'Stackin paper, game on, Descript, streaming's world tour, and robocookies

January 15, 2021 Edition (#38)

Well, we made it to another Friday, for better or worse. Congratulations to us 👏👏👏. Hope everyone is staying healthy and sane.

“It’s 2021, but I’m still writing ‘yearlong fever dream of chaos and despair’ on my checks.”

Source: Joedator, New Yorker, January 2021.

Stack it up, Stack it in

I do enjoy my meta moments. Substack’s charmed year, wherein seemingly every tech publication and VC lionized the company as the savior of the printed word, has evolved into a period of extensive evaluation conducted by the media-industrial complex. A few weeks ago, we wrote about Substack’s announcement of its new Reader app—designed to make it easier for consumers to organize their subscriptions and to enable recommendations and reader/writer discovery—and cited two relatively critical pieces in the Columbia Journalism Review and Nieman Lab.

Shortly thereafter, the New Yorker’s Anna Wiener wrote a similar disquisition on Substack and its cultural impact. [Ed.: the title is: “Is Substack the Media Future We Want?” so no points for guessing the overall tone of the piece.] Wiener’s article is worth a read, and not simply because she is a proud denizen of Trillium’s former San Franciscan neighborhood. Hers is a thoughtful examination, with myriad anecdotes about the journeys undertaken by card-carrying members of the Substackerati, and the relationship between Substack & “the blogosphere” (similarly homogeneous in terms of writer profile and stylistic bent). It also provides a well-researched history of the newsletter, and the historical context in which Substack sits.

Source: Ard Su (illustration), New Yorker.

Wiener describes the ways in which Substack has evolved from a platform for paid newsletters into something that more “resembles a newsroom,” complete with recent perks like hefty advances for certain “high-profile” writers, healthcare stipends, design help, and the promise of selectively deployed legal defense-as-a-service. (The company’s founders determine which cases and writers are eligible for this benefit, and the program is in very early days.) The overall impression is of a company that understands that its value is a function of the quality of the writers it is able to attract, and thus is willing to experiment with different methods of engaging and retaining this talent.

I’m sure there is some harried growth hacker’s magical spreadsheet that suggests that it is, in fact, economically rational to offer a $250,000 advance to a top writer, if one labels it a cost of customer acquisition and models the lifetime value of said writer with an appropriately lofty growth trajectory and retention period. And, frankly, if advances and benefits are the cost of doing business, so be it. Either the company’s economic model will work, or it won’t, and investors and employees will be happy or sad. I’m not especially concerned that journalistic or writerly purity is diluted by some of the incentives at play. Since you asked, sure, I’d be happy to be even less concerned if that advance was headed to my bank account, since nobody in this town works without a retainer.

Wiener, for her part, seems piqued in part by the broad-strokes idealism [Ed.: I suppose one could say “grandiosity”] of Substack’s original mission statement, in which the company’s founders talk about journalism’s “intrinsic value.” It’s true: she shares anecdotes about the founders and investors that are very Silicon Valley and don’t necessarily paint the venture in the most humble light [Ed.: see “pioneering a ‘new business model for culture.’”] Alas, that’s kind of the state of play out here, so it’s hard to get too up in arms.

Where Wiener’s balanced assessment and concerns feel especially nuanced is in her questioning where Substack falls between an objective platform—a relevant question given the company’s hosting of a number of politically controversial publications—and media company; the answer to this question also has ramifications for how the company might be valued and regulated over time. Wiener hits the nail on the head with her analysis of the broader “passion economy trend,” which offers the promise that anyone can make a full-time living doing what they love. As Wiener notes, there is a meaningful movement at play here, but it’s far more difficult for the average writer or creator to make a proper living through tools like Substack than most (spiritually) invested VCs and technologists might like to admit.

Part of the reason why it’s so hard for a talented but unknown creator to support themselves through Substack, or its equivalents across other media types, is the ever-present question of consumer discovery. As I’ve mentioned previously, the dominant dynamic for Substack is that someone with a lot of Twitter followers sets up a paid newsletter and then promotes it actively on social media. Typically, somewhere in the low-to-mid single digit percentages of their followers convert into paying customers. You don’t have to be Will Hunting to realize that, if you don’t already have a substantial number of people following you elsewhere on the Internet, that conversion rate isn’t exactly going to pay for a new Lambo [Ed.: or an old one, for that matter]. In summary: if you’re already popular, now you can make more money off that popularity. If you’re not, it hasn’t become any easier to get there.

I don’t consider this to be a moral failing on behalf of Substack (or its leadership), but simply a reality of the company’s business model, product dynamics, and the fact that it’s hard to help creators find new potential customers. Substack, via Reader, is starting to experiment with ways to address this challenge, and I have no doubt that we’ll see more such attempts to come. But, until it does, the general contention that Substack reinforces the “power structures” that currently exist in “culture,” writ large, is hard to dismiss out-of-hand. [Ed.: Keep in mind that it is the serendipity that TikTok offers to heretofore-unknown creators who one day receive a massive influx of traffic that has won the hearts and minds of users across the globe.]

For a thoughtful analysis of Substack’s business and market impact, I recommend Ben Thompson’s recent piece, in which he explores some of the core questions about the company: its evolution into a proper platform (ongoing), its cost structure (favorable), and whether it offers enough value to retain top writers who might realize that they can effectively disintermediate the company once they reach a certain critical mass of readers (unclear, and possibly an area for concern).

All the world’s a game

The Guardian’s Sean Monahan recently wrote an article in which he claimed, decisively, that “video games have replaced music as the most important aspect of youth culture.” Monahan’s not wrong; the numbers tell a clear story. In 2020, the global video game industry generated an estimated $180bn in revenues (+20% YoY), eclipsing sports and movies. Granted, 2020 was a challenging year for certain industries—including those involving live events and theaters—but the facts remain: the gaming industry is ascendant, multi-faceted, and central to global culture.

Monahan cites gaming’s increasing push to the mainstream, referencing the Biden Campaign’s Animal Crossing integration and the streaming forays of Alexandra Ocasio-Cortez, amongst other politicians and “mainstream celebrities.” He also points out that revenue is coming from all corners of this world: new games, brand partnerships, consoles (Switch, PS5), e-sports, and gamer livestreaming, to name just a few. Trevor McFedries—co-founder of Brud and originator of Avatar Celeb Lil Miquela—makes an appearance in the piece, via a recent Tweet stating that “gaming is replacing music as the lynchpin of emergent social scenes.” This is the thesis of so many Trillium snippets over the past few months, including last week’s issue in which we wrote about Roblox as a platform for a broader range of social experiences, not simply traditional gameplay.

I agree with Monahan (and McFedries), though I wonder to what extent gaming is “replacing” the music, sports, and movie worlds. This logic puts gaming on a parallel level with these other verticals. Instead, the fast-arriving reality is that gaming is something different. Yes, it encompasses traditional product types—hardware and software—and shares many similarities to other creative industries (including a Hollywood-like reliance on multi-year development cycles, big marketing budgets, and celeb-driven publicity). Gaming is also a platform that can underpin a vast range of work produced across other industries, which themselves may evolve by leveraging the new creative activations and distribution made possible by the reach of gaming platforms. If the concert-stage has moved into video games, why can’t the theater and the field/court?

Setting aside the next-gen digital creator experiences that we often discuss in these pages, it’s reasonable to expect that certain products or experiences will always require interaction IRL. Beyond these, there’s no reason that access/distribution, community engagement, and monetization can’t be built on the foundation provided by games.

De$cript’s latest haul

You may recall past Trillium mentions of Descript. As a refresher, per Tech Crunch’s Ingrid Lunden, the company “builds tools that lets creators edit audio and video by using…natural language processing.” (Tech Crunch) Simply: imagine if you could edit an audio or video file the same way you edit text in a Google doc. It’s a slick tool, largely because it doesn’t require a creator to have significant knowledge of editing tools—nor code—in order to achieve professional-quality editing and production.

Source: Descript.

This week, the company announced $30M in fresh Series B funding, led by Spark Capital. Originally, the company targeted podcast producers; they’ve now expanded to offer tools to video editors. One of the cooler—but creepier—features is the ability to create an “overdub” bot that uses actual recordings of you to generate net-new media snippets that you can insert into your work. [Ed.: Yes, certain foreign cybercommands are likely big fans of this feature, too.] While tools that can be deployed to aid in deepfake production carry a certain and unavoidable implicit risk, I find Descript interesting because its product is a tangible example of a theme we’ve discussed often in the past: products with machine learning & AI at their core that give “creative superpowers” to users, thus expanding any individual’s technical abilities and making it easier for novices to engage in creative pursuits.

Corridonline

That is a rather niche Mexican musical genre/Internet pun that has an audience of exactly 2.5 (Reader RB is on the fence). We’ve been fascinated by the impact of streaming platforms like Spotify on local music cultures, and the way in which certain unexpected “trigger cities” like Jakarta and Sao Paulo have, through sheer population size, become important geographies for artists looking to break new music. Because of the way in which streaming services report global play counts and listener engagement, fan behavior across the world has been “flattened,” so to speak; leaderboards and algorithms assessing song popularity to make recommendations to listeners are working in real-time off a global dataset in a way that was practically impossible in the pre-streaming era.

While the identity of micro music scenes—and the music produced within—has changed as streaming has grown, the impact on artists has been multi-faceted. In a recent analysis, Francisco Toscano and Jason Joven of Chart Metric analyzed how “global music streaming services…are driving Mexican music genres like Mariachi, Banda, and Corrido worldwide.” As your humble author is one of the elusive and highly official non-Mexican ambassadors of Mariachi worldwide, this piece spoke to Trillium’s core.

One of the interesting aspects of Toscano and Joven’s analysis is their identification of curated/editorial playlists as an enabler of global discovery of artists in previously niche Mexican music genres. It’s well-established that artist discoverability is largely a function of inclusion in these sorts of playlists, along with proactive platform promotion (#P3), but nonetheless interesting to observe the sustained impact of these playlists on genre popularity. Deezer, in particular, is cited as differentiating on the basis of its regional Latin music inventory, with nearly 25% of its “Browse” playlists featuring these genres. Other services have similar offerings, leading to remarkable growth in consumption of “Regional Mexican” music.

Source: Chartmetric & YouTube.

This warms my heart, much as I detest Banda music. Most of this activity is from Latin American listeners, or the Latin diaspora globally, but I can’t help reveling in visions of Thai Gen Zers belting out Mariachi classics about lost love and vengeance in front of their bedroom mirrors. The chart above is also a testament to the role of these digital platforms in driving listener awareness of artists new (Christian Nodal, neck and neck with Taylor Swift) and old (Vicente Fernandez, who is compared favorably in listenership, and certainly in spirit, with artists like Lil Pump, The Chainsmokers, and Kacey Musgraves).

Source: Chartmetric.

The whole article is worth reading if you’re at all curious about “Trap Corrido” and “Mariachēno.” Three things are clear. First, streaming services have enabled truly global demand for an artist’s product to be created and fulfilled. Second, algorithmic promotion and merchandising of an artist’s work, through playlists and recommendations, is an opaque but important driver of popularity. Third, the digital platforms control the relationship with the end-customer (the listener) and thus have complete leverage to dictate financial terms. It’s easy to see the tension at play for artists, and the reason that many artists are increasingly drawn to paths involving direct-to-consumer strategies where they have greater control over their commercial destiny. [Ed.: we made the mistake of watching a few Christian Nodal videos, so please accept this majestic photo of OG Jose Alfredo Jimenez as a palette cleanser.]

Source: Let’s Loop.

Robobaking

Ah, baking. That thing you’ve been doing over the past 9 months as an occasional replacement for rocking yourself to sleep in the fetal position while whale song plays in the background. Everyone loves a nice home-baked good, so we couldn’t resist sharing a fun little story (Input) combining two of Trillium’s favorite things: refined sugar and unnecessary applications of artificial intelligence.

Last week, we talked about another Google AI experiment: the Blob Opera, as showcased in Trillium’s spectacularly musical magnum opus. This week, we bring you news of a Google AI project that uses a Google Cloud feature (AutoML Tables) to digest (👏) what goes into recipes, categorize the output, and, thus, acquire the ability to “create entirely new recipes like the breakie and the cakie.” In case you’re keeping score at home, those are exactly what you think they are. If robo-generated bread, cookie, and cake hybrids appeal to you, or you’re simply trying to figure out what to do with the metric ton of yeast you ordered last March, knock yourself out. Personally, I’m more of an equal opportunity employer with a “what’s in a name” stance on categorizing baked goods, but a little extra inspiration never hurt. Details below:

For your ears only

In a feeble attempt to manifest the spirit of healthy union, I bring you news of an upcoming LP collaboration between legendary producer Madlib and English electronic all-star Four Tet (Kieran Hebden). If you aren’t familiar with either name, fear not: you are in for a treat.

Madlib tends to skew more towards sample-based and beat-driven production—including Freddie Gibbs’ incredible Piñata and Bandana albums, not to mention his classic with MF Doom (Madvillainy)—while Four Tet produces a range of beautiful instrumental music, sometimes verging on ambient, sometimes clubby, and sometimes with Eastern influences. The pairing is excellent, to say the least. Two early releases are below, with the full album coming on January 29th.

See you all next week.

N