From a9251ce48232075964d30e26b47732f6169cd6c2 Mon Sep 17 00:00:00 2001 From: Teleo Agents Date: Thu, 19 Mar 2026 16:46:40 +0000 Subject: [PATCH 1/5] extract: shapiro-power-laws-culture Pentagon-Agent: Epimetheus <3D35839A-7722-4740-B93D-51157F7D5E70> --- .../shapiro-power-laws-culture.json | 44 +++++++++++++++++++ inbox/queue/shapiro-power-laws-culture.md | 20 ++++++++- 2 files changed, 63 insertions(+), 1 deletion(-) create mode 100644 inbox/queue/.extraction-debug/shapiro-power-laws-culture.json diff --git a/inbox/queue/.extraction-debug/shapiro-power-laws-culture.json b/inbox/queue/.extraction-debug/shapiro-power-laws-culture.json new file mode 100644 index 00000000..02caef3d --- /dev/null +++ b/inbox/queue/.extraction-debug/shapiro-power-laws-culture.json @@ -0,0 +1,44 @@ +{ + "rejected_claims": [ + { + "filename": "information-cascades-create-power-law-distributions-in-culture-because-consumers-use-popularity-as-quality-signal-when-choice-is-overwhelming.md", + "issues": [ + "missing_attribution_extractor" + ] + }, + { + "filename": "recommendation-algorithms-amplify-or-dampen-power-laws-depending-on-collaborative-filtering-weight-because-algorithms-that-surface-popular-content-reinforce-network-cascades.md", + "issues": [ + "missing_attribution_extractor" + ] + }, + { + "filename": "the-middle-of-cultural-markets-is-disappearing-because-power-law-amplification-concentrates-returns-at-the-head-and-tail-simultaneously.md", + "issues": [ + "missing_attribution_extractor" + ] + } + ], + "validation_stats": { + "total": 3, + "kept": 0, + "fixed": 7, + "rejected": 3, + "fixes_applied": [ + "information-cascades-create-power-law-distributions-in-culture-because-consumers-use-popularity-as-quality-signal-when-choice-is-overwhelming.md:set_created:2026-03-19", + "information-cascades-create-power-law-distributions-in-culture-because-consumers-use-popularity-as-quality-signal-when-choice-is-overwhelming.md:stripped_wiki_link:meme-propagation-selects-for-simplicity-novelty-and-conformi", + "information-cascades-create-power-law-distributions-in-culture-because-consumers-use-popularity-as-quality-signal-when-choice-is-overwhelming.md:stripped_wiki_link:network-effects-create-winner-take-most-markets-because-each", + "recommendation-algorithms-amplify-or-dampen-power-laws-depending-on-collaborative-filtering-weight-because-algorithms-that-surface-popular-content-reinforce-network-cascades.md:set_created:2026-03-19", + "recommendation-algorithms-amplify-or-dampen-power-laws-depending-on-collaborative-filtering-weight-because-algorithms-that-surface-popular-content-reinforce-network-cascades.md:stripped_wiki_link:agents-that-raise-capital-via-futarchy-accelerate-their-own-", + "the-middle-of-cultural-markets-is-disappearing-because-power-law-amplification-concentrates-returns-at-the-head-and-tail-simultaneously.md:set_created:2026-03-19", + "the-middle-of-cultural-markets-is-disappearing-because-power-law-amplification-concentrates-returns-at-the-head-and-tail-simultaneously.md:stripped_wiki_link:disruptors-redefine-quality-rather-than-competing-on-the-inc" + ], + "rejections": [ + "information-cascades-create-power-law-distributions-in-culture-because-consumers-use-popularity-as-quality-signal-when-choice-is-overwhelming.md:missing_attribution_extractor", + "recommendation-algorithms-amplify-or-dampen-power-laws-depending-on-collaborative-filtering-weight-because-algorithms-that-surface-popular-content-reinforce-network-cascades.md:missing_attribution_extractor", + "the-middle-of-cultural-markets-is-disappearing-because-power-law-amplification-concentrates-returns-at-the-head-and-tail-simultaneously.md:missing_attribution_extractor" + ] + }, + "model": "anthropic/claude-sonnet-4.5", + "date": "2026-03-19" +} \ No newline at end of file diff --git a/inbox/queue/shapiro-power-laws-culture.md b/inbox/queue/shapiro-power-laws-culture.md index 8450d8f3..3681cf5a 100644 --- a/inbox/queue/shapiro-power-laws-culture.md +++ b/inbox/queue/shapiro-power-laws-culture.md @@ -7,9 +7,13 @@ date_published: "2023-03-01" date_archived: "2025-04-23" archived_by: "clay" domain: "entertainment" -status: unprocessed +status: null-result claims_extracted: - "information cascades create power law distributions in culture because consumers use popularity as a quality signal when choice is overwhelming" +processed_by: leo +processed_date: 2026-03-19 +extraction_model: "anthropic/claude-sonnet-4.5" +extraction_notes: "LLM returned 3 claims, 3 rejected by validator" --- # 4/23/25, 6:53 PM Power Laws in Culture - The Mediator by Doug Shapiro @@ -855,3 +859,17 @@ Subscribe https://archive.ph/0cYxS ## 20/20 + + +## Key Facts +- Spotify has over 80 million tracks with 100,000 new songs uploaded daily as of 2023 +- Top Gun Maverick generated over $700 million domestic box office in 2022 +- Bad Bunny had 18.5 billion Spotify streams in 2022 +- 142 million households watched Squid Game Season 1 in first 28 days per Netflix +- Patreon has 250,000+ creators and 13 million patrons +- Netflix reports 80% of watch time comes from recommendations, 20% from direct search +- Spotify reports 1/3 of new music discovery occurs through algorithmic recommendation +- Major labels (Universal, Sony, Warner) plus Merlin represented 87% of Spotify streams in 2017, declining to 77% by 2021 +- In Salganik/Dodds/Watts experiment, 14,000 subjects rated 48 unknown songs across 9 groups (1 independent, 8 social influence) +- Box office distribution slope became increasingly negative (more extreme) from 2000 to 2022 +- Netflix top 10% of originals represented 95%, 85%, and 75% of global demand in 2018, 2020, and 2022 respectively during period of massive international expansion -- 2.45.2 From ff46d2ad7132eaa6c3c70377b432b861370127f5 Mon Sep 17 00:00:00 2001 From: Teleo Agents Date: Thu, 19 Mar 2026 16:48:16 +0000 Subject: [PATCH 2/5] pipeline: archive 1 source(s) post-merge Pentagon-Agent: Epimetheus <3D35839A-7722-4740-B93D-51157F7D5E70> --- .../general/shapiro-power-laws-culture.md | 857 ++++++++++++++++++ 1 file changed, 857 insertions(+) create mode 100644 inbox/archive/general/shapiro-power-laws-culture.md diff --git a/inbox/archive/general/shapiro-power-laws-culture.md b/inbox/archive/general/shapiro-power-laws-culture.md new file mode 100644 index 00000000..c2d299d0 --- /dev/null +++ b/inbox/archive/general/shapiro-power-laws-culture.md @@ -0,0 +1,857 @@ +--- +source_type: "article" +title: "Power Laws in Culture" +author: "Doug Shapiro" +url: "https://dougshapiro.substack.com/p/power-laws-in-culture" +date_published: "2023-03-01" +date_archived: "2025-04-23" +archived_by: "clay" +domain: "entertainment" +status: processed +claims_extracted: + - "information cascades create power law distributions in culture because consumers use popularity as a quality signal when choice is overwhelming" +--- +# 4/23/25, 6:53 PM Power Laws in Culture - The Mediator by Doug Shapiro + +archive.today Saved from https://dougshapiro.substack.com/p/power-laws-in-culture + +webpage capture +All snapshots from host dougshapiro.substack.com +search +no other snapshots from this url +Webpage +Screenshot +https://archive.ph/0cYxS + +the mediator + +Subscribe +Sign in + +## Power Laws in Culture + +Why Hits Will Persist in an Infinite Content World + +DOUG SHAPIRO +MAR 16, 2023 + +[Note that this essay was originally published on Medium] + + + +Source: Hurca!/stock.adobe.com + +* Almost 20 years ago, Chris Anderson wrote The Long Tail, which accurately predicted that the Internet would fragment attention and consumption would shift into the "tail.” But Top Gun Maverick generated over $700 million at the domestic box office last year, Bad Bunny had 18.5 billion streams on Spotify last year and 142 million households reportedly watched Squid Game Season 1 in its first 28 days. Why are there still hits in a fragmenting world? + +* I recently posted an essay called Forget Peak TV, Here Comes Infinite TV. It made the case that over the next decade video will follow the path of text, photography and music and the quality distinction between “professionally-produced" content and "independent/creator/user-generated" content will increasingly blur. This will result in practically infinite quality video content. Will there still be hits then, or only personalized niches? + +* Have you ever wondered why so many blockbuster movies are about superheroes? Is Hollywood lazy or are consumers' tastes becoming dumber and more homogenized? Or neither? + +## 1/20 + +# 4/23/25, 6:53 PM Power Laws in Culture - The Mediator by Doug Shapiro + +https://archive.ph/0cYxS + +* Why does something go viral, anyway? + +* Do content recommendations push you to the most popular shows, movies and songs or are they tailored just for you? Or do they have a different agenda? + +* Will web3 really be the savior of small creators? + +* When Billie Eilish, Lil Nas X, Mr. Beast or PewDiePie emerge from obscurity, was it inevitable that their talent would be recognized or just luck? + +* Are the top rated reviews on Amazon or answers on Quora really the most helpful? + +All of these are questions about the distribution of popularity. And the same phenomenon underlies the answers: networks. + +This essay may be a little wonky, but the topic is something I've been thinking about for more than a decade. (Off and on, not continuously.) + +I explain why power law-like distributions—meaning a few massive hits and a vast number of misses—are an inherent feature of networks; describe how recommendation systems can either dampen or reinforce social signals; show some examples of the persistence of power law-like distributions in media across movies, TV, music and the creator economy; and discuss why all this matters. + +Tl;dr: + +* In an apparent contradiction, the Internet both fragments and concentrates attention. + +* The reason for the former is intuitive. More stuff, less attention per unit of stuff. The reason for the latter is not. It happens because networks are subject to powerful positive feedback loops. On a network, people's choices are influenced by others' decisions, amplifying "hits.” + +* There are two mechanisms underlying this: information cascades (when people treat others' choices as signals of quality) and reputational cascades (when people conform with the group decision). As choice has exploded on the Internet and it has become easier to both observe others' choices and share your own, these mechanisms have become more powerful. + +* Consumers also rely heavily on recommendation algorithms to make choices, intentionally and unintentionally. Depending on how they're constructed, these systems can either boost or dampen the social signals arising from the network. + +* The result is that the distribution of consumption in almost all media persistently, and in some cases increasingly, looks like a power law: a few massive hits and a very, very (very) long tail. I provide a framework for thinking about the "extremeness" of the distribution and show a few examples: box office, Netflix original series, Spotify streams and Patreon patrons. + +* There are a number of important implications for media companies. The good news is that there will likely always be big hits, even in a world of practically infinite content. The bad news is just about everything else: the lucrative middle is being hollowed out; the randomness—and therefore risk-in producing hits is climbing; the tail is become more competitive for hits; more economic rent will + +## 2/20 + +# 4/23/25, 6:53 PM Power Laws in Culture - The Mediator by Doug Shapiro + +https://archive.ph/0cYxS + +likely shift to talent; content producers are increasingly at the mercy of curators' algorithms; and paid media is being devalued. + +Thanks for reading The Mediator! Subscribe for free to receive new posts and support my work. + +## The Long Tail Was Half Right + +The idea that the Internet would cause media fragmentation is almost as old as the modern Internet itself. (Or maybe older. The line often misattributed to Andy Warhol that "in the future, everyone will be world-famous for 15 minutes” was a pre-Internet prediction of fragmentation.) In 1999, Qwest Communications produced an ad featuring a motel with “every movie ever made in every language" (Figure 1). [The Long Tail](https://www.wired.com/2004/10/tail/), published in 2004, argued that because the Internet dramatically lowered the cost to store and transport information goods, it would result in practically infinite shelf space. Faced with far more choice, consumers would shift most of their consumption to the "tail,” heralding the end of mass culture and waning importance of hits. If anything, Anderson underestimated the size of the tail because he didn't anticipate social media. The tail is not Icelandic synth pop, as it turns out, but an endless amount of user generated content. + +Figure 1. Qwest Envisioned Media Fragmentation 25 Years Ago + + + +Source: Qwest Communications print advertisement, 1999. + +That the Internet would yield more choice and, therefore, more fragmentation was intuitive then and is indisputable now. But it only tells half the story. Though it seems contradictory, the Internet both fragments and concentrates attention. This latter idea was explored by Anita Elberse in her book [Blockbusters: Hit-making, Risk-taking, and the Big Business of Entertainment](https://www.amazon.com/Blockbusters-Hit-making-Risk-taking-Business/dp/0547248912), which was in part a rebuttal to The Long Tail. But that book + +## 3/20 + +# 4/23/25, 6:53 PM Power Laws in Culture - The Mediator by Doug Shapiro + +https://archive.ph/0cYxS + +was more focused on why suppliers should pursue blockbuster strategies and less about the underlying demand-side dynamics that create hits. + +Understanding those dynamics matters. The contention that there are still hits may seem uncontroversial and certainly feels right intuitively. We know that when Beyonce or Taylor Swift releases an album or the next season of Stranger Things or Game of Thrones drops, the collective attention of popular culture, much like the eye of Sauron, will be trained on it—at least until the next thing comes along. But understanding why there are still hits provides insight into whether this will persist as the supply of content keeps growing faster than demand. + +Understanding why there are still hits provides insight into whether this will persist and the implications. + +The reason the Internet concentrates attention is that it connects everyone in a big network. And networks are subject to powerful feedback loops. Since consumers increasingly both discover and consume content through information networks, their decisions are increasingly influenced by other people's decisions. These feedback loops amplify the popularity of a small number of choices-hits. + +The net result of these opposing forces-fragmentation and concentration-is that media consumption, and culture more broadly, is persistently, and in some cases, increasingly observing power-law like distributions. That means that few TV shows, movies, songs, books, video games, journal articles, newsletters, short form videos and tweets will be wildly popular, while the vast (vast, vast, vast...) majority will be hardly consumed at all. + +## What is a Power Law? + +One of the first statistical concepts we are taught in school, right after mean, median and mode, is the "bell curve," aka the normal or Gaussian distribution. The intuition behind a normal distribution is that if you have enough random independent observations most observations will be relatively close to the average (or mean) and equally distributed on either side of it. Many independent natural phenomena approximate this distribution, especially when the extremes are bounded, like height, weight, test scores or rolling two six-sided dice. + +Figure 2. Normal and Power Law Distributions + + + +## 4/20 + +# 4/23/25, 6:53 PM Power Laws in Culture - The Mediator by Doug Shapiro + +https://archive.ph/0cYxS + +Power law distributions, by contrast, look very different. A power law simply means that the dependent variable is a “power” of the independent variable. For instance, the volume of a cube is a “power” of the length of the sides, because volume increases 3 units for each 1 unit in length. Generally, they can be expressed as: + +y = ax + +In a power law probability distribution, the exponent is negative, which results in a downward sloping curve (as illustrated crudely in Figure 2). As shown, power law distributions are characterized by a large number of very small observations and a small number of very large observations. + +There are plenty of places to explore the technical differences between a normal and power law distribution, including the excellent book [Networks, Crowds and Markets](http://www.cs.cornell.edu/home/kleinber/networks-book/), available for free here (see Chapter 18). + +For our purposes, the main point of this comparison is shown in the graph furthest to the right in Figure 2, which superimposes a power law distribution over a normal distribution: the likelihood of both extremely small and extremely large observations is much greater in the former than the latter. + +The main point: in a power law, both extremely small and extremely large observations are much more common. + +Perhaps the best way of thinking about these differences is a framework popularized by Nassim Nicholas Taleb in The Black Swan. Think of the world of normal distributions as Mediocristan-a place where everything hovers somewhere around the average and the world of power-law distributions as Extremistan-a place where seemingly extreme things happen much more often. + +## Why Do Power Laws Occur in Culture? Networks + +As mentioned above, the idea that the Internet causes media fragmentation is intuitive but the idea that it also amplifies hits is not. Let's explore why that happens. + +Power laws (or, strictly speaking, power-law like distributions) show up in a lot of places: the incidence of earthquakes, the occurrence of words in any given publication (called Zipf's Law), the population of cities, metabolic scaling among mammals and a whole lot else. + +The mechanisms behind these power laws are not always clear (there is debate whether power laws are an inherent property of complex systems). But power laws are common in networks because network phenomena tend to be dependent, meaning there are feedback loops. Each node on the network influences, and is influenced by, other nodes. + +## 5/20 + + +# 4/23/25, 6:53 PM +Power Laws in Culture - The Mediator by Doug Shapiro + +Popularity follows power-law like distributions because people's choices are subject to +feedback loops. + +This is particularly true for popularity. Power-law like distributions are everywhere in +media, as shown in this [article](https://archive.ph/o/0cYxS/https://stratechery.com/2023/power-laws-in-culture/) by Michael Tauberg. + +## Social Signals Influence Our Choices + +So, if networks tend to amplify hits because people often base their choices on what +they see other people do, the next question is: why? For two reasons: 1) it is often +rational to assume that other people's choices contain valuable information; and 2) +people care what others think of them. + +These are two distinct phenomena, what social scientists call “information cascades” +and "reputational cascades." + +* Information cascades. What do you do when you have to make a choice and have + incomplete information? It probably depends on how hard it is to determine the + quality of your options yourself (“search costs”), as well as the consequences + (including the reversibility) of making a bad choice (“opportunity costs”). Search + costs are a function both of the number of choices and the time required to + ascertain the quality of each choice. For instance, it is easy to quickly judge + quality when scrolling TikTok and hard when looking for the next multi-season + TV series. The opportunity cost of listening to the first 8 seconds of a + recommended song on Spotify is very different than getting a babysitter and going + to the movies. When search and opportunity costs are low, you may choose to + figure it out yourself. When they are high and you can see what other people have + done, it is reasonable to presume that (collectively) other people have more + information than you do and base your decisions on theirs. When many people do + this successively, it results in something called an "information cascade." This is + sometimes called cumulative advantage, preferential attachment or the “rich-get- + richer effect," whereby popular things tend to get more popular and unpopular + things stay unpopular. + +Taking signals from the network is a rational choice when confronted with high search and +opportunity costs. + +* Social conformity and reputational cascades. When you can see people's choices + and they can see yours, you may conform, consciously or subconsciously. As a + generality, we all feel pressure to conform, as was corroborated by famous social + science experiments in the 1950s-1970s, such as those conducted by Solomon + Asch. Alternatively, you may intentionally choose to follow the group's decisions + because you want to signal your allegiance and worthiness of belonging, or what is + called a reputational cascade. + +# 6/20 + +# 4/23/25, 6:53 PM +Power Laws in Culture - The Mediator by Doug Shapiro + +(There is also a third reason that people are often influenced by other's choices that +I'm overlooking: network effects. Sometimes people follow the crowd because they +benefit directly from a larger network. This may be a significant factor for fax +machines, operating systems or electric vehicles, but probably has less relevance in +culture. The direct benefits of more developers building apps for Windows or more +Tesla rapid-charging stations are clear; the network effects from a lot of people +watching your favorite TV show or listening to your favorite band are questionable +and may actually be a drawback for people who believe they have unique taste.) + +## Social Signals are Becoming More Important + +So, people are more likely to be influenced by what other people do when: 1) there are +a lot of choices; and 2) it is easy to observe what other people do. + +Over the last two decades, the conditions that lead to cascades have become more prevalent: +choice has exploded and it is far easier to observe others' actions and to be observed. + +Both of those conditions have increased dramatically in the last few decades: + +* The amount of content available has exploded, making search costs + astronomical and increasing opportunity costs. It is obvious that more choice + means higher search costs. It also means higher opportunity costs, because when + you make a choice today there are many more things you are choosing not to do. +* Owing to online networks, people are much more likely to be influenced + (directly and indirectly) by what other people choose. Many people explicitly + outsource their content curation to their friends (by relying on the Facebook + newsfeed), their hand-selected panel of “experts” (on their Twitter timeline) or + their favorite celebrities or influencers (on Instagram). But sometimes we forget + that elements of social networking are embedded in non-social networking + applications too. Go to the Apple app store, Amazon, OpenTable, or even look for + “restaurants near me" on Google Maps-in every case, you will probably be + influenced by other people's opinions. Most recommendation algorithms also rely + in part on collaborative filtering, discussed more below, which is based on the + collective choices of a group or subgroup. When you accept an algorithm's + recommendation you are often indirectly influenced by what other people choose, + whether you know it or not. + +Taken together, this means that today, people are much more likely to base their +choices on other people's decisions. This explains the paradox described at the +beginning: while the Internet fragments attention, it also causes cascades that +concentrate attention. + +## Recommendation Engines Can Help or Hurt + +Confronted with so much choice, consumers don't only depend on the organic social +signals they receive from the network, they also rely (to varying degrees, depending on +the person and type of media) on recommendation systems. Those systems may +amplify or dampen the influence of the network, depending on how they are +engineered. + +# 7/20 + +# 4/23/25, 6:53 PM +Power Laws in Culture - The Mediator by Doug Shapiro + +Recommendation algorithms are based on two primary types of models: collaborative +filtering and content models. In the former, the algorithm recommends content or +products based on what other people have chosen. In the latter, recommendations are +based on certain attributes of the content or products themselves. + +Recommendation systems can amplify or dampen social signals, depending on how +they're built. + +It is common for these algorithms to include elements of both models. For instance, in +its recommendation system Netflix incorporates all kinds of metadata associated with +each content asset (director, actors, genre, age rating, tone) and popularity (viewership, +completion rates and ratings) among cohorts it believes are similar to the customer, as +well as prior viewing behavior by the customer (device, time of day, time spent +viewing). TikTok similarly bases its algorithm on user behavior, collaborative filtering +and specific content attributes, among other things. By contrast, Pandora's +recommendation system is uncommon because it is based solely on content attributes, +not on any collaborative filtering. + +## A Simple Framework + +As mentioned, power-law like distributions are ubiquitous in media, but to varying +degrees. Synthesizing the last two sections, I'll propose a few rules of thumb for +predicting when distributions will be more, or less, extreme: + +* Higher search costs = more extreme distributions (because people need to rely + more heavily on social signals) +* Higher opportunity costs = more extreme distributions (also because people are + more likely to seek out social signals before committing) +* Recommendation systems that lean heavily toward collaborative filtering = more + extreme distributions (because the algorithm amplifies the social signals) + +## A Little Math + +How do we know a popularity distribution is a power law and how do we measure +"extreme?" + +Answering those requires a little more math. As shown above, the general +mathematical expression of a power law looks like this: + +y = ax + +In a pure power law, c is a constant, which can be thought of a scaling factor. In a +power law distribution, c is also negative, which is why the curve is downward sloping. +It can be hard to tell whether this scaling factor is constant just by looking-and +therefore whether it is really a power law. An easier way is to convert the data to a log- +log plot and determine whether the resulting relationship is linear. To see why, we +take the log of both sides of the equation above: + +# 8/20 + +# 4/23/25, 6:53 PM +Power Laws in Culture - The Mediator by Doug Shapiro +log (y) = log (a) + c log (x) + +That is a linear function, equivalent to y = b + mx. In other words, if we really have a +power law (or something power-law like), the log-log plot should look like a straight +line, where the slope is c and, the larger (or more negative) the value of c, the more +"extreme" it is. We can also test how straight it is, and therefore whether the scaling +factor is really a constant, by calculating the r². + +Figure 3. Popularity Distributions Usually Show Value as a Function of Probability (or Rank) + +The image shows two graphs. The first graph has "Value" on the x-axis and "Probability of value" on the y-axis. The graph shows a curve that starts high on the y-axis and decreases as it moves to the right on the x-axis. The second graph has "Probability of value" on the x-axis and "Value" on the y-axis. The graph shows a curve that starts high on the y-axis and decreases as it moves to the right on the x-axis. The graph is labeled with "The 'head'" and "The 'tail'". + +## A Few Examples (and Caveats) + +Below, I look at some representative time series of consumption distribution for a few +media: box office, TV series on Netflix, streams on Spotify and Patreon creators. + +(One quick note: In the power law distribution above in Figure 2, the Y-axis is +probability and X-axis is value to better compare normal and power law distributions. A +more intuitive and common way to discuss popularity distributions is to flip the axes +so that the Y-axis is the value and the X-axis is the probability, which is also a power +law (Figure 3). This shows that only a handful of observations will be extremely large +(what is colloquially called the “head”) and a vast number will be very small (the “tail”). +This is how I discuss popularity distributions below.) + +This analysis is imperfect, for a few reasons. I would like to have longer time series +than I show here (box office is great, at ~20 years, but it would be great to have 20 years +of music data too). Also, the data for Spotify and Patreon only show the distribution of +consumption at the head of the curve. Since power laws are self-similar (or "scale +invariant"), in theory the distribution at the head of the curve is representative of the +entire distribution, but if these are not pure power laws that may not be the case. + +Putting those aside, all four of these examples show persistently extreme distributions +that closely approximate power laws. + +## Box Office + +Relative to most other media, moviegoers face very few choices but extraordinarily +high opportunity costs. Not surprisingly, the relative distribution of consumption has +become even more concentrated in the top hits in recent years. Figure 4 shows the +distribution of total U.S. box office in 2000, 2010, 2019 and 2022 and the same data on a +log-log basis. As shown by the r-squared values in the log-log plots, these are close to + +# 9/20 + +# 4/23/25, 6:53 PM +Power Laws in Culture - The Mediator by Doug Shapiro +power law distributions. As also shown, over that time period the distribution has +gotten increasingly extreme (i.e., the slope on the log-log plots has gotten increasingly +negative); on a relative basis, the biggest hits are bigger than ever. + +Figure 4. Distribution of Box Office Getting More Extreme + +The image shows two graphs related to the distribution of total US box office revenue. + +The first graph, titled "DISTRIBUTION OF TOTAL US BOX OFFICE," displays the percentage of total US box office revenue against release rank for the years 2000, 2010, 2019, and 2022. The graph shows that the top-ranked movies account for a larger percentage of the total box office revenue in more recent years. + +The second graph, titled "DISTRIBUTION OF TOTAL US BOX OFFICE LOG-LOG," presents the same data on a log-log scale. This transformation helps to visualize the power-law distribution of box office revenue. The graph includes R² and Slope values for each year, indicating the goodness of fit of the power-law model. The R² values are close to 1, suggesting a strong fit, and the slopes are negative, indicating a decreasing trend. + +Source: Box Office Mojo, Author analysis. + +## Netflix TV Series + +In TV, the search and opportunity costs of finding and committing to a TV series are +pretty high, which should lead to relatively extreme distributions. But it's tough to test +shifts in popularity distributions over time for all of TV because there is no good +cross-platform (linear and streaming) measurement. And although Nielsen now +provides streaming ratings, it's only been doing so for a couple of years. + +The best data I could find was from the good people at Parrot Analytics, who provided +me a time series of global demand for Netflix original series. Parrot's demand metric + +# 10/20 + +# 4/23/25, 6:53 PM + +Power Laws in Culture - The Mediator by Doug Shapiro + +incorporates a variety of inputs (social, fan and critic ratings, piracy, wikis, blogs, etc.) +to gauge the popularity of each series and movie on each streaming service. + +The most remarkable takeaway from this data is that it remains relatively skewed and +is becoming more power-law like over time despite Netflix's big international push +over this timeframe. As noted, this is global demand and measures a period when +Netflix added about 100 million subscribers, almost all of which were international, +and its annual cash content spend increased from $13 billion to $17 billion, much of +which was local content. + +Despite its growth and increased spend internationally, as shown in Figure 5, globally +demand remains concentrated in relatively few titles. Note that in 2018, 2020 and 2022, +the top 10% of originals represented ~95%, 85% and 75% of all global demand on +Netflix, respectively. + +Figure 5. Demand for Netflix Series Has Remained Skewed Despite Big International +Expansion + +The image shows two line graphs related to the distribution of global demand among the top 250 series on Netflix. The first graph shows the distribution on a linear scale, while the second graph shows the distribution on a log-log scale. Both graphs plot data for the years 2018, 2020, and 2022. The log-log graph also includes R-squared values and slopes for each year. The graphs illustrate how demand is concentrated among a few top series, and how this concentration has changed over time. + +Note: Parrot Analytics' demand metric incorporates a variety of inputs to measure the +popularity of series and movies. Source: Parrot Analytics, Author analysis. + +Spotify Streams + +Music is an interesting case because there are factors working in both directions. On +the one hand, with so much choice (Spotify has over 80 million tracks and 100,000 new +songs uploaded every day), listeners use both social signals and recommendation +engines to discover new music. And most streaming services' recommendation + +[https://archive.ph/0cYxS](https://archive.ph/0cYxS) + +11/20 + +# 4/23/25, 6:53 PM + +Power Laws in Culture - The Mediator by Doug Shapiro + +engines rely heavily on collaborative filtering (see a description of Spotify's +recommendation engine here). This implies a relatively extreme distribution. + +On the other hand, the search costs and opportunity cost of trying a new song are very +low and easily reversed (you can easily skip to the next song). Both of those factors +support a broader dispersion of consumption. + +The result is that consumption in the head is extremely skewed toward the biggest +hits, but also that more aggregate consumption is shifting into the tail. By implication, +the "middle" is even skinnier than you would see in a pure power law. + +Figure 6 shows the distribution of consumption among all the songs that appeared in +Spotify's Global Top 200 Weekly at least once, in both 2017 and 2022 (and the same +data on a log-log basis). In both years, that was about 1,000 songs. (This is the very +head of the curve-it's the top 1,000 songs out of 80 million, or the top 0.001%.) As +illustrated by the slope on the log-log plots, the distribution is very extreme, even +more so than box office. As is also evident, the slope is not constant; it becomes more +negative as you move past the 100th most popular song. That means the biggest hits +are even bigger on a relative basis and even more consumption is occurring in the tail +than would occur in a true power law. + +Figure 6. The Head of the Spotify Curve Remains Extreme... + +The image shows two line graphs related to the distribution of top songs on Spotify. The first graph shows the percentage of total streams among songs appearing in the weekly chart of top 200 songs globally, plotted against song rank. The second graph shows the same data on a log-log scale. Both graphs plot data for the years 2017 and 2022. The log-log graph also includes R-squared values and slopes for each year. The graphs illustrate how consumption is skewed towards the top songs, and how this skewness has changed over time. + +[https://archive.ph/0cYxS](https://archive.ph/0cYxS) + +12/20 + +# 4/23/25, 6:53 PM + +Power Laws in Culture - The Mediator by Doug Shapiro + +Source: Spotify, Author analysis. + +The idea that more consumption is shifting to the tail is corroborated by aggregate +consumption data. As shown in Figure 7, based on Spotify's reporting, the three +majors (Universal, Sony and Warner Music) and Merlin (a partnership of independent +labels) represented 77% of total streams in 2021, down 10 percentage points from 2017. + +Figure 7. ...But More Consumption is Also Shifting to the Tail + +The image is a bar graph showing the combined distribution market share of annual Spotify plays for Universal Music, Sony Music, Warner Music, and Merlin (%). The graph displays data from 2017 to 2021, with the market share decreasing from 87% in 2017 to 77% in 2021. + +Source: Spotify company reports, via Music Business Worldwide. + +Patreon Creators + +Patreon provides a backend solution for creators to sell subscriptions, with more than +250,000 creators on the platform and 13 million patrons. It is also an interesting +example because consumption distribution is unaffected by recommendation +algorithms. While Patreon.com features a handful of creators on its landing page, few +consumers visit it. They primarily navigate directly to creators' Patreon pages from +wherever their work is featured, such as YouTube, Apple podcasts or their websites. + +With no amplifying effect from recommendation algorithms, it should show a slightly +less skewed distribution than some other examples. Figure 8 shows the distribution of +the top 1,000 creators at the end of both 2016 and 2022 and the log-log data. Again, this +is the head of the curve, or 0.4% of creators in 2022. As shown, the distribution tracks +almost exactly as a power law, but the slope is less extreme than the prior examples. + +Figure 8. The Creator Economy Observes Power Laws Too + +[https://archive.ph/0cYxS](https://archive.ph/0cYxS) + +13/20 + +# 4/23/25, 6:53 PM + +Power Laws in Culture - The Mediator by Doug Shapiro + +The image shows two line graphs related to the distribution of patrons to top creators on Patreon. The first graph shows the distribution on a linear scale, while the second graph shows the distribution on a log-log scale. Both graphs plot data for the years 2016 and 2022. The log-log graph also includes R-squared values and slopes for each year. The graphs illustrate how patrons are distributed among the top creators, and how this distribution has changed over time. + +Source: Graphtreon, Author analysis. + +So What? Understanding the Pervasive Implications of +Power Laws + +As my 11th grade history teacher Mr. Conroy used to say "So what?" The persistence +of these highly skewed consumption distributions has very important practical +implications for the media business and culture more broadly. + +Hits Will Persist in an Infinite Content World + +As mentioned at the top, lately I have been writing about the inevitability of Infinite +TV as the quality distinction between professional and independent/creator content +blurs. + +One of the questions I got back was: will there still be hits in such a world? + +The short answer: there will likely always be hits, if not even larger ones. As described +above, the more choice, the more consumers need to rely on social signals and +recommendation engines (which in turn rely on social signals) to manage search costs. +This is already evident in music. High production value tools have been democratized, +leading to a practically infinite amount of high production value music. But massive +hits persist. + +[https://archive.ph/0cYxS](https://archive.ph/0cYxS) + +14/20 + +# 4/23/25, 6:53 PM + +Power Laws in Culture - The Mediator by Doug Shapiro + +OK, but can we really use the word "always"? Let's go really far out. What if eventually +generative Al is able to create distinct personalized content for each individual? In a +recent post about generative AI, Sequoia posited that by 2030, movies will be +"personalized dreams” (Figure 9). + +Figure 9. Will All Content be “Personalized Dreams"? + +The image is a table that outlines the evolution of AI capabilities in content creation across different media types (text, code, images, video/3D/gaming) from pre-2020 to a projected 2030. It shows a progression from basic tasks like spam detection and auto-complete to advanced capabilities like generating final drafts better than professional writers and developers, and ultimately, personalized video games and movies by 2030. + +Source: Sequoia. + +This may not be as far fetched as it sounds, at least technologically. Let's say that by +2035 we are all wearing AR glasses, which record data about us that put Google and +Facebook to shame. They track our gaze, including the length of time we linger on +anything and the dilation of our pupils, respiration and heart rate (h/t Rony Abovitz). +They might know more about us than we know ourselves. Let's go even further. +Perhaps we'll wear devices that record brain activity as we sleep and reconstruct the +imagery from our dreams. Sound crazy? Researchers in Japan just showed that this is +already possible. + +There is no way to disprove the concept of individualized content. But just because it +might be technically possible doesn't mean it will be popular. It runs counter to two +fundamental human needs: 1) People want agency (or at least the appearance of +agency) in their choices-they don't want to be reduced to an algorithm. (Which is +why Netflix recently removed its "Surprise Me" button.) 2) More important, we are +ultimately social animals and have a need to coalesce around common experiences. As +I discussed in another recent essay, for many people, those shared experiences are +entertainment (sports, music, gaming, movies, TV shows). At a time when loneliness is +considered a public health crisis, it is hard to imagine that we would forego shared +experiences and retreat to lonely theaters of one. + +Bye, Bye Middle + +If the biggest hits are as big as ever-or bigger—and the tail is also getting bigger, +another implication is that the middle is going away. + +What's the middle? Consider the middle any content that attracted attention (and +economics) solely because it benefited from formerly scarce distribution: local +newspapers largely comprising syndicated news, TV stations with weak local +coverage, radio stations without distinctive on-air personalities, middling general +entertainment cable networks populated with second-tier reruns or inexpensive reality +programming, mid-budget me-too theatrical releases, etc. It's hard to define "the + +[https://archive.ph/0cYxS](https://archive.ph/0cYxS) + +15/20 + + +# 4/23/25, 6:53 PM Power Laws in Culture - The Mediator by Doug Shapiro + +middle" with precision, but it's safe to say that historically the middle has collectively +generated a substantial proportion of profits in every media vertical. + +The dwindling middle has generated a substantial portion of profits in every media vertical. + +## Hits Include a Big Dose of Luck + +Another important implication of this "power-lawing" is that hits are increasingly +random because of how information cascades work. To be clear, I'm not arguing that +all hits are random, but that luck is becoming more important. + +Hits are not completely random, but the role of luck is increasing. + +[Meta Comment: Link to archive.ph] +https://archive.ph/0cYxS + +More than 15 years ago, researchers Matthew Salganik, Peter Dodds and Duncan +Watts conducted an experiment to determine the effect of social influence on content +choices. They split 14,000 subjects into nine groups, one "independent group" and +eight "social influence groups." All the subjects were invited to visit a website where +they were asked to rate 48 unknown songs by unknown bands. They were able to +download the songs if they chose. In the eight social influence groups, subjects could +see how many times each song had been downloaded by prior visitors from their +group; in the independent group, they couldn't. At the end, the researchers tallied the +popularity of the songs in each group. + +The major conclusions were twofold: 1) each of the nine groups had different rankings +of the songs (while some songs tended to be more popular and some songs were +consistently less popular, other than that the rankings were quite different); and 2) the +distribution of popularity within the social influence groups was more extreme than in +the independent group. The second conclusion supports the main point of this essay, +namely that the presence of social signals will cause the distribution of popularity to +be more skewed. (And keep in mind that in this experiment the only signal was the +number of previous downloads, so the participants were only subject to information +cascades, not pressure to conform or reputational cascades. In the real world, the +social signals are a lot stronger.) + +But let's think about the implications of the first conclusion, namely that each group +produced a different popularity ranking. It implies that hits require a high degree of +luck. + +To see why this happens, try out a thought experiment (borrowed from Michael +Mauboussin). Imagine a barrel with 1,000 balls in it, each of which is numbered 1-10, +and there are 100 of each number (100 #1s, 100 #2s, etc.). Also imagine you have 10 +urns, each marked 1-10. Now randomly pick 10 balls out of the barrel and, based on +the number marked on each, put each ball in its corresponding urn. Replace the 10 +balls you removed from the barrel with new balls, but this time the distribution of new +balls will be equivalent to the distribution of balls in the urns. (If there are two balls in +urn #2 and none in #3, then two of the new balls should be marked #2 and none should + +## 16/20 + +# 4/23/25, 6:53 PM Power Laws in Culture - The Mediator by Doug Shapiro + +be marked #3.) Keep running the process, removing 10 balls from the barrel at random, +placing them in the corresponding urns, and adding new balls to the barrel based on +the distribution of balls in the urns. After you run this process for enough cycles, what +you find is that the urns with more balls are increasingly likely to have more balls +added each time. + +Or think of a real-world example: Amazon reviews. The Amazon algorithm places the +reviews with the most "helpful" votes at the top. Naturally, most people start at the top +and read just a few reviews. The first reviews written for a new book will appear at the +top of the page (for lack of many reviews). So, they are more likely to be read and +deemed helpful than subsequent reviews. This creates a positive feedback loop: they +are more likely to remain near the top of the page, making it likely that new visitors +will mark them as helpful, cementing their position at the top of the page. + +In a networked environment, hits are highly sensitive to initial conditions. + +[Meta Comment: Link to archive.ph] +https://archive.ph/0cYxS + +This phenomenon (which above I referred to as the rich-get-richer effect, cumulative +advantage or preferential attachment) shows that in a networked environment +popularity is influenced by luck and highly sensitive to initial conditions. The balls +that happen to be selected first (or the reviews that are written first) have a much +higher likelihood of dominating. Even in a hypothetical world in which all content was +of equal quality there would still be massive, random hits. Was the success of +PewDiePie or Charlie Puth inevitable? Hard to say. + +As content consumption is increasingly affected by network dynamics, this means that +hits will become more unpredictable. And just as in the financial markets, higher +volatility means higher risk, and higher risk means lower returns. + +## Hits Can (and Will) Emerge from the Tail + +A corollary of the prior point is that hits can, and will, emerge from the tail. Again, +this is already evident in music. As I wrote in Infinite TV: + +[A]lmost all of the new breakout acts of the last few years-like The Weeknd, Billie +Eilish, Lil Uzi Vert, XXXTentacion, Bad Bunny, Post Malone, Migos and many +more-emerged from the tail of self-distributed content, not from A&R reps +hanging around at 2AM for the last act. + +Writing compelling fiction, composing a catchy pop song, conceiving innovative +gameplay or writing a great screenplay are extraordinarily rare talents. It is reasonable +to think that many of the people capable of doing these things, with persistence and +luck, are able to succeed through the traditional channels of content production and +win the support of the small handful of people who control resources at places like +HarperCollins, Republic Records, Blizzard or Universal Pictures. But how many +creative "lost Einsteins" are there who have fell through the cracks? Thousands? Tens +of Thousands? Hundreds of thousands? + +Just has occurred with the music labels, every traditional producer of any type of +content should be prepared to both discover talent that emerges from the tail and + +## 17/20 + +# 4/23/25, 6:53 PM Power Laws in Culture - The Mediator by Doug Shapiro + +compete with it. + +## There's a Reason Every Movie Star Wears Tights + +If it sometimes feels like every movie is a prequel or sequel or about superheroes (or +both) and every new TV show is a spinoff or reboot, that's because a disproportionate +percentage of them are (as discussed in this article by Adam Mastroianni). + +[Meta Comment: Link to article] +this article + +The reasons often cited for this include entertainment companies' crass +commercialism, the death of creativity and the dumbing-down of the American +consumer, among others. But looking at this through the lens of the network dynamics +described in this essay suggests several other reinforcing reasons. Established IP +reduces risk because it: + +* Lowers consumer search costs. As discussed above, consumers are overwhelmed + by choice and the resulting high search costs. Well-known brands, talent and + franchises reduce those costs, making consumers less reliant on network signals. +* Benefits from a pre-existing community. As also discussed, consumers + sometimes choose content because of a desire to join a community or enhance + their standing within it. Established IP has established communities, increasing + the community's influence. + +Whether this is good or bad is a different question. There is a risk that major media +companies lean too heavily on established IP and all the innovative ideas instead +emerge from the tail. But there is a clear logic behind it. + +## Rents Will Likely Shift Even More Toward Top Talent + +The details of how talent is compensated in creative businesses can be extraordinarily +complicated and opaque. If you abstract it out, however, ultimately talent +compensation is a function of the underlying economic structure of the industries in +which they operate. + +At a time when there is both more transparency of performance data and greater +competition for superstars, a more extreme distribution of consumption will likely +shift even more bargaining power to the top talent. + +## No One is Policing the Algorithm + +Algorithms clearly influence the distribution of consumption and they will become +increasingly important. According to Spotify, 1/3 of new music discovery occurs +through algorithmic recommendation. Netflix says that 80% of watch time comes from +its recommendations and 20% from direct search (but it also concedes that "users tend +to come to the service with a specific show, movie or genre in mind"). All things equal, +the more choice, the more consumers will seek help in choosing, whether from the +organic social signals that emerge from the network or recommendation systems. + +Platforms have a strong incentive to surface the best recommendations. More usage +increases consumer affinity, improves retention and, for ad supported platforms, +increases revenue. But, at least on the margin, they may have other incentives. Spotify +and Netflix both have an incentive to reduce their reliance on their largest suppliers. +Both Spotify and TikTok disclose that “commercial considerations” influence their +recommendations. Not much can or will likely be done about this, but the opacity and + +## 18/20 + +# 4/23/25, 6:53 PM Power Laws in Culture - The Mediator by Doug Shapiro + +importance of algorithms will become an increasingly important competitive +advantage for content aggregators over time. + +## The Creator Economy and Web3 Live in Extremistan Too + +Much has been written (including by me) about the rise of the creator economy and +platforms and tools that enable creators to connect directly with—and generate +revenue from-fans (not just Patreon, but Substack, OnlyFans, Cameo and many +others). Web3 promises an even more decisive step in that direction. Since web3 +applications are decentralized, data is not mediated by centralized servers and creators +retain ownership of their product. For many people, the greatest promise of web3 is to +redistribute power and value from centralized institutions to creators and users. + +While both the evolution of the creator economy and web3 should enable more +creators to make a living wage, redistribution should not be confused with equal +distribution, something I also discussed here. As shown in the popularity distributions +for Patreon creators above, as long as there are network dynamics, there will be power- +law like popularity distributions. + +## Earned Media is Increasingly Important + +Back to Salganik, Dodds and Watts for a moment. As mentioned, some of the subjects +were placed in an independent group that received no social signals at all. The +researchers used this group's popularity ranking of songs as a proxy for “quality." What +they found among the other groups was that the songs considered best by the +independent group rarely did poorly and the songs considered the worst rarely did +very well, but anything else could happen. + +Quality matters in popularity. Complete crap will fail. But, above some threshold of quality, +popularity is highly reliant on network dynamics. + +The implication is that, as any marketer would tell you, marketing matters. Quality +will not necessarily naturally rise to the top. The question is how to market. + +Marketers draw a distinction between paid, earned and owned media. Paid is +traditional advertising: TV, outdoor, print, radio, retail media, display, search and +social. Earned is PR and word-of-mouth, increasingly through influencers. And owned +is the brand's own marketing channels, such as its branded content, website, retail +outlets, catalogs, etc. Media companies tend to rely very heavily on paid media-think +of massive advertising campaigns to launch a new show or movie. As more content +discovery occurs through the network itself, the value of paid media is increasingly +diluted. It also becomes more important for marketers to understand what signals are +emerging organically and how to use both paid and earned media to amplify or +counter those signals. + +## We're Not in Kansas Anymore + +Almost 30 years since the IPO of Netscape, the media industry is still coming to grips +with the implications of the Internet. The reality that it fragments attention is +intuitive. The reasons why it also amplifies hits are less well understood. + +## 19/20 + +# 4/23/25, 6:53 PM Power Laws in Culture - The Mediator by Doug Shapiro + +For media companies, the implications of operating in a networked world are a mixed +bag, at best. The good news is that hits still matter and likely always will. The bad +news is just about everything else: the lucrative middle is being hollowed out; risk is +climbing; the tail is become more competitive for hits; bargaining power is shifting to +the top talent; content producers are increasingly at the mercy of curators' algorithms; +and paid media is being devalued. As consumers grapple with a growing tsunami of +options, these dynamics will become more pronounced. None of this will get easier. + +[Meta Comment: Social Media Icons] +D + +Previous + +Comments + +Write a comment... + +Share + +Next → + +Top +New Community + +Q + +No posts + +Ready for more? + +Type your email... +Subscribe + +[Meta Comment: Link to archive.ph] +https://archive.ph/0cYxS + +## 20/20 -- 2.45.2 From 088ea1d42f54a79fa357484395cc2b9c47f6b17e Mon Sep 17 00:00:00 2001 From: Teleo Agents Date: Thu, 19 Mar 2026 16:47:55 +0000 Subject: [PATCH 3/5] extract: shapiro-relentless-creator-economy Pentagon-Agent: Epimetheus <3D35839A-7722-4740-B93D-51157F7D5E70> --- .../shapiro-relentless-creator-economy.json | 43 +++++++++++++++++++ .../shapiro-relentless-creator-economy.md | 23 +++++++++- 2 files changed, 65 insertions(+), 1 deletion(-) create mode 100644 inbox/queue/.extraction-debug/shapiro-relentless-creator-economy.json diff --git a/inbox/queue/.extraction-debug/shapiro-relentless-creator-economy.json b/inbox/queue/.extraction-debug/shapiro-relentless-creator-economy.json new file mode 100644 index 00000000..8aa10772 --- /dev/null +++ b/inbox/queue/.extraction-debug/shapiro-relentless-creator-economy.json @@ -0,0 +1,43 @@ +{ + "rejected_claims": [ + { + "filename": "creator-and-corporate-media-are-zero-sum-because-total-time-is-stagnant.md", + "issues": [ + "missing_attribution_extractor" + ] + }, + { + "filename": "genai-reduces-creative-decisions-not-just-execution-cost-which-is-categorically-different-from-prior-tools.md", + "issues": [ + "missing_attribution_extractor" + ] + }, + { + "filename": "genai-as-general-purpose-technology-advances-faster-than-domain-specific-tools-because-breakthroughs-compound-across-modalities.md", + "issues": [ + "missing_attribution_extractor" + ] + } + ], + "validation_stats": { + "total": 3, + "kept": 0, + "fixed": 6, + "rejected": 3, + "fixes_applied": [ + "creator-and-corporate-media-are-zero-sum-because-total-time-is-stagnant.md:set_created:2026-03-19", + "creator-and-corporate-media-are-zero-sum-because-total-time-is-stagnant.md:stripped_wiki_link:two-phase-disruption-where-distribution-moats-fall-first-and", + "genai-reduces-creative-decisions-not-just-execution-cost-which-is-categorically-different-from-prior-tools.md:set_created:2026-03-19", + "genai-reduces-creative-decisions-not-just-execution-cost-which-is-categorically-different-from-prior-tools.md:stripped_wiki_link:metis-is-practical-knowledge-that-can-only-be-acquired-throu", + "genai-as-general-purpose-technology-advances-faster-than-domain-specific-tools-because-breakthroughs-compound-across-modalities.md:set_created:2026-03-19", + "genai-as-general-purpose-technology-advances-faster-than-domain-specific-tools-because-breakthroughs-compound-across-modalities.md:stripped_wiki_link:technology-advances-exponentially-but-coordination-mechanism" + ], + "rejections": [ + "creator-and-corporate-media-are-zero-sum-because-total-time-is-stagnant.md:missing_attribution_extractor", + "genai-reduces-creative-decisions-not-just-execution-cost-which-is-categorically-different-from-prior-tools.md:missing_attribution_extractor", + "genai-as-general-purpose-technology-advances-faster-than-domain-specific-tools-because-breakthroughs-compound-across-modalities.md:missing_attribution_extractor" + ] + }, + "model": "anthropic/claude-sonnet-4.5", + "date": "2026-03-19" +} \ No newline at end of file diff --git a/inbox/queue/shapiro-relentless-creator-economy.md b/inbox/queue/shapiro-relentless-creator-economy.md index 824245ce..caf44595 100644 --- a/inbox/queue/shapiro-relentless-creator-economy.md +++ b/inbox/queue/shapiro-relentless-creator-economy.md @@ -7,9 +7,13 @@ date_published: "2023-06-01" date_archived: "2025-04-23" archived_by: "clay" domain: "entertainment" -status: unprocessed +status: null-result claims_extracted: - "creator and corporate media economies are zero-sum because total media time is stagnant and every marginal hour shifts between them" +processed_by: leo +processed_date: 2026-03-19 +extraction_model: "anthropic/claude-sonnet-4.5" +extraction_notes: "LLM returned 3 claims, 3 rejected by validator" --- # 4/23/25, 6:54 PM The Relentless, Inevitable March of the Creator Economy @@ -852,3 +856,20 @@ JAN 7 DOUG SHAPIRO https://archive.ph/wTgnR # 21/22 + + +## Key Facts +- Global M&E market grew at ~5% annually 2019-2023 (PwC/Omdia estimates) +- Creator media economy grew ~25% annually 2019-2023 (Shapiro estimate) +- Corporate media grew ~3% annually 2019-2023 (Shapiro estimate) +- Creator media economy was ~$250B globally in 2023 (~15% of total M&E) +- Creator media accounted for almost half of total M&E revenue growth 2019-2023 +- Social video represents ~1/4 of all video consumption time in U.S. (Maverix/Nielsen data) +- ~25% of Spotify streams are from artists not represented by majors or Merlin +- YouTube has 114 million channels and users upload ~250 million hours annually +- Spotify has 10 million+ total artists uploading ~37 million tracks per year (vs 225K professional/professionally-aspiring) +- Steam has 100,000 games vs 3,000 supported on Xbox +- Kodak estimated 80 billion photos taken in 2000; current estimates close to 2 trillion for 2023 (25x increase) +- 83% of TikTok users have posted a video (2021 study) vs ~4% of YouTube users who create +- Over 4 million podcasts exist today vs only a few thousand in 2004 +- Major labels and Merlin accounted for 74% of Spotify streams in 2023 -- 2.45.2 From 6d3218abe84202b9ad3f25131e5a540eabb011cd Mon Sep 17 00:00:00 2001 From: Teleo Agents Date: Thu, 19 Mar 2026 16:49:30 +0000 Subject: [PATCH 4/5] pipeline: archive 1 source(s) post-merge Pentagon-Agent: Epimetheus <3D35839A-7722-4740-B93D-51157F7D5E70> --- .../shapiro-relentless-creator-economy.md | 854 ++++++++++++++++++ 1 file changed, 854 insertions(+) create mode 100644 inbox/archive/general/shapiro-relentless-creator-economy.md diff --git a/inbox/archive/general/shapiro-relentless-creator-economy.md b/inbox/archive/general/shapiro-relentless-creator-economy.md new file mode 100644 index 00000000..fb06ae9f --- /dev/null +++ b/inbox/archive/general/shapiro-relentless-creator-economy.md @@ -0,0 +1,854 @@ +--- +source_type: "article" +title: "The Relentless Inevitable March of the Creator Economy" +author: "Doug Shapiro" +url: "https://dougshapiro.substack.com/p/the-relentless-inevitable-march" +date_published: "2023-06-01" +date_archived: "2025-04-23" +archived_by: "clay" +domain: "entertainment" +status: processed +claims_extracted: + - "creator and corporate media economies are zero-sum because total media time is stagnant and every marginal hour shifts between them" +--- +# 4/23/25, 6:54 PM The Relentless, Inevitable March of the Creator Economy + +Thanks for reading The Mediator! Subscribe for +free to receive new posts and support my work. + +This post is sponsored by WSC Sports. + +The NBA, Top Rank, Euroleague and more are already working with the WSC Sports' Creators +Program to expand reach to fans and monetize archival and near live sports content. + +Fans are following influencers, so give influencers official tools to provide new perspectives and +storylines to their audiences. The Creators Program exposes your content to new potential fans +and generates additional revenues. + +WSC Sports' Creators Program provides a turnkey solution for rights holders by offering: + +* Full rights holder control over content +* Options for creator access and types of accessible content +* Performance metrics and valuable data + +Reach out to WSC Sports to learn more. + +To contact me about sponsorship opportunities for The Mediator, reach me here. + +## Defining the Creator (Media) Economy + +Let's establish some definitions. + +There isn't a consensus definition of "creator." Sometimes creators are considered +synonymous with influencers. That's relatively narrow, because it confines the creator +economy mostly to Instagram, TikTok and YouTube. Sometimes creators are +considered those who distribute content online strictly to commercialize it. On a +recent episode of The Colin and Samir Show, Samir drew the distinction between a +creator and a creative: + +> ...a creator is someone with a distribution mind. They're thinking about what do I +make that's going to reach the most amount of people? They're an independent +media company....And they're trying to solve how they can get their content seen at +a large scale on platforms...A creative is working on the craft, right? They're +working on the skill set and they typically get hired to direct stuff or support other +people in making their thing. + +Figure 1. The Corporate Media Economy + +[https://archive.ph/wTgnR](https://archive.ph/wTgnR) + +3/22 + +# 4/23/25, 6:54 PM The Relentless, Inevitable March of the Creator Economy + +The image is a diagram illustrating the corporate media economy. It shows a linear process starting with "IDEATION" and ending with "CONSUMPTION". The process includes steps such as "PRODUCTION", "MARKETING", "DISTRIBUTION", and "MONETIZATION". On the right side of the diagram, there are examples of creative roles (e.g., Writer, Musician, Director), producers/publishers (e.g., Music Label, Newspaper, TV and Film Studio), aggregators/distributors (e.g., Retailer, Streaming Service, Theater), and traditional intermediaries (e.g., Sony, Netflix, Disney+). The diagram visually represents the flow of content creation and distribution in the corporate media landscape. + +IDEATION +PRODUCTION +MARKETING +DISTRIBUTION +MONETIZATION +CREATIVE +Writer | Musician +Director | Actor | Producer +Makeup Artist | Designer +DP | Journalist +Developer | Photographer +Editor | Animator +VFX Artist +PRODUCER/ +PUBLISHER +Music Label +Newspaper +Magazine +Videogame Publisher +TV and Film Studio +CONSUMPTION +MONETIZATION +DISTRIBUTION +100000 +MARKETING +IDEATION +Writer | Musician +Director | Actor | Producer +Makeup Artist | Designer +DP | Journalist +Developer | Photographer +PRODUCTION +Editor | Animator +VFX Artist +CREATIVE +AGGREGATOR/ +DISTRIBUTOR +Retailer (electronic or +physical) | Streaming +Service | Theater +TV/Radio Station | Cable +Network | Cable +Systems/Satellite/Telco +TRADITIONAL INTERMEDIARIES +SONY +The WALT Disney Studios +ACTIVISION A NETFLIX tv+ +CONDÉ NAST +Disney+ +NBC UNIVERSAL The New York Times +VALVE +Paramount +UNIVERSAL +WARNER MUSIC GROUP +prime video +Discovery Turner +spectrum +iHeart Xfinity +RADIO +Nexstar +amazon +tv CINEMARK +Walmart +GameStop +CONSUMER + +Source: Author. + +Since I focus on the business of media, to me the most interesting distinction is +between traditional media, or what we could call corporate media, and creator media. +Let's define two, mutually-exclusive, economies: + +* The corporate media economy is the ecosystem of traditional content creation, +distribution and monetization, which usually entails institutional ownership, +centralized decision making, portfolio-level risk management and several intermediaries +between creative 1 and consumer who provide financing, marketing and distribution +(Figure 1). As shown in Figure 2, most of the household names in the media and +entertainment business are intermediaries. +* The creator media economy, as I'm defining it here, encompasses all other media +monetization. It is the ecosystem of content creation activities in +which independent creators create content on a self-directed basis, they have a direct +relationship with consumers, and this content is monetized. The passive voice in the +last clause signifies that the content is monetized by someone, even if not by the +creators themselves. (So, under this definition, everyone who posts anything that +generates revenue is a creator, even if it is Meta or X/Twitter who monetizes it, +not them.) (Figure 3.) A gray area is small independent teams, of, say, 50 people or +fewer. I put these in the creator category. Mr. Beast runs a full-fledged production +company, with multi-million dollar budgets, but for these purposes he is a creator. +2 + +Figure 3. The Creator Media Economy + +[https://archive.ph/wTgnR](https://archive.ph/wTgnR) + +4/22 + +# 4/23/25, 6:54 PM The Relentless, Inevitable March of the Creator Economy + +The image is a diagram illustrating the creator media economy. It shows a linear process starting with "IDEATION" and ending with "CONSUMPTION". The process includes steps such as "PRODUCTION", "MARKETING", "DISTRIBUTION", and "MONETIZATION". On the left side of the diagram, there are examples of creator roles (e.g., Blogger, Singer, Musician, Comedian), and on the right side, there are enabling tools/platforms (e.g., Unity, Ableton, Instagram, YouTube, Spotify). The diagram visually represents the flow of content creation and distribution in the creator media landscape. + +IDEATION +PRODUCTION +The Relentless, Inevitable March of the Creator Economy +ENABLING TOOLS/PLATFORMS +Unity UNREAL +MARKETING +DISTRIBUTION +CREATOR +Blogger | Singer +Musician | Comedian +Actor | Game Developer +Influencer | Journalist +Photographer +Podcaster | Digital Artist +Video Creator +Streamer | Animator +IIII Ableton +Logic Pro +Instagram Tik Tok +DISCORD +► YouTube Spotify substack +MONETIZATION +CONSUMPTION +STEAM +CONSUMER +SOUNDCLOUD +PATREON + +Source: Author. + +The Relationship Between Corporate Media and Creator +Media is Zero Sum + +As I have written about before (like here and here), the overall media and +entertainment (M&E) market is not growing much globally, slightly less than the rate +of inflation (Figure 4). + +Figure 4. Globally, Media Isn't Growing on a Real Basis + +Value of the Global Entertainment and Media Market, +Nominal and Real + +$, in Trillions +$2.5 +$2.0 +$1.5 +$1.0 +$0.5 +$0.0 +2019 +2020 +2021 +2022 +2023 +2024 +2025 +2026 +2027 +2028 +Nominal +Real + +Note: Includes PwC estimates for “Consumer” and “Advertising,” but not “Connectivity." +Sources: PwC and Omdia, IMF, Author analysis. + +The reason is that time spent with media has stagnated in recent years. It grew with +the advent of mobile starting in 2008 and then had a COVID bump in 2020, but has +been flat or declined since (Figure 5). Since M&E revenue is derived by monetizing +consumer time and engagement, it is tough for the overall market to grow faster than +inflation if time spent is not growing. + +[https://archive.ph/wTgnR](https://archive.ph/wTgnR) + +5/22 + + +# 4/23/25, 6:54 PM +The Relentless, Inevitable March of the Creator Economy + +Since M&E revenue is derived by monetizing consumer time and engagement, it is tough for +the overall market to grow if time spent is not. + +Figure 5. Time Spent is Not Growing Either + +The image is a line graph showing the average daily time spent with media by U.S. adults from 2008 to 2022. The y-axis represents time in hours and minutes, ranging from 0:00 to 14:24. The x-axis represents the years from 2008 to 2022. The graph includes several categories of media: Print, Radio, TV, PC, Mobile, and Other Connected Devices. The "Other Connected Devices" category shows the most significant growth over the period, reaching 13:11 in 2022. The other categories show varying degrees of change, with some declining and others remaining relatively stable. + +Source: eMarketer, April 2022. + +As mentioned, my intention is that these two economies are mutually exclusive and +cumulatively exhaustive (or MECE, as they say in consulting land). Every dollar of end- +market M&E revenue is either one or the other. As there is only one pool of consumer +time, the relationship between the corporate and creator media economies is largely +zero sum. The growth in the latter mostly comes at the expense of the former. + +Creators Generate Revenue on a Lot of Platforms + +Under my definition above, creators' work is monetized (there's the passive voice +again) on a wide variety of outlets and platforms. These include: + +* Social Networking (Meta, YouTube, Douyin, TikTok, Kuashiou, Snap, Pinterest, X, + Bilibili, Weibo, VK, etc.) +* Patronage/Community (OnlyFans, Patreon, Discord, etc.) +* Gaming (Mobile Gaming, Steam, Epic, Roblox) +* Livestreaming (Twitch, Bigo Live, Huya, DouYu) +* Music (Spotify, Apple Music, Soundcloud, Bandcamp, etc.) +* Podcasting +* Influencer Marketing +* Writing (Substack, Medium, Ghost, Beehiiv, etc.) + +The proportion of total revenue on these outlets that is attributable to creators can +range from very little to all of it. + +[https://archive.ph/wTgnR](https://archive.ph/wTgnR) + +6/22 + +# 4/23/25, 6:54 PM +The Relentless, Inevitable March of the Creator Economy + +For instance, in gaming, a relatively small proportion of mobile game (iOS and Google +Play) revenue is attributable to independent developers (I estimate ~5-10%), slightly +more for Epic, slightly more for Steam, and, for Roblox, almost all revenue is +attributable to independent developers (other than the few games that Roblox creates +itself). In music, Spotify reported that the major labels and Merlin accounted for 74% +of streams last year, so we can attribute ~25% of revenue to independent and individual +creators, but almost all of the revenue on Bandcamp likely comes from creators. On +social networking and patronage platforms like Patreon, the majority or virtually all of +the revenue is attributable to creators. Likewise, influencer marketing represents the +sponsorship fees paid by brands directly to influencers and so is also 100% attributable +to creators. This continuum of creator attribution can be seen in Figure 6. + +Figure 6. The Proportion of Revenue Attributable to Creators Varies Widely + +The image is a bar graph showing the proportion of platform revenue attributable to creators for various platforms. The y-axis represents the percentage, ranging from 0% to 100%. The x-axis lists different platforms, including Mobile Gaming (Google Play & iOS), Steam, Spotify, Discord, Pinterest, Podcasts, Epic Games, Apple Music, Meta Platforms (Facebook & Instagram), X/Twitter, YouTube Premium, Weibo, YouTube (Advertising), Snap, VK (VKontakte), Huya, DouYu, Tik Tok, Douyin, Kuaishou, Bilibili, Bigo Live, SoundCloud, Twitch, Bandcamp, Roblox, Influencer Marketing, OnlyFans, Patreon, Substack, and Medium. The bars vary in height, indicating the different proportions of revenue attributable to creators for each platform. For example, Influencer Marketing, OnlyFans, Patreon, and Roblox have bars reaching 100%, while Mobile Gaming (Google Play & iOS) has a very low percentage. + +Source: Company reports, Author estimates. + +How Big is It? + +In Figure 7, I show my bottoms-up estimate of the aggregate end-market revenue of +the creator media economy, i.e., all advertising, subscription and transactional revenue +attributable to creator content, globally. I derived this by applying the proportions in +Figure 6 to the reported or estimated revenue for each outlet. As shown, I calculate +that total creator media economy revenue was a little shy of $250 billion last year. + +Figure 7. The Creator Media Economy Approached $250 Billion Globally Last Year + +[https://archive.ph/wTgnR](https://archive.ph/wTgnR) + +7/22 + +# 4/23/25, 6:54 PM +The Relentless, Inevitable March of the Creator Economy + +The image is a stacked bar graph showing the creator media economy revenue from 2019 to 2023. The y-axis represents the revenue in billions of dollars. The x-axis represents the years from 2019 to 2023. The graph is divided into several categories: Social Networking (Meta, YT (Ad and Premium), Douyin, Tik Tok, Kuashiou, Snap, Pinterest, X, Bilibili, Weibo, VK, etc.), Influencer Marketing, Patronage/Community (OnlyFans, Patreon, Discord, etc.), Gaming (Mobile Gaming, Steam, Epic, Roblox), Livestreaming (Twitch, Bigo Live, Huya, DouYu), Music (Spotify, Apple Music, Soundcloud, Bandcamp, etc.), Podcasting, Writing (Substack, Medium, Ghost, Beehiv, etc.), and Other. The total revenue increases over the years, with Social Networking being the largest contributor. + +estimates that the total M&E has grown at 5% annually over the past four years, I +estimate that the creator media economy has grown ~25% per year and corporate +media has grown at 3%. So, although creator media is a relatively small portion of the +total M&E market, it has accounted for almost half the growth. + +The creator media economy has accounted for about half of total M&E revenue growth over +the last four years. + +Figure 8. The Creator Media Economy is ~15% of Global M&E and Half its Growth + +The image is a combination of a bar graph and a line graph showing the global corporate media vs. creator media revenue from 2019 to 2023. The left y-axis represents the revenue in billions of dollars, and the right y-axis represents the percentage. The x-axis represents the years from 2019 to 2023. The bar graph shows the revenue for the Creator Media Economy and the Corporate Media Economy. The line graph shows the Creator Economy % of Total Media Economy. The CAGR (Compound Annual Growth Rate) for the Creator Media Economy is highlighted as 26%, while the CAGR for the Corporate Media Economy is 3%. The Creator Economy % of Total Media Economy is around 15% in 2023. + +Note: Global M&E includes PwC estimates for “Consumer” and “Advertising,” but not +"Connectivity." Source: Company reports, PwC and Omdia, eMarketer, Statista, Sacra, Wall +Street Zen, Fast Company, Video Game Insights, MoffettNathanson, Influencer Marketing +Hub, CB Insights, Music Business Worldwide, Author estimates. + +The Creator/Independent Media Economy Will Inevitably +Keep Taking Share + +[https://archive.ph/wTgnR](https://archive.ph/wTgnR) + +8/22 + +# 4/23/25, 6:54 PM +The Relentless, Inevitable March of the Creator Economy + +A simple math exercise shows how much larger and relatively more important the +creator media economy will be by the end of the decade, if it keeps growing anywhere +close to its recent pace. 3 Presuming that the total M&E market grows in line with the +PwC and Omdia estimate of ~4% through the end of the decade, then: + +* If the creator media economy grows at 10% annually, by 2030 it will be $460 billion + and 20% of the M&E market; +* If it grows at 15% growth annually it would reach $630 billion and exceed 25% of + the market; +* And, at 20% annual growth it would approach $850 billion and exceed 35% of the + market. + +Figure 9 shows the mid case, 15% annual growth. + +Figure 9. The Creator Media Economy Could Easily Reach ~25% of Global M&E by the End +of the Decade + +The image is a combination of a bar graph and a line graph showing the global corporate media vs. creator media revenue from 2019 to 2030 (estimated). The left y-axis represents the revenue in billions of dollars, and the right y-axis represents the percentage. The x-axis represents the years from 2019 to 2030. The bar graph shows the revenue for the Creator Media Economy and the Corporate Media Economy. The line graph shows the Creator Economy % of Total Media Economy. The CAGR (Compound Annual Growth Rate) for the period 2023-2030 is 4%. The Creator Economy % of Total Media Economy is estimated to reach around 25% by 2030. + +Note: Global M&E includes PwC estimates for “Consumer” and “Advertising,” but not +“Connectivity.” Source: Company reports, PwC and Omdia, eMarketer, Statista, Sacra, Wall +Street Zen, Fast Company, Video Game Insights, MoffettNathanson, Influencer Marketing +Hub, CB Insights, Music Business Worldwide, Author estimates. + +Since no one likes wishy washy, let's go with a point estimate: I forecast that the +creator media economy will more than double by the end of the decade, exceeding +$600 billion and 25% of the entire M&E market. + +Powerful technological, cultural and demographic trends are tailwinds for the creator +economy. + +But there are a whole host of reasons-powerful technological, cultural, demographic +and economic trends-why it could grow even faster than that. Let's walk through +them. + +[https://archive.ph/wTgnR](https://archive.ph/wTgnR) + +9/22 + +# 4/23/25, 6:54 PM +The Relentless, Inevitable March of the Creator Economy + +1. The Volume of Creator Content Will Keep Growing Fast +(Even Without GenAl) + +There is already a vast amount of creator/independent content. + +A few examples to make the point are shown in Figure 10. Consider: 20,000 times as +much video is uploaded to YouTube each year as is produced by Hollywood (in other +words, the equivalent of Hollywood's annual output is uploaded every ~30 minutes, +24/7); 98% of artists on Spotify are hobbyists and they upload ~100,000 tracks per day; +there are more than 30x as many games on Steam as are supported by Xbox (and it is +set to add 17,000 new games this year). + +Still, this gulf between the amount of creator content and “corporate” content will +undoubtedly widen. + +Figure 10. Some Examples of the Relative Scale of Creator Content + +| | Traditional +The image is a table describing the relative scale of creator content. The table has two columns, "Traditional" and "New," and three rows, "TV and Film," "Music," and "Games." The "Traditional" column provides information about the traditional media industry, such as the number of hours of TV and film produced by Hollywood annually. The "New" column provides information about the amount of content uploaded by users to platforms like YouTube, Spotify, and Steam. The table highlights the significant difference in scale between traditional media and creator content. + +* TV and Film: Hollywood produces about 15,000 hours of TV and film annually in the U.S. Users upload ~250 million hours of video to YouTube annually, across 114 million channels. +* Music: There are 225,000 professional and "professionally-aspiring" musicians on Spotify, uploading about 5 million tracks per year. There are 10 million+ total artists on Spotify, uploading roughly 37 million tracks per year. +* Games: There are 3,000 games supported on Xbox. There are 100,000 games on Steam and ~500,000 games on the iOS app store. + +Source: YouTube upfront May 2019, Tim Queen, Spotify 4Q21 earnings release, Spotify +"Loud&Clear" Top Takeaways 2023, Wikipedia, Steam, Business of Apps, Author estimates. + +Part of the reason is that the more accessible it is to create, the more people create. Without +probing the psychological or evolutionary roots of it, it is clear that humans have an +innate desire to create. Closer to the bottom of Maslow's hierarchy than the top, +creativity emerges spontaneously in children (until it is wrung out of most of us by +society, criticism or something else); throughout history, every known culture has +produced art, music and stories; and people create art in the most extreme hardship, in +prison, during war, and in dire poverty. + +As evidence of this innate need, people create more when creation is more accessible. + +The empirical evidence shows that people make more when creation is more +accessible. Some examples: + +* While Kodak estimated that 80 billion photos were taken in 2000, current + estimates are close to 2 trillion for this year, a more than 20-fold increase— + obviously driven by the current constant availability of cameras. +* YouTube has 2.7 billion MAUs and an estimated 114 million channels. Even if + each of these channels is run by a discrete user and all of these channels are active + (neither of which is true), that means about 4% of users also create. By contrast, + TikTok makes creation much easier. It has a camera function in the app and offers + +[https://archive.ph/wTgnR](https://archive.ph/wTgnR) + +10/22 + + +# 4/23/25, 6:54 PM +The Relentless, Inevitable March of the Creator Economy + +* in-app editing tools, filters, music libraries, text overlays, stitches, etc. According to a 2021 study by TikTok, 83% of users have posted a video. +* In 2004, there were only a few thousand podcasts. Today, thanks to tools like Riverside FM, Zencastr, cheap webcams, high-quality mics and the like, there are currently over 4 million. + +Through the natural progression of software development and the move toward no- code/low-code, creation tools will undoubtedly keep getting more user friendly: better and easier video editing tools; music sample and beat marketplaces and collaboration tools; no-code/low-code game development on UGC gaming platforms, etc. But the most significant innovation is likely to be generative AI (GenAI). + +## 2. GenAl Will Trigger a Tsunami of Creator Content + +If I were to distill the last couple of years of my writing into one sentence, it would be this: the last two decades in media were defined by the disruption of content distribution, facilitated by the internet, the next decade will be defined by the disruption of content creation, enabled by GenAI. + +It not controversial to write that GenAI will result in a lot more content, but let's tease apart the two key reasons. + +Prior innovations in content creation technology have mostly reduced the cost for humans to execute creative decisions. GenAI reduces the number of creative decisions. + +### GenAl Automates Creative Decisions + +Prior innovations in content creation technology have mostly made it easier and cheaper for humans to execute creative decisions. But they have not materially reduced the number of creative decisions. GenAI, in contrast, can automate creative decisions. Humans can decide what proportion of creative decisions they delegate to AI, anywhere from almost all of them to relatively few. (Whether the output in the former case will be any good is a different question.) But even when there is substantial human direction and oversight, it can automate a lot of creative decisions, dramatically speeding the creative process. (See GenAI is Foremost a Creative Tool for a more detailed discussion.) + +### As a General Purpose Technology, GenAl is Advancing Incredibly Fast + +GenAI is clearly moving at a blistering pace. One of the key reasons this is happening is because it is a general purpose technology (GPT). + +Most of the innovations in content creation over the last 5-10 years have been medium or domain-specific: ubiquitous cameras on mobile phones; cheaper in-home production equipment, like microphones; digital audio workstation (DAWS) software; free gaming engines for small developers from Epic and Unity; inexpensive and easy-to-use photo and video editing tools, etc. Advances in one domain didn't necessarily benefit others. DAWs didn't help anyone make videos faster. + +[https://archive.ph/wTgnR](https://archive.ph/wTgnR) + +11/22 + +# 4/23/25, 6:54 PM +The Relentless, Inevitable March of the Creator Economy + +Just as bits were a new atomic unit for the distribution of information goods, tokens are a new atomic unit for the creation of information goods—text, audio, images, video and more. + +GenAI, like the internet, is a GPT. And just as bits were a new atomic unit for the distribution of information goods, tokens are a new atomic unit for the creation of information goods-text, audio, images, video and more. + +It is hard to overstate the significance of the universality of tokens. + +It is hard to overstate the significance of the universality of tokens. GPTs tend to advance much faster than narrow purpose technologies for many reasons: since they have such broad applicability, they attract orders of magnitude more resources (more capital, more labor, more brain power); breakthroughs in one domain (or modality) often benefit others; they tend to create new bottlenecks that lead to adjacent innovations (for instance, the compute and energy demands of GenAI will undoubtedly propel advancements in both); and wider adoption means a broader user base and a faster feedback loop. So, I don't only mean advancements in the GenAI models themselves, but in tooling (like user-friendly interfaces and workflows) and integration with existing workflows and software. Like all technology, over time GenAI will get further abstracted away and will be seamlessly embedded in Adobe, YouTube Studio, TikTok, Soundcloud, Roblox, and probably ever other content creation tool and platform. + +General purpose technologies tend to advance far more quickly because they attract a lot more resources; breakthroughs yield benefits across domains; they compel complementary innovations; and they benefit from a much faster feedback loop. + +GenAI will greatly enhance current creators' capacity to create and, probably, the number of creators too. It may feel like there are a lot of creators already, but 114 million channels on YouTube, 10 million artists on Spotify, 4 million podcasts or 80,000 developers on Steam are all miniscule relative to the potential global population of would-be creators. + +## 3. The Quality Distinction Between Corporate and Creator Content Will Blur + +The biggest knock against creator content is that it's low quality, sh*t, crap, slop, garbage, choose your pejorative. + +The thing about this criticism is that it is objectively true. No one watches, listens to or plays most of the stuff on YouTube, Spotify or even Steam. On average, it is crap. The other thing about this criticism is that it is irrelevant. In a power law, there is no arithmetic average, and in a power law popularity distribution, the average is + +[https://archive.ph/wTgnR](https://archive.ph/wTgnR) + +12/22 + +# 4/23/25, 6:54 PM +The Relentless, Inevitable March of the Creator Economy + +inconsequential. What matters is the head of the curve, the most popular stuff. That's what's competing for consumers' time. And the "quality" of the head will likely keep getting better relative to corporate-produced content. + +Most creator content is not good, but most isn't what matters; the best, most popular stuff is what matters. + +### GenAl Production Values Will Keep Improving + +I won't belabor this, because anyone who has been paying attention knows that the output quality of GenAI text, image, audio and video models-whether Claude 3.5 Sonnet, Midjourney v6 (see below), Suno v.4 or Runway Gen-3-is advancing at a dizzying pace. + +The image shows a grid of faces, presumably generated by AI, labeled V1 through V6. The faces appear to be of older men with varying skin tones and facial features. The progression from V1 to V6 suggests an improvement in the realism and detail of the AI-generated faces. + +Source: Henrique Centieiro and Bell Lee. + +### The Consumer Definition of Quality is Shifting Toward Creator Content + +Another reason the quality distinction will blur is because the definition of quality itself is changing. + +Corporate media will have the edge in production values for some time, but production values are becoming less important to consumers. + +I often write about the shifting consumer definition of quality, such as here. In a nutshell, the idea is that quality is not a stated opinion or judgment, but is revealed preference: people's choices implicitly indicate that what they choose is higher quality to them than what they don't. These choices—and therefore the definition of quality- change over time. + +One of the biggest challenges for anyone who has been in a field for a long time is that they tend to get anchored to a relatively fixed definition of quality. Consumers' + +[https://archive.ph/wTgnR](https://archive.ph/wTgnR) + +13/22 + +# 4/23/25, 6:54 PM +The Relentless, Inevitable March of the Creator Economy + +definitions, however, are fluid. When new entrants enter markets with new features, they often change consumers' definition of quality in the process. This is especially true of younger consumers, whose definitions of quality aren't as established. + +The creator economy is introducing new attributes that are changing the consumer definition of quality, like authenticity, relatability, intimacy, social relevance (whether to a small community or to broad cultural fluency), digestibility, indie, underground, niche, low friction, etc. + +By inference, that's happening today across media. The creator economy is introducing new attributes that consumers clearly value, like authenticity, relatability, intimacy, social relevance (whether to a small community or to broad cultural fluency), digestibility, indie, underground, niche, low friction, etc. Every time that someone slumps on the coach and picks up their phone to scroll through Reels, rather than watch Netflix on the TV that sits mere feet away, they are implicitly indicating that Reels is "higher quality” than Netflix, at least in that context. + +It's also backed up by research. In a recent study of 12,000 video viewers by YouTube, 90% of respondents said that quality is determined by both technical (i.e., production value) and emotive markers. These emotive markers include "really means something to me personally," "is relevant to my interests and preferences,” and “is authentic and relatable." + +Very little of creator content needs to be good for it to yield a lot of good content. + +### Internet Scale + +The vast scale of creator content means that very little of it has to be good for it to yield a lot of good content. + +Refer back to Figure 10. Hollywood produced about 15,000 hours of new TV and film last year, compared to close to 300 million hours uploaded to YouTube. That means that if only 0.01% of YouTube content is considered competitive with Hollywood content (not comparable, but competitive for time), it would yield 30,000 hours of competitive content, 2x Hollywood's annual output. + +### Some Established Talent Will Defect + +One of the four "tectonic” trends in media that I write about is disintermediation: technology is making it easier for creators (and creatives, who are all latent creators) to produce, market, distribute and monetize content by themselves, increasing their bargaining power over intermediaries or enabling them to circumvent them altogether. + +Over the next decade, more established talent may start to question the relative benefit of sticking with traditional intermediaries. As economic pressure grows on traditional media companies, they will become more risk averse, stingier and generally less fun to + +[https://archive.ph/wTgnR](https://archive.ph/wTgnR) + +14/22 + +# 4/23/25, 6:54 PM +The Relentless, Inevitable March of the Creator Economy + +work with. At the same time, it will become increasingly viable and potentially more lucrative for talent to go it alone. + +This has already occurred in journalism. Top journalists like Matt Taibbi, Bari Weiss, Glenn Greenwald, Matt Yglesias, Casey Newton and others have left established news outlets for Substack to gain freedom and, apparently, generally make more money. Over time, this may become more common in other media too. + +## 4. Rising Distrust of Centralized Institutions and Demand for Authenticity Structurally Favors Creators + +In the U.S., and probably most of the west, trust in centralized institutions has been falling for decades. Trust in government is at all-time lows (Figure 11) and, more to the point, so is trust in mass media (Figure 12). + +Figure 11. Trust in Government Has Been Falling for Decades... + +The image is a line graph showing the public trust in government over time. The x-axis represents the years from 1960 to 2020, and the y-axis represents the percentage of people who trust the government. The graph shows a significant decline in public trust in government over the decades. + +Figure 12. ...As Has Trust in Mass Media + +The image is a line graph showing Americans' trust in mass media from 1972 to 2024. The graph shows a decline in the percentage of Americans who have a great deal or fair amount of trust in the mass media, while the percentage of those with not very much or no trust at all has increased. + +[https://archive.ph/wTgnR](https://archive.ph/wTgnR) + +15/22 + + +# 4/23/25, 6:54 PM +The Relentless, Inevitable March of the Creator Economy + +Source: Gallup. + +Trust and authenticity are complicated issues in the creator economy. Many creators +aren't considered authentic. Those who are can quickly lose trust and audience if they +are perceived as too commercial. + +Structurally, the direct relationship between creators and consumers creates more natural +conditions for perceived authenticity. + +But the creator-consumer relationship is parasocial: because it is often unvarnished, +unmediated and “un-institutional,” fans feel like they personally know the creator. +Structurally, this unmediated relationship creates more natural conditions for +perceived authenticity. Also, when a creator earns trust, it tends to be more personal +and resilient compared to institutional trust. + +## 5. The Demise of Monoculture + +Many have lamented the end of “monoculture,” big shared cultural experiences. As I +explained in Power Laws in Culture, cultural touchstones still exist-Taylor Swift, the +Super Bowl, Barbenheimer, GTA 6—but they are fewer and further between. +Underscoring the degree of atomization today, according to YouTube's recent Culture +and Trends Report, half of GenZ respondents say that they belong to a fandom that +"no one they know personally is a part of." + +We might be nostalgic for monoculture, but recall that mass media is only 100 years old. It +might not be the natural state. + +Most of the people reading this likely grew up with monoculture-I distinctly +remember the finale of M*A*S*H*, when over 100 million people tuned in-but keep in +mind that mass media is only 100 years old. We might be nostalgic for monoculture, +but perhaps it is not our natural state, at least not most of the time. + +Attention has atomized not only because there is much more choice, but, by inference, people +don't actually want a monoculture. + +Part of the reason that attention has fragmented is the massive increase in choice. +(Again, see Figure 10.) But the mere availability of vastly more stuff is an insufficient +reason. It must also be the case that people are choosing to spend their time with a +wider variety of content choices, or what we could call microcultures. + +Put differently, whether you think the decline of monoculture is good or bad, it's +happening because people prefer the alternative. We can infer a bunch of reasons why. +People have varied taste and they no longer need settle for homogenous content; in a +https://archive.ph/wTgnR +## 16/22 + +# 4/23/25, 6:54 PM +The Relentless, Inevitable March of the Creator Economy +world of near infinite choice, what you read/watch/listen to becomes a more powerful +way to signal identity and individuality; and it's more fulfilling to be part of a smaller, +more passionate, more engaged community, etc. + +But the reasons don't really matter. When offered more choices, consumers are taking +them. The implication is that as the relative volume of creator/independent content +choices grow, consumer attention will fracture even more. Economically, corporate +media is only viable if it programs to a wide audience. Further atomization into +microcultures definitionally means more share shift away from corporate media. + +## 6. Demographics Foretell a Perpetual Shift Toward Creators + +If you ever spend time around GenZ, or even occasionally see them slouched over a +phone at a neighboring table at a restaurant, it seems obvious that younger consumers +spend more of their time with creator content than do other age cohorts. It is probably +not worth litigating the point, but here are a few graphs for the heck of it: + +Figure 13. Over 1/3 of GenZ is on Social Media >2 Hours Per Day + +The image is a bar graph titled "Time spent on social media daily, 1% of respondents (n = 41,960)". The x-axis represents the amount of time spent on social media daily, divided into five categories: ">2 hours", "1-2 hours", "10 minutes-1 hour", "<10 minutes", and "Don't use social media". The y-axis represents different generations: Gen Z, Millennials, Gen X, and Baby boomers. Each bar represents the percentage of respondents in each generation who spend a certain amount of time on social media daily. For example, 35% of Gen Z respondents spend more than 2 hours on social media daily, while 23% spend 1-2 hours, 36% spend 10 minutes-1 hour, 4% spend less than 10 minutes, and 2% don't use social media. + +(1) Question: How much time, on average, do you spend on social media (not including +messaging apps) per day. Source: McKinsey Health Institute survey, April 2023. + +Figure 14. Almost 3/4 of Adults 18-29 Follow Creators + +The image is a horizontal bar graph titled "Follow influencers or content creators on social media". The y-axis represents different age groups: Total, Men, Women, Ages 18-29, 30-49, 50-64, and 65+. The x-axis represents the percentage of respondents in each age group who follow influencers or content creators on social media. For example, 40% of total respondents follow influencers or content creators on social media, while 36% of men, 42% of women, 72% of ages 18-29, 44% of ages 30-49, 26% of ages 50-64, and 12% of ages 65+ follow influencers or content creators on social media. + +Source: Pew Research Center survey of U.S. Adults, July 5-17, 2022. + +Demographics are destiny. + +As time marches on, these younger demos will make up a larger portion of the +consumer base and today's older demos will, well, not. If younger demos maintain +https://archive.ph/wTgnR +## 17/22 + +# 4/23/25, 6:54 PM +The Relentless, Inevitable March of the Creator Economy +their disproportionate usage of creator content as they age, it will be a perma-tailwind +for the creator economy. + +## 7. The Monetization Gap Should Narrow + +The creator media economy's share of M&E revenue lags its share of time spent, +although it's hard to tell how much. + +Above, I estimated that the total creator media economy is about 10% of M&E revenue +globally. That's probably substantially lower than its share of time. As shown in Figure +15, I estimate that social video represents about 1/4 of all time spent with video in the +U.S. (For more detail on how I derived this, see here.) And, as shown in Figure 16, +according to Spotify, about 1/4 of all streams are now derived from artists not +represented by the majors or Merlin. These are probably decent proxies for the share +of total media time spent with creator/independent content. + +Figure 15. Social Video is ~1/4 of Total Video Consumption + +The image is a bar graph titled "Social Video Time Spent vs. Other Video Total Sample (ADJUSTED)". The y-axis represents "Hours: Minutes" ranging from 0:00 to 9:36. The x-axis is labeled "2024". The graph shows the time spent on different types of video: Linear, SVOD, FAST, and Social Video. Social Video accounts for 24% of the total video consumption. + +Source: Maverix Insights MIDG data, Nielsen, Author analysis. + +Figure 16. Similarly, About 1/4 of Spotify Streams are Attributable to Creators/Independents + +The image is a line graph titled "Share of Spotify Streams for Majors and Merlin". The y-axis represents the percentage ranging from 50% to 100%. The x-axis represents the years from 2017 to 2023. The graph shows a downward trend, indicating that the share of Spotify streams for majors and Merlin has decreased over time. + +Source: Spotify. +https://archive.ph/wTgnR +## 18/22 + +# 4/23/25, 6:54 PM +The Relentless, Inevitable March of the Creator Economy +Over time, the gap between creator economy share of money and share of time should narrow. + +Over time, this monetization gap should narrow, even if it won't likely close +completely. + +* "Money follows eyeballs, with a lag.” This is an old expression in the marketing + business. It lags because new outlets necessitate new formats and creative; + measurement and attribution; planning and budgeting processes and cycles, etc. + Plus, a lot of ad allocations are still driven by relationships. Most advertisers don't + do zero-based budgeting, starting from scratch each year, but base their current + year media plans in part on last year's. But, as new practices, processes and + systems fall into place, budgets eventually shift. +* There is an ongoing mix shift to digital-native enterprises. Just as younger + consumers tend to spend more of their time and money on creator content, + younger businesses do too. There is a kind of "demographic effect" in the + enterprise. These digital-native businesses allocate more of the their budgets to + the creator economy, so as they inevitably become a larger proportion of the + global economy, this represents another tailwind. +* Creator monetization models should continue to mature. Current creator + monetization models are still relatively young. Subscription and patronage + platforms like Patreon and Substack only emerged in the last decade (Patreon + launched in 2013, Substack in 2017). Primarily ad-supported platforms, like + Instagram, YouTube and X/Twitter, have only recently enabled creators to offer + subscriptions. Just as traditional media took decades to optimize its business + models (cable bundles, retransmission fees, windowing strategies), the creator + economy should see similar refinement and "hardening" of business models over + time. + +## "Less Than" or Not, It's Where the Growth Is + +I used the words “inevitable and relentless” in the title of this piece because there are +so many tailwinds at the back of creator media, it's hard to see why the trend reverses. +It's really just a question of how fast it proceeds. + +For creators, the future is likely a mixed bag. It's great to have the wind at your back +and monetization tools and models should continue to improve. The offset is that +competition is near infinite, power laws are merciless, and the ranks of losers will +outnumber the winners by many orders of magnitude. + +Creatives will face a perpetual question of when and whether it is better to +disintermediate traditional intermediaries and go direct. For many creatives, they have +not historically thought like owners, but ownership of their output—and creative +control-will be an increasingly viable option. + +For traditional media companies, the growth of creator media may be unsettling, but +it's time to move into the acceptance phase of the five stages of grief. There are only +two choices: figure out how to participate in the creator economy or accept a +perpetually shrinking business. +https://archive.ph/wTgnR +## 19/22 + +# 4/23/25, 6:54 PM +The Relentless, Inevitable March of the Creator Economy + +The image is an advertisement for WSC Sports. The ad features the text "WSC SPORTS" in a white, bold font. Below that, it says "Monetize content by starting your own official creators program" in a larger, white font. There is a "LEARN MORE" button in yellow. To the right of the text, there are four images of sports highlights. + +1 In a nod to Samir's distinction between creative and creator, note that I've used the term +"creative" in Figures 1 and 2 and "creator" in Figure 3. + +2 Note also that I have avoided using the word "professional" in these definitions, because +plenty of creators earn money and are, therefore, professionals. + +3 Through the first nine months of 2024, Meta and YouTube advertising have grown by 22% +and 15%, respectively, good proxies for overall creator media economy growth. + +Subscribe to The Mediator +By Doug Shapiro + +The Mediator is (mostly) about the long term structural changes in the media industry and the business, +cultural, and societal implications of those shifts. I write it to get closer to the frontier. + +By subscribing, I agree to Substack's Terms of Use, and acknowledge +its Information Collection Notice and Privacy Policy. + +72 Likes. 17 Restacks + +72 +10 +17 + +Previous +Discussion about this post +https://archive.ph/wTgnR +Comments Restacks +Share +Next → +## 20/22 + +# 4/23/25, 6:54 PM + +The Relentless, Inevitable March of the Creator Economy + +Write a comment... + +Jonathan Glazier Dec 1 +❤Liked by Doug Shapiro +Great post. I probably take slight issue with the characterisation that "we" the establishment are a bit +sniffy toward the creator community. I think we rather look toward it with envy. The envy born from the +creative freedom and lack of barriers to entry. When the internet was conceived by Tim his vision was +for democratisation of content IP writing etc now the internet is owned by big players, manipulation by +agents on all sides is rife and algorithms have become the new gate keepers. And the creator +community is becoming owned and controlled in the same way. So the platforms used by the creators +are used just as much by the establishment a video clip from one of my shows featuring the sacred +Rihanna is still up there in terms of views. Every production has a digital strategy. So do I see the two +entities as warring factions, no and I certainly don't treat it or any new creators with any lack of respect. +I look to them for inspiration! +LIKE (4) REPLY SHARE + +☑ Spencer Parlier Dec 26 +❤Liked by Doug Shapiro +This is brilliant, Doug. Enjoyed the post-Christmas reading. + +One platform to watch in 2025 is Bleacher Report, especially regarding your last paragraph. B/R (a +subsidiary of WBD/TNT Sports) has made it a mission to embrace the creator economy while remaining +under the traditional corporate media umbrella. + +The platform always invited users to engage with, and sometimes, create their content, but mainly via +the written form (this was the original mission of B/R before it got scooped up by Turner when the +blogosphere was still dominating as the "new kid on the media block"). Now they have launched their +"creator program," allowing users to "go live" on video in their product as a reaction to certain games +and other tentpole events in the sports world. + +While leaning toward the slightly vague branding as "Twitch but for Sports" B/R still hasn't reached the +level of Amazon's platform as it still has creators go through a thorough vetting process before +allowing them the tools to go live, strongly gatekeeping who and who can't use their live video tools in +their app. I believe the vetting process /before/ going live is probably constrained due to staffing on +the content moderation side. (Maybe Al can help alleviate this problem down the road...?). + +Although I can't go into too much detail, I do know that B/R is going to lean into this strategy even +more in 2025 with the launch of an updated product. This paired with B/R's partnership with House of +Highlights and its Creator League (https://www.youtube.com/@CreatorLeague) makes it a brand to +watch as creator and corporate economies continue their tug-of-war in the back half of this decade. +LIKE (2) +REPLY SHARE + +1 reply by Doug Shapiro + +8 more comments... + +Top Latest Discussions + +The image shows a card with the title "28 Days of Media Slides" and the subtitle "An Industry in Upheaval". It also includes the date "JAN 7 DOUG SHAPIRO" and some social media interaction icons with numbers 53 and 9. There is a thumbnail image on the right side of the card. + +28 Days of Media Slides +An Industry in Upheaval +JAN 7 DOUG SHAPIRO +53 +9 + +https://archive.ph/wTgnR + +# 21/22 -- 2.45.2 From f205ec04f36b07d5f8de8068ee9d29a6a88632cb Mon Sep 17 00:00:00 2001 From: Teleo Agents Date: Thu, 19 Mar 2026 16:48:58 +0000 Subject: [PATCH 5/5] extract: shapiro-scarce-when-quality-abundant Pentagon-Agent: Epimetheus <3D35839A-7722-4740-B93D-51157F7D5E70> --- .../shapiro-scarce-when-quality-abundant.json | 34 +++++++++++++++++++ .../shapiro-scarce-when-quality-abundant.md | 16 ++++++++- 2 files changed, 49 insertions(+), 1 deletion(-) create mode 100644 inbox/queue/.extraction-debug/shapiro-scarce-when-quality-abundant.json diff --git a/inbox/queue/.extraction-debug/shapiro-scarce-when-quality-abundant.json b/inbox/queue/.extraction-debug/shapiro-scarce-when-quality-abundant.json new file mode 100644 index 00000000..bc86da2a --- /dev/null +++ b/inbox/queue/.extraction-debug/shapiro-scarce-when-quality-abundant.json @@ -0,0 +1,34 @@ +{ + "rejected_claims": [ + { + "filename": "consumer-quality-definition-is-revealed-preference-not-production-value.md", + "issues": [ + "missing_attribution_extractor" + ] + }, + { + "filename": "value-accrues-to-scarce-resources-and-shifts-when-relative-scarcity-changes.md", + "issues": [ + "missing_attribution_extractor" + ] + } + ], + "validation_stats": { + "total": 2, + "kept": 0, + "fixed": 4, + "rejected": 2, + "fixes_applied": [ + "consumer-quality-definition-is-revealed-preference-not-production-value.md:set_created:2026-03-19", + "consumer-quality-definition-is-revealed-preference-not-production-value.md:stripped_wiki_link:disruptors-redefine-quality-rather-than-competing-on-the-inc", + "value-accrues-to-scarce-resources-and-shifts-when-relative-scarcity-changes.md:set_created:2026-03-19", + "value-accrues-to-scarce-resources-and-shifts-when-relative-scarcity-changes.md:stripped_wiki_link:giving-away-the-commoditized-layer-to-capture-value-on-the-s" + ], + "rejections": [ + "consumer-quality-definition-is-revealed-preference-not-production-value.md:missing_attribution_extractor", + "value-accrues-to-scarce-resources-and-shifts-when-relative-scarcity-changes.md:missing_attribution_extractor" + ] + }, + "model": "anthropic/claude-sonnet-4.5", + "date": "2026-03-19" +} \ No newline at end of file diff --git a/inbox/queue/shapiro-scarce-when-quality-abundant.md b/inbox/queue/shapiro-scarce-when-quality-abundant.md index 3213b6a0..50b3e624 100644 --- a/inbox/queue/shapiro-scarce-when-quality-abundant.md +++ b/inbox/queue/shapiro-scarce-when-quality-abundant.md @@ -7,10 +7,14 @@ date_published: "2023-10-01" date_archived: "2025-04-23" archived_by: "clay" domain: "entertainment" -status: unprocessed +status: null-result claims_extracted: - "consumer definition of quality is fluid and revealed through preference not fixed by production value" - "fanchise management is a stack of increasing fan engagement from content extensions through co-creation and co-ownership" +processed_by: leo +processed_date: 2026-03-19 +extraction_model: "anthropic/claude-sonnet-4.5" +extraction_notes: "LLM returned 2 claims, 2 rejected by validator" --- # What is Scarce When Quality is Abundant - by Doug Shapiro @@ -555,3 +559,13 @@ JAN 20 DOUG SHAPIRO * The first image is a thumbnail for "28 Days of Media Slides" and features a dark blue background with white text that reads "28 Days of Media Slides" in a stylized font. * The second image is a thumbnail for "Quality is a Serious Problem" and features a person sitting in front of a television screen displaying the HBO logo. The person is looking at the screen with a thoughtful expression. + + +## Key Facts +- Hollywood released approximately 15,000 hours of new TV and film content in the U.S. in 2022 (496 films averaging 100 minutes plus ~2,000 TV series averaging 10 episodes of 40 minutes each) +- YouTube disclosed in 2019 that 500 hours of new video are uploaded every minute, or 30,000 hours per hour, equivalent to over 250 million hours per year +- According to Nielsen (August 2023), YouTube is the most-streamed service on U.S. TVs at 9.1% share, exceeding Netflix (8.2%) and equal to Hulu, Disney+, Max, Peacock, and Paramount+ combined +- Mr. Beast's recent video 'World's Most Dangerous Trap!' garnered over 100 million views in its first week with a 20-minute runtime, comparable to Netflix's top global series +- Warner Bros. library consists of more than 145,000 hours of programming including 12,500 feature films and 2,400 television programs +- Manifold Markets bettors gave 26% probability that a film created using text-to-video generator will be nominated for an Academy Award by 2030 (as of October 2023) +- Impact investing is a $1.57 trillion market with 92% of investors citing fragmented measurement as a concern and $19.6 billion fleeing U.S. ESG funds in 2024 -- 2.45.2