The X factor — human instinct, undefinable and unknowable — still matters in a world where people increasingly rely on machines to do their thinking for them.
A January 29 episode of the HBO Max show “The Last of Us” has sparked a resurgent interest in a song sung by Linda Ronstadt that was released 53 years ago. “Long, Long Time,” written by Gary White and released by Ronstadt in 1970, features prominently as the motif for a tender relationship between two survivors of an apocalypse, which forms the heart of “The Last of Us” Episode Three. The episode, also named “Long, Long Time,” has received strong praise from critics and audiences alike, in particular for the performances of Nick Offerman and Murray Bartlett, who play the two survivors who find love amid the ruins of a zombie apocalypse. The song is used in a pivotal scene between the two, and then plays over the episode’s closing credits. Within one hour of the episode being aired January 29, streams for “Long, Long Time” shot up 4900% according to Spotify.
It’s certainly not the first time or last time that a popular TV show has brought back a song into circulation. What I love about the story is how “Long, Long Time” was even used in the first place. As reported in Variety, executive producer and director Craig Mazin was looking for a song for the two characters to connect over — something that “needed to hit certain things about longing and aching and endlessly unrequited love.” He could have used an AI-powered assistant such as ChatGPT to do the research for him, but instead he asked a friend, Seth Rudetsky, who is the host of Sirius XM on Broadway. Rudetsky texted him the song name in a few seconds.
To be sure, algorithms that suggest music based on our listening habits — most famously — Spotify — work brilliantly to expand a consumer’s listening experience. Algorithms build off what they know about you based on the data (your song preferences) that you’ve accumulated. They’re less helpful in making less obvious connections — finding the wild cards in the deck. By contrast, for “The Last of Us,” two people through one text exchange came up with a less obvious suggestion to enrich an artistic expression in another medium, television, for a mass audience of viewers, each of whom bring their own tastes to the experience.
Out of curiosity, I asked ChatGPT to suggest a song using the same parameters that Craig Mazin used — something that would express longing, aching, and endlessly unrequited love. ChatGPT suggested Adele’s “Someone Like You” — an excellent suggestion, to be sure, but an obvious one. ChatGPT had no other suggestions.
So, how did two people know a song released more than 50 years ago would resonate for a mass audience whose music tastes were unknown to the show makers?
Well, they didn’t, really. They made a judgment call based on the X factor: their instincts. Specifically, their instincts that a song many people have forgotten about today would work well because it connected with universal human feelings such as longing and aching.
Music supervisors rely on their judgment all the time to choose music for shows, movies, and games. A more famous example than “The Last of Us” occurred in 2022, when Kate Bush’s 1985 song “Running up That Hill (A Deal with God)” became a viral sensation after being featured in Netflix “Stranger Things.” The song became the top-streamed song on Spotify and reached Number 8 on the Billboard U.S.-based Hot 100 — marking Kate Bush’s first-ever appearance in the Billboard Top 10. Music supervisor Nora Felder said she chose “Running up That Hill” because it resonated with the pain and loss afflicting one of the show’s young characters, Max (Sadie Sink), and “could be very special for its powerful melodic flow and very poignant themes.” She told The Washington Post, “I’ve always felt that this song was so timeless and deserves to be heard for years to come. I think it’s struck a chord for so many people because it really touches on the alienation and emotional struggle that so many of us go through at one point or another in life, especially as teenagers. Music gives us validation and strength, especially when we aren’t feeling supported or understood by others.”
Here again, one person relied on her instincts — her understanding of art and the way it connects to human truths — to choose a song. She did not get led around the nose by an algorithm. She trusted herself. The song not only blew up commercially, it also became a sensation on TikTok — nearly 40 years after its release.
When I asked ChatGPT to recommend a song that would “touch on the alienation and emotional struggle that so many of us go through at one point or another in life, especially as teenagers,” the assistant recommended REM’s “Everybody Hurts.” Not a bad choice. And this time around, ChatGPT recommended a selection older than 2011 (the song was released in 1992). But here again, the choice was a bit obvious — not the recommendation you don’t see coming as “Running up That Hill” is.
This, by the way, his how successful music supervisor Mary Ramos and legendary director Quentin Tarantino collaborate when they choose songs for soundtracks to Tarantino’s movies. Their process does not work as quickly as the text exchange between Rudetsky and Mazin, but it’s grounded in the human touch: specifically, playing music (on Tarantino’s record player) and finding the right song for a scene. In doing so, they both contribute their perspectives about the interplay between music, film, and their own lives — so appropriate for a deeply personal expression such as art. Their collaboration has resulted in landmark soundtracks such as “Once upon a Time in Hollywood” and “Kill Bill,” Volumes 1 and 2.
I am not suggesting that people in the creative field rely on instinct alone. Rather, instinct and technology can work together. Lianna Kissinger Virizlay, an executive who manages a content team for an agency, states the case for how generative AI tools such as ChatGPT can assist people in content creation steps such as ideation without replacing human judgment. Many other creatives have demonstrated as well how generative AI in particular can play a valuable role.
To make the breakthroughs that no one sees coming, though, you’ll need to lean into instinct.