I revived an 1820s sea shanty with AI, and it’s a banger

My kids have been really into sea shanties lately (my family has eclectic musical tastes.)

There are a surprisingly large number of modern shanties on YouTube and TikTok. But one historic song, The Wellermen, really spoke to me.

Going down a rabbit hole of the song’s history, I learned that it was written in 1966 by a New Zealander. But the whaling classic was inspired by a much older song from 1820.

Eventually, I found the lyrics to the original. But there was a problem–the words were cryptic and the melody was lost to the sands of time, making it impossible to sing.

So, I decided to leverage today’s most powerful music-generating AI to bring it back. The result is a modern shanty that draws word for word on the 19th century original.

Spoiler alert: it’s a banger.

Here’s how I made it–and what I learned about the future of AI music.

The rise of ShantyTok

During the pandemic, sea shanties had an odd cultural moment. 

The trend was known as ShantyTok. Modern creators discovered centuries-old shanties, and started adapting them for young, streaming audiences.

Shanties work surprisingly well on social media. They’re often simple, repetitive songs, designed to be sung communally. 

They’re dramatic. And they’re highly story-driven, which encourages listeners to stick around and listen to the whole thing, rather than swiping away. TikTok and Youtube’s algorithms love that kind of engagement.

With their messages of struggle and resilience, shanties were also perfect for the Covid-addled moment. The result is that Shantytok yielded really fantastic modern renditions of ancient classics.

Wellerman is a perfect example. Collected and put to music by the folk musician Neil Colquhoun in 1966, it was adapted by modern musician Nathan Evans and went viral on social media in 2021.

The song tells the story of a ship’s captain and his crew, locked in a mortal, never-ending battle with an elusive and powerful whale.

If that sounds a lot like the story of Moby Dick, that’s no coincidence. The website New Zealand Folk Song has an excellent history of Wellermen. Colquhoun apparently based his modern version on a historical 1820 whaling shanty that his schoolteacher wife found in an old book. 

That shanty, titled Mocha Dick, is way darker than Wellermen–the entire crew dies in this one, instead of simply pursuing a whale for all eternity.

Mocha Dick is based on a real whale that reportedly drowned 100 men off the coast of Chile while evading capture, and the Smithsonian says that real-life whale inspired Melville’s iconic novel.

When I finally found the lyrics to Mocha Dick on New Zealand Folk Song’s site, I was initially excited, and then disappointed.

In its current form, the song is basically unsingable. The melody has been lost to history and the lyrics are tough to interpret–dark, lacking a discernible rhyme scheme, and filled with 19th century colloquialisms like “bully boys” and loads of references to very specific parts of whaling ships.

I wanted to bring it back to life and sing it with my kids. So, I turned to AI to see if I could revive the song–and make it as much of an earworm as Wellerman.

AI pirates

To bring the shanty back from the dead, I turned to Suno, the most powerful music-generating AI on the market, plunking down $10 for a month of Pro access.

I sing in a choir and can hold my own with a ukulele. But you can use Suno even if you have zero musical skill.

Suno’s system adapts to whatever inputs you provide. You can give it something as simple as a written concept (“death metal lullaby” or “acapella Python-themed polka”) and it will spit out a fully-produced song, complete with vocals, instrumentation and cover art.

But if you know more about music–or have source material to start from–the system is also happy to play Elton John to your Bernie Taupin, crafting music to match your lyrics or even remixing your original song.

In my case, I pasted the exact, 1820 lyrics of Mocha Dick into Suno’s interface. I then specified that I wanted a “rousing sea shanty.” In less than a minute, it had produced four different song variants.

Two were vocal-forward, Irish-inflected tunes that sounded fine, but not like anything special. I immediately fell in love with the third one, though.

After a lilting start with a single, gravely voice, the song launches into a lively ballad, complete with multi-part vocal harmonies–all imagined by Suno’s generative AI. My kids think it sounds like a choir of pirates. 

I named it William of Tyre after the doomed ship in the original song, spun up a graphic with Google’s Nano Banana, and uploaded it to YouTube.

Automated bangers

There’s a lot to like, musically, about Suno’s creation. After its quiet and subdued start, the song slowly swells in intensity–with the pirate choir providing subtle vocal backing–until the doomed crew confront the white whale.

At that climactic moment, the tempo suddenly quickens, echoing the speed and drama of the chase, before slowing and adopting a mournful tone as the fictional William of Tyre is ultimately sunk.

There’s an odd lyrical section at the end of the original shanty in which the whale begins speaking directly to the crew. Suno does a terrific job making musical sense of this, bringing in multiple voices singing in a lower register with an almost monastic tone to suggest the voice of a massive, seafaring leviathan.

Likewise–because the original shanty lacks the kind of catchy, repeating chorus you’d find in a modern song–Suno turns its single repeating line (“Blow my bully boys, blow”) into a powerful refrain that caps off each verse.

The original lyrics also lack a clear ending–the song just kind of stops, and there isn’t even a final “Blow my bully boys!” to see us off. 

Again, Suno handles this musical ambiguity surprisingly well, finishing the piece with a series of shouted “OY OYs!” and a climactic drum solo.

To be sure, there are problems. For no reason at all, Suno’s imagined singer pauses in the middle of the first “Bully boys” line. My best guess is that Suno saw a comma in the original lyrics, and interpreted it as a spot to randomly pause.

The AI also makes mistakes that a human would probably catch. There’s a line about the “bow” of the ship. Anyone with rudimentary nautical knowledge would know how to pronounce that word in a ship-y context.

But Suno makes it sound like the “bow” in “bow and arrow”. It’s a subtle mistake, but one that’s very telling of the song’s AI origins.

The future of music?

Despite its flaws, though, there’s a strange appeal to the song. I find myself listening to it again and again.

So, is this good for the world of music, or not?

Suno is already notorious in the music industry for flooding streaming platforms like Spotify with millions of songs that established musicians describe as “slop.” 

A coalition of these musicians have already launched a Say No to Suno campaign. And the RIAA has reportedly filed suit against the company over copyright allegations.

I’m sure a real historical musicologist would listen to my AI-generated shanty and conclude that there’s nothing whatsoever authentic about it. But after listening to the song a lot, I’d be leery of dismissing it too quickly. 

Sure, the original shanty probably sounded a lot different when it was sung in the 1820s. There was likely no choir of pirates with electric instruments providing dramatic musical backing.

Still, there’s something powerful about hearing the song’s original, historic lyrics set to music–even if that music is imagined by AI.

Many lines in Mocha Dick made no sense to me when I simply read the lyrics. Hearing them performed gives them an emotional power and resonance I never thought I’d find in the two-century-old original.

A line that reads “Come raise your hand/My bully boys/And swear you’ll not flinch or fear/While there’s a spar to keep afloat” doesn’t hit very hard when you first read it. 

But when it’s sung in a pleading yet resolute tone by overlapping, harmonizing voices (even imagined ones) you suddenly realize that the fictional crew knows the grim fate that awaits them as they confront the whale, yet chooses to proceed anyway. It completely changes the song.

That’s the real power of platforms like Suno. Creating AI-powered bangers is nice. But the ability to breathe new life into a long-forgotten set of lyrics–and to create an emotional bridge across centuries of time–is an impactful one.

Perhaps William of Tyre isn’t historically accurate or authentic to the original. But all music has a unique and magical ability to stir the soul, lending resonance and power to our frail, human words and struggles. 

As I learned from my experiment, that applies even if it’s written by a computer.

View the full article