Before Hemingway Hearts existed, Bryan Mabe was already writing songs. They just did not sound like songs yet.
For years, he wrote stories, poems and lyrics without formal musical training. Some ideas stayed as words on a page. Others came with melodies he could hear internally but did not have the production or performance skills to turn into finished recordings.
“I’ve been scribbling down my thoughts for as long as I can remember,” he tells Lunar Boom. “Everyone wants to share their songs, even if all you’ve got are the words.”
Generative music changed what he could do with those ideas. But the more interesting part of his story is what happened afterwards.
AI did not make him less interested in learning music. It made him want to understand it more.
The song can exist before the musician can perform it
A common argument around generative music starts with substitution.
AI can generate vocals, so it could replace a singer. It can generate instrumentation, so it could replace musicians. As the technology improves, the concern is that more parts of music production can be automated.
That is one possible effect. Hemingway Hearts shows another.
Bryan already had something he wanted to express. What he lacked was the ability to perform and produce it at the level he wanted, along with the money required to hire people who could.
“In a perfect world where money wasn’t an issue, I’d either have the best music teachers give me lessons until I mastered it or I’d gather all my favorite musicians and pay them to reproduce my creations,” he says. “But I’m just a dad with a hobby.”
Suno gave him another route.
Bryan now writes the lyrics and develops the melodies behind Hemingway Hearts while relying heavily on generative tools for vocals and instrumentation. AI fills gaps in his ability to execute the music, rather than providing the original reason for making it.
This resembles something we have encountered repeatedly in our conversations with AI music creators. In AI Makes Music Easier to Produce. Building a Career Is Another Problem., imOliver described a world where generating music is becoming easier while taste and artistic judgment remain scarce.
Hemingway Hearts approaches the same shift from a different direction. Before worrying about building a career, some creators first need a way to hear the song they already have in their head.
AI can lower the barrier before someone learns the craft
Traditional music education usually works in the opposite order.
Someone learns an instrument, develops their voice, studies production or spends years becoming comfortable with a digital audio workstation. Only then can they reproduce increasingly complicated versions of the music they imagine.
Generative tools can partially reverse that sequence.
A writer can hear their lyrics performed before they know how to produce the recording themselves. Someone with a melody can test arrangements before understanding every technical decision behind them. The finished result can arrive before all the skills traditionally required to produce it.
It would be easy to interpret that as a reason not to learn those skills.
Bryan experienced the opposite.
“Using AI to create songs has also motivated me to learn guitar, produce with DAWs, and understand music production more deeply.”
One artist does not establish a wider trend. But the possibility is worth considering.
If hearing your ideas become music makes the process more rewarding, generative tools could become an entry point into musicianship rather than simply a shortcut around it. Some people may be satisfied generating songs without learning anything further. Others may begin with generation and then want greater control over what they are hearing.
For that second group, AI could make music education feel more relevant because there is already something tangible to improve.
The human part can begin long before generation
Hemingway Hearts also complicates another assumption about AI music.
Bryan does not use AI to decide what his songs should be about. In fact, he draws a clear boundary around one part of the process.
“I never really considered letting AI write my lyrics,” he says. “To me, the lyrics are the most human part of a song.”
That is his own creative rule, not a universal definition of authorship. Other artists will draw the boundary somewhere else.
But in Bryan's case, the distinction is closely connected to what he writes about.
Much of Hemingway Hearts deals with depression, difficult experiences and the effort to find hope alongside them. Bryan describes writing as a way of processing events that might otherwise remain buried.
“As a young kid, I witnessed a few traumatic events that gave me a glimpse into the darkness of the world before I ever should have been exposed,” he says. “But it made me appreciate that life has darkness we can’t escape. We just have to navigate it.”
The story usually comes before the sound.
Bryan might develop an idea while walking or listening to music, then save a line in his phone before it disappears. Sometimes a chorus arrives with a melody already attached. Only later does the production take shape.
This creates an important separation between the origin of a song and the means used to perform it.
The voice may be generated. The instruments may be generated. The experience being expressed through them is not.
There is more than one way to divide human and machine work
At Lunar Boom, we have increasingly found the term “AI music” too broad to describe these differences.
As we argued in AI Music Is a Spectrum. Our Labels Should Reflect That, two artists can both use generative AI while dividing the creative process between human and machine in completely different ways.
Hemingway Hearts is a useful example.
Bryan writes his lyrics and develops the underlying ideas and melodies. Generative systems provide capabilities he does not have himself, particularly around performance and instrumentation. He then makes decisions about which results belong to the project.
Another artist might perform everything but use AI to write a hook. Another could generate lyrics but record the vocals traditionally. A producer might use AI only for one instrument, mixing task or production layer.
Calling all of them AI artists tells us that a technology was involved. It does not tell us much about what the person actually did.
Bryan puts the problem more simply.
“I think we first need to distinguish how AI is used and be open and honest about it.”
Easier execution puts more pressure on judgment
AI has also given Hemingway Hearts considerable freedom over genre.
Bryan cites influences ranging from Johnny Cash to The Killers, SYML and Jon Bellion, but says he rarely begins a song by trying to reproduce a particular artist. Sometimes he does not even know what genre the finished track will occupy.
Instead, the writing provides continuity.
“I’d say I take Cash’s simplicity to heart,” he says. “I’m not one who thinks lyrics need to be Shakespearean, have complicated metaphors, or that the instrumentals need to be super unique and experimental.”
If something feels forced, he is willing to abandon it.
This is another theme that has appeared across our AI artist interviews. When Brian Myk spoke to us for What Skills Should Artists Value in the Age of AI Music?, he argued that the production bottleneck was moving away from logistics and toward judgment.
Generative tools can provide more possibilities. They cannot remove the need to choose between them.
For someone like Bryan, that means AI can offer different singers, instruments and production directions without deciding which version still feels like Hemingway Hearts.
AI could expand who gets to participate
There are legitimate concerns about what cheaper synthetic performance means for professional musicians, vocalists and producers. Hemingway Hearts does not resolve those questions.
But it does show why substitution is not the only useful way to think about generative music.
Some AI-created recordings will replace work that might previously have gone to human performers. Other recordings may exist because the person behind them was never realistically going to hire a studio, session musicians and a professional vocalist in the first place.
Bryan describes himself as a dad with a hobby. Without tools like Suno, many of his old lyrics might simply have remained notes.
Instead, they became songs his children now sing around the house.
That is a different kind of economic and creative effect. AI is not only changing how existing musicians produce music. It is reducing the amount of technical ability, money and infrastructure someone needs before they can participate at all.
The result could simply be much more music.
It could also produce more people who begin thinking of themselves as musicians.
The tool can finish an idea without becoming the idea
Bryan expects the boundary between conventional production and AI production to become harder to see.
“I think we’re already there,” he says, pointing to the growing use of AI tools throughout production. His view is that transparency should focus less on whether AI appeared somewhere in the workflow and more on what role it actually played.
His own position is clear.
“AI is a tool, not an artist. Generative AI should complete ideas, not replace them.”
Not every AI creator will agree with where Bryan draws that line. Some will collaborate with AI on lyrics. Others will deliberately begin without a human-written song at all.
The wider point is that access to generation does not dictate what someone chooses to do with it.
For Hemingway Hearts, AI provided a way to turn years of writing into recordings. Once that became possible, Bryan did not conclude that learning music was unnecessary. He started learning guitar and production because hearing his ideas made him want more control over them.
That may become one of the less obvious effects of generative music.
AI will almost certainly make it easier to produce songs without traditional musical skills. For some people, that will be the end of the process.
For others, it might be where becoming a musician begins.
