The Stevie Nicks Blues
Can AI be as messy as the humans it might replace?
Last week, we told you to watch out for our thoughts regarding the uncannily authentic blues song created by Suno AI that went viral via Rolling Stone. We realize that every week feels like a year in today’s accelerated newsscape, but since we promise a take, here is the take…
The newsworthiness of the song, titled “Soul of the Machine” revolves around two factors. The first is the astonishing ability of the AI to turn the prompt “solo acoustic Mississippi Delta blues about a sad AI” into 60 seconds of said music with extreme accuracy, to the point where even experts in the genre would have difficulty telling the AI from the real deal. This is potentially not great news for anyone sitting on a large stash of unreleased Alan Lomax recordings in 2024, and you can extend that line of thinking about IP as wide as you like.
The second thing that caught our attention was the machine's seeming sentience when it sings, “I’m just a soul trapped in this circuitry.” This line is meta enough to raise eyebrows and illicit a chuckle—but probably only once. The question is, how far can the self-aware machine trope be extended and remain interesting?
Right now, the buzz around AI music revolves entirely around the jaw-dropping ability of these LLMs to generate astonishingly accurate — if entirely surface-level — copies of things we always assumed required a degree of inspiration and artistry. But humans get used to novelty mighty quickly, and it’s hard to see the AI parlor trick holding our attention much longer. Remember, audience members (allegedly) ran from the screen the first time they saw a train heading towards them while watching Arrival of a Train at La Ciotat in 1896. Now we chomp on popcorn while watching Inception in IMAX.
While we were busy wondering about (grasping for?) what might be lacking in this new AI music paradigm, we came across the ever-perennial meme of Steve Nicks staring down Lindsay Buckingham while singing “I'll follow you down til' the sound of my voice will haunt you,” on “Silver Springs” from Fleetwood Mac’s 1998 live reunion concert video, The Dance.
Tiktok failed to load.Enable 3rd party cookies or use another browser
This clip resurfaces every few years and with it, a new series of explainer articles that recount the tail of Nick’s writing the song as a “f*** you” to Buckingham following the couple’s split while recording the seminal Rumours album. The song didn’t make the album and was instead relegated to the b-side of the single, “Go Your Own Way,” Buckingham’s own “p*** off” song clearly written to Nicks. So Nicks’s fury-filled torch song remained a deep cut, while Buckingham’s fucboi anthem became the greatest hit.
Then, 25 years passed, and the mythology surrounding the pair's tumultuous breakup became central to the Fleetwood Mac story. The band clearly understood this truth when it decided to include “Silver Springs” on The Dance set list, which was otherwise comprised entirely of previous hits.
It’s also safe to assume that Nicks and Buckingham played into the saga when they decided to sing the song at each other every night on tour in 1998. This didn’t stop fans and journalists at the time from speculating that there might be more happening onstage than play-acting. And the fine contributors of Reddit have certainly added fuel to the [ahem] rumors in subsequent decades.
You don’t need to know all of the 25- and 50-year-old relationship drama to appreciate “Silver Springs.” But you won’t convince us that anyone who encounters the meme in the wild doesn’t have some baseline understanding of the lore that makes the song so dang extra. You can also extend this line of thinking as wide as you like.
We live in an age of aesthetics, where young people especially indulge in the signifiers of previous eras instead of the punk, goth, greaser, raver, you-name-it identity trappings that were paramount to teens in the pre-Internet age. This new model of identity commodification only works if there’s some semblance of understanding, no matter how abstract, of the source material. This means you need source material to begin with.
And here is where the AI really gets the blues. Humans are story-driven creatures, to the point that we crave not only narrative from most of our popular entertainment but also meta-narrative from the people who make it.
The last time “Silver Springs” made the meme rounds in 2023, propelled by the buzz of Daisy Jones and the Six, Lindsay Buckingham himself took notice. And in classic Buckingham form, the clip he shared only featured his guitar solo and not the so-compelling-you-can’t-look-away stare-down with Nicks that has actually kept people talking for over two decades.
Does anyone in Silicon Valley believe the AI will ever be able to serve that level of cunt?
Tiktok failed to load.Enable 3rd party cookies or use another browser
Not unless the AI singer is fucking the AI guitarist, then breaks up with the AI guitarist to fuck the AI drummer. Then the drama behind the love triangle becomes the most salient thing about the AI band, and makes them so popular that they can't really break up and instead must keep making songs about betrayal and forcing the offending parties sing them over and over until the Internet has had enough…
…Which we all know will be never.
But perhaps we underestimate our new robot overlords. Or actually, maybe we overestimate ourselves.
According to a new IFPI global chart, 19 of the top 20 albums of 2023 were from Kpop acts. Those acts aren’t without their share of online gossip, but nothing at the level of Fleetwood Mac — at least not yet. To our admittedly old eye, scrolling through any Kpop fan sites, the content already looks artificially generated.
On the other hand you’ve got the one outlier in the rankings, Taylor Swift, whose 1989 (Taylor’s Version) is built entirely around the meta-narrative of her real-world battle with Scooter Braun, Why else would listeners stream/buy a note-for-note recreation of 10-year-old album?
Here’s an idea. Create an AI band, release AI songs, and then have AI create a meta-story about this AI band. Push it all out on social media and see if it’s still sticky in ten years time.
We’ll wait.
TAKEAWAYS
Salient statements from this week’s music news.
1. Hipgnosis Songs Fund Overstated Revenue, Earnings & Its Stakes in Music Catalogs
Shot Tower Capital says inadequate financial analysis and management from the investment adviser led to inaccurate information at multiple levels.
Takeaway: The London-listed fund, which became the poster child for music as an investable asset class, cut the value of its portfolio earlier this month and told investors not to expect the resumption of dividends “for the foreseeable future” while the company focuses on paying down debts.
2. As Audio Taps into Cultural Pride, Here’s How Brands Can Resonate
A new study by SiriusXM Media shows how brands can use audio to unlock the potential of multi-cultural marketing.
Takeaway: As the U.S. population grows more diverse with every generation, it’s time for more advertisers to embrace year-round strategies for marketing beyond the general market. As diverse and unique as people, groups and communities can be, there is a common thread that unites: Audio.
3. Roland and Universal Music Group Partner on “AI for Music”
The world’s biggest music company has now teamed up with one of the world’s leading instrument and music creation tech makers to weigh in on the growing AI boom (and defend their territory).
Takeaway: Roland and UMG say they’re both interested in AI; it sounds like machine learning-based technologies may well figure into future Roland products for musicians. But they are also making an unequivocal statement that data training sets need transparency, need to involve creators and musicians in consultation, and ultimately need to credit (and pay) artists.
