Academy Awards doing it my way.”

Doug Shapiro

Ryan Coogler just did a deal with Warner Bros. where he’s going to get back the copyright to his movie in 25 years. I think the bargaining power of intermediaries in media is diminishing anyway. As all of these tools get democratized, the role that the studios or labels or publishers or whatever have always played between creators and consumers is easier to circumvent. And so they don’t have the bargaining power to say, “You have to do it this way.”

Robert Fishman

And you just kind of hit on this a little bit. But I’m curious if you want to go and make a controversial statement for us that the strikes and the settlements that came out of that, did that actually enhance the ability to allow AI to take over or where do you think things shook out after all the settlements and strikes?

Doug Shapiro

No, I mean, I think that the settlements were good in the sense that it at least establishes rules of the road. But there was nothing in there that wasn’t common sense. Like, you can’t use someone’s likeness, you can’t recreate someone’s likeness digitally without getting their approval. That’s not a really shocking concept. But at least it sets the stage to move forward with AI.

I think there’s things like you can’t force a writer to use AI. So there’s this element — you can’t force anybody to do anything they don’t want to do. I think the bigger issue is really more the social issues, that’s the bigger problem.

Robert Fishman

So, we talked about cost a little bit. Maybe digging in there a little bit further. Jeffrey Katzenberg has been out there vocally talking about that AI can eliminate as much as 90% of the cost of making an animated film. Curious to hear your thoughts, given what you have seen and what you’re able to share on that in terms of the labor intensity piece of making an animated movie and the amount of time that goes into that.

Doug Shapiro

I think animation is one of the first places you’re going to see it, both because of the labor demands and because the bar is a lot lower if you don’t have photorealistic humans. If you have a talking cat or something, then it’s a lot easier to do. There’s no uncanny valley with cartoon cats.

But I think the math of it is probably not that crazy in the sense that I think in — rough justice — if you look at a blockbuster, the average big film, something like 15–25% is above the line costs, and then about 50% is below the line costs and then the remaining 25–35% is post, most of which is VFX.

So, if we were to assume that you still need all of the above the line people, you still need the same creative talent, that means that the rest of that is up for grabs (75–85%). And right now, again, if you take the average budget, when you figure out just the below the line and post costs, it’s probably like $1 million a minute for most big movies.

And the question is: over time, does that cost converge with the cost of compute? Could that be $100,000 or $10,000 or $10 a minute? I think that’s kind of the trajectory. That’s what I was saying before. What happens if you bring Moore’s Law to this, as opposed to rising labor costs.

Michael Nathanson

But just stepping back, I wonder what happens longer term to the above the line talent. Katzenberg hasn’t yet said, “Hey, they’re all out of work.” But given what you’ve said, just the math of the size of Hollywood’s output versus what’s on YouTube every day, it feels like eventually those gates are going to fall too.

How would you advise the upper end talent, the most creative people, to prepare themselves for what’s coming?

Doug Shapiro

Well, I think there’s another question that you’re maybe asking, but I’m not sure, which is what is the role of humans in the future? If the question is about how much AI is too much AI, I think that’s really more about consumer acceptance than it is about technology. And you can think of these kind of four scenarios of how much AI consumers will accept. It’s not a technological issue, it’s a consumer acceptance issue.

So, the first scenario is that there’s some arbitrary level of AI that people are like, “Nope, too much. You crossed the line.” And I think we can dismiss that one because VFX has taken over the film business and people still go see Guardians of the Galaxy.

The next level would be that people draw the line at artificial humans — that anything that’s supposed to be a person has to be a person. I think that’s viable.

The next level would be that consumers draw the line at artificial ideas — that you need to have a human, or “human in the loop” as they say in the AI world. A story is supposed to evoke an emotional response. And a LLM (large language model) may understand the semantic relationship between tokens, but it doesn’t understand emotions. So, will you always need a human to be able to overlay their own judgment on what is emotive or not?

And then the last scenario is there’s no line. And in that case you just have this dystopian doom loop of infinite AI generated content, which does not sound like a fun place to live. So for me, I think that we’re going to fall somewhere between scenarios two and three.

You’re still going to need real humans in a lot of stuff. And you’re still going to need, I firmly believe, some kind of creative overlay, some human to apply their judgment.

Michael Nathanson

But here’s where I’m going with this, which is, Ben Thompson’s career has existed because of the open internet. Ben Thompson would not be on a stage with us if it wasn’t for his ability to reach us. 20 years ago, it would be the New York Times media columnist David Carr or Jim Stewart — someone who was in a gatekeeper position that pretty much had a moat around him because of the structure.

I think the question I meant to ask you is, with the ability of recommendation engines and just the scale of TikTok and YouTube, is it possible that today’s talent will not be next year’s talent? Because the walls are coming down. The moat that created the star system and Chris Nolan may fall five years from now, just from recommendation engines and the ability to create at a much cheaper cost.

That’s actually my question — does it reorder talent in a way that Ben Thompson would have been the beneficiary of the open internet?

Doug Shapiro

I think to a degree, I think that what’s happening is that all of media is gravitating to increasingly extreme power law distributions. If you go back and read The Long Tail from 1994, the idea was that hits were this artificial construct, an artificial result of scarce distribution. Once there was infinite distribution, there’d be this long tail. And that’s where everyone would be consuming ’80s synth pop or Icelandic ska or something like that.

What The Long Tail missed was that on these massive networks with infinite choice, there are these very powerful positive feedback loops which make the strong stronger.

What happens is when consumers are confronted with infinite choice, they use popularity as a signal of quality, which only feeds on itself. And that’s how you end up with Taylor Swift. That’s how you end up with these ever-bigger cultural phenomena driven by the feedback loops on the network. So, I think that what that does in some way favors the very tippy top. The network entrenches their position — like this Drake, Kendrick Lamar beef or whatever.

Those kinds of things will feed on themselves. And so the top will be the top. But I think as the barriers to entry fall, it means that anybody who’s not in the top is facing pretty much infinite competition.

I think this is the case for creators too. They’ve been somewhat deluded into thinking that they can become the next Mr. Beast. But the truth is that 99.9% of people are going to be stuck in the tail. So yeah, I don’t know. It’s not very… I don’t know if there’s any other way around that.

Robert Fishman

So, let’s take you to the other side of that then. On the industry right now, I think it’s very clear that there’s less capital being committed to scripted content from the studio side, clearly lots of talk, even more so I think as we’re speaking in terms of the future of industry consolidation. So, curious — does all of this that we’re talking about actually slow down disruption?

Or does it accelerate? If you were put in charge of a studio today, how would you approach the business differently?

Doug Shapiro

I think those are two separate questions. I don’t think it slows down disruption because I think the disruption will come from the bottom. I think that the studios, again, because of the very challenging position that they’re in, which is not their fault, but just the structure really works against them adopting these tools.

So, I think that the disruption will come from the bottom, and I think a loose analogy is to think about CGI and Pixar, that it really took Pixar coming in with Toy Story to shake up the industry. And that heralded the end of hand-drawn animation. I think that the studios are probably going to be mostly dragged kicking and screaming into this world.

So no, I don’t think that any of that slows down disruption. In terms of what I would do if I was running a studio, I think there are two broad things. One is to figure out what is still scarce as quality becomes infinite. And it’s things like IP and libraries and the ability to market, to create stars, the ability to foster fandom. I also wrote a piece called What is Scarce When Quality is Abundant that addresses this.

But the other thing is that the model has to change. The business of making content is getting riskier. The studios manage their capital like a private equity investor, but we’re in a venture capital world. The returns are moving more toward power law returns. You need more swings at bat, at lower cost, you need to give the talent more equity, which all ties together. You need to use the network as a focus group, not as a marketing tool. You need to have a dialog. You have to seed stuff onto the network, see what works, and then kill the stuff that doesn’t work and feed the stuff that works.

You need technology and development to sit pari passu next to each other, not the technology guys in the basement. You have to really embrace these tools. I don’t know, I think it’s going to require a lot of cultural change in a very short period of time.

Robert Fishman

So just one quick follow up would be, do you see the value of a library go up or down in this world of AI?

Doug Shapiro

I think probably up in the sense that there’s going to be such a volume of crap. I didn’t mention this, but the vast, vast majority of this stuff is going to suck. As I mentioned before, I made a 2 minute movie in two hours and I don’t know what I’m doing. There’s going to be such a tsunami of bad content that I think one of the interesting lenses to look at this whole thing, and we talked about it a little bit in this in the power law discussion, is as a consumer, what filters do you use when you’re confronting infinite choice?

One filter is the organic signals that arise from the network, like what is popular, but also what I recognize, what has some resonance, what brand have I heard of before? And that’s where the libraries come in.

The other question with libraries is what is the value of the training data embedded in those libraries? That value is twofold. One is, what’s it worth to license that stuff to OpenAI or whomever, but also, secondly, if you were to create a fine-tuned model that was trained on your own content, how quickly could you spin up new iterations of your existing library?

Now I’m talking myself into it. I think library values are going up.

Michael Nathanson

I want you to talk yourself into something here. When we first started covering the music business, GarageBand was introduced on iPads. You go back to Steve Jobs’ big keynotes like, “look, this is going to give every musician in the world the power of a recording studio,” which would be a hard thing to replicate because it’s so hard to get into a studio.

So, during that wave of technology, it didn’t really alter the hierarchy of the industry. I would argue stars are still the stars, and a lot of people made a lot of bad music and a couple of Billie Eilishes have risen to the top. So, why is this different than what we saw in music, where we worried about GarageBand and all the tools that people got in their iPads and MacBooks, but nothing really changed the structure of the industry?

Doug Shapiro

I think it is a good compare and contrast. The way I would put it is that the music business changed a lot, but the labels were still able to maintain their primacy in the value chain. A label, a studio, a video game publisher or a news publisher, they’re all intermediaries in a different way.

And the music business changed a lot. But to your point, Billie Eilish makes a platinum record in Finneas’ closet, and the first thing when she hits it big — or Roddy Ricch or Doja Cat or I think most of the new music breakouts in the last five, six, seven years who all started in the world of self-distribution on YouTube or they blow up on TikTok or whatever. But the first thing they do after they have a hit is they go sign a major label deal. So the question is, if the labels have been able to maintain their primacy in the value chain, could studios do the same thing?

I think the answer is maybe — if the studios are able to replicate what the labels have done. But there are certain idiosyncratic things in music. For one thing, music is incredibly complex and global. And they have the data and you wouldn’t otherwise know that you’re blowing up in Tokyo and you should now launch a Japanese tour or whatever.

But the bigger idiosyncrasy, I think, is how important catalog is. And catalog is still 75% of listening. The labels all own the catalog. And so they have tremendous bargaining power over the DSPs (digital service providers).

And so if you’re Taylor Swift on the best year, she might have like 2% of streams or 1.5% of streams on Spotify. But UMG (Universal Music Group) has 35% of streams. So, you’d rather be in their stable than negotiating on your own. And so I think the labels have positioned themselves really well because of that.

To explore the analogy, let’s stick with Taylor Swift — if anybody could thrive without a label, it’s her. But she’s with Universal, right? When she went to make her movie of the Eras Tour, though, she didn’t go to a studio. She said, “I’ll spend my own 15 million bucks, make this movie and go directly to AMC and distribute it myself and keep all the money.” And so I think the challenge for the studios is that most of their value has come from their checkbook.

Their checkbook is not going to matter as much in the future. So could they pivot? Could they lean into these other things? Again, it’s this question of what is still scarce? So yes, I think that’s a very hopeful analogy. But I don’t think it’s entirely apples to apples

Robert Fishman

Any questions for Mr. Shapiro? Yeah, in the back there.

[question inaudible]

Doug Shapiro

I can’t recite for you chapter and verse what the WGA (Writers Guild of America) and the studios’ agreement was. But it was things like there’s still a minimum number of writers in a writers room, and you can’t force a writer to use AI, and a writer can use AI if they want to.

There’s some inside baseball stuff about how writers get paid. For instance, they get paid differently if the work is original or derivative, so the studios can’t put an idea into a prompt and say, “Here’s our script, can you go fix it?” and then claim it’s now a derivative work. There’s a lot of stuff like that. But I don’t think these issues are the main inhibition. The general problem is that the studios can’t tell anybody to use AI.

They can’t say, “We’re forcing you to use this.” What they can do is they can negotiate budgets and they can try to control it that way. One thing you’ve seen just recently is there’s scuttlebutt in Hollywood that most of the majors are reevaluating the way they structure their content deals, because it used to be that the talent had an economic interest in the back end.

Then Netflix showed up and they instituted this new cost-plus structure, which basically says “We will own all rights in perpetuity in all regions and whatever that thing costs we’ll give you 25% more than that.” And so what it did is shifted all of the risk to Netflix and to the industry, and away from the talent.

Now, you’re starting to see a move back the other way toward sharing risk with talent and giving them incentives to be under budget. And so that could be the way that you force the change, saying “We will incentivize you to reduce budgets” and that might be the catalyst that gets the talent to embrace some of these tools.

Robert Fishman

One more in the back. Marketing and advertising.

Doug Shapiro

I’m not as familiar with that. I think that what you are seeing is just a lot more in-housing. A lot of brands now have their own capabilities in-house. So they’re much more likely to say, “We’re going to take a stab at doing some of the stuff that’s simpler, we’re going to try to do this ourselves, we don’t need a creative agency anymore.”

I think that’s still early days, but that kind of is the question — what is the value? I know someone who runs a small agency and is using these tools to create content and he’s being asked for massive discounts now relative to what he used to charge. The brands are coming in and saying, “Oh, well, we know that you’re saving money, it used to take you three weeks and now you can do it in three days. So, we want this thing to cost one fifth of what it used to cost.” If that is a harbinger of what’s going to happen, it could potentially put a lot of downward pricing pressure on agencies.

Michael Nathanson

The same day that Doug appeared at the BCG conference, we had a BCG partner. And I think they’re trained not to be bombastic, but one of the partners said, what he’s saying in his early consulting is that creative agencies inside holding companies are seeing that pressure. It’s pretty much coming out now that they just can’t hold on to price because of just how easy it is to use these tools to create multiple, multiple, multiple amounts of content.

Doug Shapiro

I think that was one of the things that Ben (Thompson) was talking about at lunch is this idea that we’re going to have so many iterations of stuff. And so I think that the brands are better able to say, “Well, we can just throw tons of shit against the wall and use that to generate our ideas.”

Robert Fishman

Can I leave you with one last question, maybe for the room — if we change the question, “GenAI in Hollywood — threat or opportunity?” If we say “GenAI for Netflix — threat or opportunity?” Because one of the takeaways from what you pointed out before with your graphs, Netflix shifts, right? They go from disruptor to disruptee. So would it be the same answer that you’ve given on the studio front, or do you see this as a different threat or opportunity for Netflix specifically?

Doug Shapiro

Well, you should ask Spencer (Wang) tomorrow, but I think if anyone’s in a position, if anyone has enough bargaining leverage in Hollywood to force change, it’s them. They changed the model once. They probably have more leverage to try to strongly encourage talent to use these tools. And because they do all this stuff cost-plus and they actually line item budget every single production, they’re better empowered to exert pressure surgically inside the budget.

They’re a very progressive company that is good at pivoting and good at leading the way. I think if anyone’s in a good position to weather this, they are.

The bear case is that in five years there’s just a lot of really good content available for free. Again, if the cost structure collapses, it means that you can now generate returns in a new way. You might actually be able to make money by putting a TV show on YouTube. There are going to be all kinds of new models that evolve out of that. So, if you went out five years and there’s actually good quality stuff on YouTube or TikTok decides to extend out the length of videos — do people at some point look up and go, “I don’t know why I’m still paying, by that point, $30, $35 a month for this walled-garden service on which I don’t always find something to watch.”

Michael Nathanson

Hence moving to sports.

Doug Shapiro

Sports, it’s all sports.

Michael Nathanson

Mr. Shapiro, as we call you. Thank you so much for being here. Thank you.

Doug Shapiro

No, thank you guys.



Doug Shapiro

Writes The Mediator: Site: Ind. Consultant/Advisor; Sr Advisor BCG; X: Turner/TWX; II-ranked Wall Street analyst