The reality is that we are all addicted to convenience — and streaming platforms make it damn convenient for us to not think about the artists and how they manage to survive. As streaming becomes more pervasive, the sad reality is that every track, every artist, every album is reduced to just data, served up by the algorithm. It only continues to devalue our emotional relationship with the creators.
Though universally revered, Martin Scorsese is sometimes viewed as an old-fashioned relic as he digs in his heels against changes in contemporary media. Previously, he got lots of nerdy flack for referring to superhero franchise films as “theme parks” rather than “cinema.” And, recently, in an essay on Federico Fellini, Scorsese went off on algorithms and the overuse of the word “content” to describe artistic output. He’s mainly referring to visual media, of course, and how “the art of cinema is being systematically devalued, sidelined, demeaned, and reduced to its lowest common denominator” when we refer to it all as “content.” Here’s Scorsese:
“Content” was used more and more by the people who took over media companies, most of whom knew nothing about the history of the art form, or even cared enough to think that they should. “Content” became a business term for all moving images: a David Lean movie, a cat video, a Super Bowl commercial, a superhero sequel, a series episode. […] … it has created a situation in which everything is presented to the viewer on a level playing field, which sounds democratic but isn’t.
A platform’s reliance on algorithms that can’t separate artistic intention from specious cash grabs exacerbates this perception. There’s so much talk about freeing ourselves from the gatekeepers, but perhaps ‘old-fashioned’ human curation is a gatekeeping we need. Scorsese again:
Curating isn’t undemocratic or “elitist,” a term that is now used so often that it’s become meaningless. It’s an act of generosity—you’re sharing what you love and what has inspired you. Algorithms, by definition, are based on calculations that treat the viewer as a consumer and nothing else.
“Scorsese is right,” tweeted music critic Ted Gioia. “Anyone who refers to film, music, or writing as ‘content’ is simply not a trustworthy custodian of anything of cultural value. Unfortunately, these are the key decision makers in media right now.”
I don’t have too much of a problem with media companies calling the music or movies they stream “content.” It’s like a politician using blatant dog whistle language — at least you know who’s in this for the right reasons and deserving of trust. What’s insidious is when we, as artists, are convinced to start using the word “content” instead of “art” or even “our work.” A musician creates a beautiful song, puts sweat into editing an accompanying video, and then thinks, “here’s some content for YouTube” — that’s distressing.
Language is powerful, and the words we use in our heads change our behaviors. If we start replacing words like “art” with “content” — even just internally — our intentions shift. We start feeding the companies hungry for content. Instead of making music and films for the fans or the human curators, we’re producing content for the algorithms.
Seth Godin must have read Scorsese’s rant. Soon after the essay’s publication, Seth wrote his own rant on his daily blog:
Publishing to an algorithm is not the same as publishing to an audience. And living in a culture that’s driven by profit-seeking algorithm owners is different as well. Because without curation, who is responsible? Who is guiding the culture? Who pushes the boundaries or raises the standards? […] …we benefit when we realize that the algorithm isn’t rooting for us and quite probably is working against us. The only winning approach is to earn permission and a direct connection with our fans and then act as curators for ideas (and as our own publishers).
Getting back to the power of language, I touched on this topic on the blog a few years ago when I commented on Cherie Hu’s idea that “The word ‘creator’ does more harm than good.” (Cherie’s original essay is offline, but I wholeheartedly recommend her Water & Music platform, where you can find many of her enlightening pieces.) I wrote this in my blog post:
It may seem like semantics, but the way we adopt and use language rewires our thinking. Hu’s point— which I never considered — is that the more we refer to ourselves as ‘creators,’ the easier it is to submit to the notion that our creations are in fealty to others. Notice how the services almost all use ‘creator’ — a sampling of examples Hu points out include YouTube Creators, Facebook for Creators, Spotify’s “Creator Marketing.’ So when a platform sneakily claims ownership of our work we’re desensitized against protest.
“Content” is the same. The language implicates employment, that we’re delivering goods in a fiefdom. Responsibility, leverage, and agency shift to the “content provider.”
Buckle down, folks. Dig in your heels like Martin. You’re artists making art. Don’t let anyone tell you anything else.
Sites that only obey the Long Tail and the primacy of the algorithm have fewer standards. They view curation as a last resort, and if mass is the standard, then mass is all that will be rewarded.
Spotify’s infamous recommendation algorithm is a hot topic on this blog, under fire for pay-to-play schemes and encouraging saccharine content. Stereogum’s Nate Rogers touches on both aspects while exploring how an obscure Pavement b-side became the band’s most popular song on the streaming platform. No one is certain of the reason for this — fucking algorithms, how do they work? But the song did start collecting massive play counts in early 2017 when Spotify switched Autoplay ‘on’ for everyone by default.
The Autoplay feature on Spotify plays a stream of songs automatically once you’ve finished listening to an album, its selection based on that album’s sound. Autoplay also is enacted when you launch Spotify’s ‘radio’ function. That function is also based on a band or a song’s sound — you could specify ‘LCD Soundsystem Radio,’ for example. I’ve written before about how Spotify uses Autoplay to keep you listening to the platform in a way that pays fewer royalties.
The theory goes that, for whatever mysterious reason, Spotify’s algorithm loves Pavement’s “Harness Your Hopes.” Whenever, in Autoplay mode, the algorithm selects a song from Pavement, that’s the one it picks.
Damon Krukowski has noticed something similar. The Galaxie 500 song “Strange” is similarly the most popular song on Spotify from the band by a wide margin. Damon was puzzled as the song was never a single and “not particularly popular in the past” (which I’ll dispute as I’ve always loved that song). But its rise on Spotify coincided with the ascendance of “Harness Your Hopes” — January 2017. That pesky Autoplay algorithm.
It’s nice that these deep cuts get thrust in the Spotify spotlight, even though Autoplay streams pay much lower royalties than intentional streams. But why are these songs sticking out? It’s argued that “Harness Your Hopes” is a quintessential Pavement song — not as crazy or weird or (and I don’t mean this disparagingly) memorable as other titles in their catalog. I’m sure the band agrees. Krukowski wonders about this, too, with regards to “Strange”:
“‘Strange’ is a touch faster, louder, with a more regular backbeat and a more predictable song structure than most Galaxie 500 songs,” he pointed out on his blog. “Might an unintended result of Autoplay, then, be the separating out and rewarding of the most ‘normal’ songs in each band’s catalog…? … As albums are increasingly supplanted by playlists, and intentional listening of all kinds is increasingly replaced by algorithmic recommendations, ‘Play Galaxie 500’ may really come to mean, ‘Play the song by Galaxie 500 that most resembles songs by others.'”
That sounds worrying, but keep in mind that Autoplay is a passive listening mode. It’s playing in the background for most listeners. So keeping the crazy or weird or memorable at bay is desirable. The music shouldn’t linger or provoke by design.
The problem is the list of ‘top songs’ on an act’s Spotify artist page. These Autoplay ‘passive’ listens are treated the same as intentional listens. Though purposefully selecting to listen to a song or album holds more weight for the artist — both in royalty and fan-building — it’s treated the same as a passive, in-the-background listen. One hundred passive Autoplay streams are identical to one hundred intentional plays when determining a band’s top songs. So, when you go to Galaxie 500’s Spotify page, you’ll see “Strange” as the top song at 11,680,597 plays.
“When Will You Come Home” is probably a song more beloved by fans, and it’s certainly more indicative of Galaxie 500’s sound, but it’s stuck at 1,439,734 streams. That seems measly compared to the top song’s count. But, assuming those million-and-a-half streams are intentional plays as opposed to Autoplay-ed, that song has a lot more relevance than an algorithm’s inscrutable choice.
Update: Damon Krukowski got in a spirited discussion with Spotify’s Glenn McDonald over the issues raised in the Stereogum article Check out the thread on Twitter.
Here’s an informative video that inspired a spirited conversation in my household:
I can understand the overwhelming temptation to appease and filter for THE ALGORITHM when you’re making a living off your YouTube and online efforts. But there’s a loss of voice, akin to the loss of agency that casual listening creates. Rather than choosing we’re being chosen.
There’s an SEO plug-in installed on my blog, and it tells me that I should optimize my titles and my content for traffic-catching metrics. I was paying attention to its demands for a while, changing snappy short titles to longer (less fun) ones that complied with SEO-recommended character limits. I was told to insert keywords into the content and always attach eye-catching featured images (I was never good at that aspect). I’d get stressed out when the plug-in told me that I wasn’t adhering to the internet’s mysterious ordinances.
But, here’s the thing: there are many blogs and newsletters that I love, and none of these follow the rules. Some of these authors have a voice that flies in the face of these rules — one-word titles, blog posts with only a few sentences, the minimalist of
There are those who are going for mass and, according to their goals, choose to follow the rules. That’s cool. Derek Muller, in the video above, is doing so grudgingly but he’s in the machine where 5 million vs. 10 million views have a direct impact on his project’s budget. Veritasium is a channel with — at the least — staff and travel requirements, so there’s a good reason to keep the views and funding high.1And it’s a well-done science channel, worth your subscription in my opinion. But my needs — and the needs of the bloggers and channels I follow — are different. We’re looking for that longterm connection and an audience that’s on board because of our way of saying things. It’s not that Derek doesn’t want that — it’s evident from his video that his voice and point of view are essential and he’s struggling with maintaining these — but the success of his platform is affected by more important things. On the other hand, at this point, I measure my success by people enjoying what I do and, to paraphrase Seth Godin, missing me when I’ve gone quiet.
I’m thinking about the direction of this blog all of the time. I change my mind about it constantly. But what always sticks is that I want it to be fun for you and — most importantly, if I’m honest — fun for me. I’m using this to find my people, to have conversations, and make those connections. And I’m using it to find potential friends, collaborators, clients, and employers. An SEO plug-in’s suggestions won’t be much help there. I see these imposed guidelines as a detriment and I’m happy to resist. So expect the blog to get more personal, more specialized, and obtuse — my private-made-public playground. No holding back the freak flag.
Dani Deahl in The Verge:
The word “human” does not appear at all in US copyright law, and there’s not much existing litigation around the word’s absence. This has created a giant gray area and left AI’s place in copyright unclear. It also means the law doesn’t account for AI’s unique abilities, like its potential to work endlessly and mimic the sound of a specific artist. Depending on how legal decisions shake out, AI systems could become a valuable tool to assist creativity, a nuisance ripping off hard-working human musicians, or both. […]
If [an AI] system then makes music that sounds like Beyoncé, is Beyoncé owed anything? Several legal experts believe the answer is “no.” […] “There’s nothing legally requiring you to give her any profits from it unless you’re directly sampling,” [Public Knowledge policy counsel Meredith] Rose says. There’s room for debate, she says, over whether this is good for musicians. “I think courts and our general instinct would say, ‘Well, if an algorithm is only fed Beyoncé songs and the output is a piece of music, it’s a robot. It clearly couldn’t have added anything to this, and there’s nothing original there.’”
I’m not so sure. It could turn out that the controversial “Blurred Lines” ruling laid the groundwork for litigating AI-mimicry.
From an insightful piece by Philip Cosores in Uproxx:
From track lengths to chord progressions to song structures, the amount of math involved in what sounds good to the ears is the least sexy aspect of music, right up there with the language of recording contracts and the cleanliness of tour buses. But it wasn’t until the rise of services like Spotify and Apple Music that the mathematics of music felt so dangerous. Namely, the math involved in streaming. […]
It’s been music critics who have been beating the drum about the dangers of streaming algorithms lately … but most of the time the criticism is less about well-researched investigations and more about gut feeling call outs, directed at music that is often simultaneously commercially successful and critically derided. Over the course of the last year, you’d be hardpressed to find a negative album review that didn’t at some point evoke the idea of The Algorithm being to blame for the music’s perceived lack of quality — it has become this specter hovering above popular music, ready to sink its talons into anything that finds commercial success. […]
Of course, the music world has changed because of streaming, and many artists and labels will always look to trends when creating their own strategies and aesthetics. But blaming streaming for the music that you don’t like feels increasingly closed off from reality, where streaming is, in fact, influencing most of the music that is being consumed, regardless of quality. This is no better or worse than it has ever been, it’s just a recent mode of consumption that musicians are learning how to work with.
It’s impossible to argue that in the history of commercial music — even before recording technology — there was a time when the means of delivery wasn’t an influence on songcraft. Whether it’s writing an opera with
Of course, there are artists creating music specifically to exploit Spotify as a platform — the ‘poop song’ guy immediately comes to mind — but I agree with the thesis of this piece. It’s easy to accuse music we don’t like of solely catering to ‘the algorithm’ just as we once derided songs made specifically for pop radio or albums in the ‘70s that seemed so serendipitous they were obviously capitalizing on a trend.
The favored target of the music critic is ever-changing (and I love music criticism and feel it’s necessary, so don’t take this as a slam). The identity of that target is a gauge of where music stands and the ways we, as music fans, feel uncertain in its progress. Emerging trends create a widening feedback loop, making it increasingly difficult for the critic to separate the calculated from the cultural. Yesterday’s disparaged made-for-MTV band is today’s algorithm-friendly artist. And, soon enough, probably tomorrow’s A.I. assisted songwriter.
Spotify … announced a major change to how its playlists will operate, with the news that some of its previously human-curated [editorial] playlists will now be personalized based on listeners’ tastes. […] “Some playlists will now be personalized for each listener based on their particular taste. This means that for those specific playlists, no two will be the same,” the company shared in a blog post.
Spotify says it decided to make this change after finding that users listened longer to the personalized playlists, during a trial of the new system. It also notes that the new system will increase the number of artists featured on playlists by 30 percent and the number of songs listened to by 35 percent — metrics that artists will surely like.
I’m a huge fan of increasing discovery opportunities, so I welcome and am intrigued by this news. Though I wonder if Spotify’s creating a musical version of the ‘Facebook bubble,’ where listeners with narrow tastes don’t get introduced to artists outside of their established spectrum. The algorithmic playlist change could be beneficial for new artists among listeners with an already broad predilection and great for classic, already well-known catalog artists with everyone else.
Matty Karas isn’t having it. He wrote this rebuttal in the 3/28/19 edition of the Music
But sometimes I really, really don’t want personalization. Like when I decide to click on my preferred ANTI. I don’t want to hear the re-sequenced version of the album that Spotify thinks would be best for me, and I don’t want to start on track 2, no matter how great track 2 is. If I wanted that, I would’ve clicked directly on it. I want to hear the album Rihanna actually made, sequenced and mastered. That was the intention of my click. Likewise any of the playlists that I follow. I follow them because I like the music and the flow and/or I like and trust the curator. The unspoken agreement between me and them is they’ll put thought and effort into the playlist and I’ll listen. Period. […]
Labels love this, I’m told, because it’s a way to get more tracks and therefore more labels on any given playlist. But who wants that kind of democracy? I want the four most interesting, pertinent, appropriate tracks you’ve got, not one from each of the three major labels and one from a token indie. Does anybody not want that? […] I want my curators to lead. If they’re just passively following me, why exactly am I following them?
Luckily there are many curated third-party playlists out there, but those are for the ‘broad’ listeners mentioned above. Maybe we’re selfishly expecting the majority of listeners — the ones who, in the past, mainly listened to music via commercial radio on car commutes — to explore and embrace new artists. Spotify’s giving the majority of its users what they want (and I won’t lie —algorithmic playlists are fascinating and fun) while the rest of us can dig into curated niche selections like this. Or this.
Elephant, get into that room. Let’s talk about Apple Music. The industry is expecting the company to copy Spotify and start introducing their own sophisticated algorithmic playlists. However, I’d like to see them lean into curation. Apple Music has flirted with playlists compiled by influencers and other notables, but they are hardly visible — the ones that exist are sort of difficult to find. If Apple can get Oprah and Spielberg on stage to promote its TV offering, then why not enlist playlists from heavy hitters? And I’m not talking lazy extensions of Beats 1 shows. Perhaps Frank Ocean’s ‘Songs I Listened To Growing Up.’
Let Spotify have the algorithms. Apple probably won’t be able to catch up anyway. Apple Music already subtly differentiates itself by being friendlier to the album format — they should go all in on the taste-making curator as well.