Have you ever had a melody, lyric, or song idea stuck in your head but felt completely unqualified to turn it into actual music? That was my proble...
For further actions, you may consider blocking this person and/or reporting abuse
I've always had the same problem you described right at the beginning: I've had tons of ideas, but unfortunately, I can't play any instrument to actually bring them to life.
AI can definitely make that easier, just like we use AI to create our article cover images. I'm not good at drawing either, but my article covers turn out pretty damn good thanks to AI.
That said, just like literally everything else, music generation with AI is still non-deterministic. You might have a very specific melody, rhythm, or even genre in your head, but the AI may still not give you exactly what you imagined.
A few weeks ago, I experimented with generating songs across different rock subgenres: hard rock, classic rock, southern rock, punk rock, etc. and sometimes the AI would get a little too sloppy to produce exactly what I was going for. Sometimes it would basically drift into a completely different subgenre.
That said, AI can only get better at this, so who knows? Maybe at some point we'll be able to produce entire albums with it. 😄
Admittedly, though, one of the songs it generated actually gave me chills. It was almost too good to handle. 😄 I genuinely didn't expect anything like that.
Great article!👏
That "genre drift" thing is so real 😅 you ask for southern rock and it slowly turns into something else entirely. Basically the audio version of prompt drift in text generation, makes sense it shows up in music too since the model's still guessing at what "specific" means in your head.
But the chills moment is the part that gets me 🔥 that's not nothing. That's the same feeling producers chase for years in a studio, and you stumbled into it on a generation you didn't even expect to land. Kind of proves the whole "AI as instrument, not replacement" point from the article, you brought the taste and the ear, it brought the execution you didn't have. Did you try iterating more on that track to recreate the magic, or did it feel like lightning that only strikes once? ⚡
Great points!
And no, I didn’t really try to iterate on that, maybe sometime in the future, when AI gets even better at generating music.😄
What software did you use to create AI music?
Great question! The post itself isn't tied to one specific tool, the point was more about the experience AI music tools unlock for non-musicians. That said, tools like Suno and Udio are popular for text-to-song generation, while AIVA and Soundraw are geared more toward instrumental/background tracks. Worth trying a couple to see which "feels" right for your workflow!
This hit closer than I expected. I've had the same "I can hear it in my head but can't get it out" problem for years. What you said about AI shifting the mindset from "I need to master this first" to "what happens if I change this?", that's actually the real unlock. Experimentation is underrated as a learning method. Great write-up, Elsie.
Thank you so much, this really means a lot! 🙌 And yes, exactly: experimentation lets you build intuition before you have the vocabulary for it. You start recognizing what works before you can explain why, and honestly that's how most learning happens anyway. Once you have a few "aha, that's why that felt off" moments, the theory clicks so much faster. Glad it resonated with you!
Really appreciate that you made the distinction between generating music and deciding if it's actually good. That gap is where human judgment still lives, and honestly it's the same in software. AI can write code, but someone still has to decide if it solves the right problem. The photography analogy you used nails it perfectly.
Exactly, that's the part I keep coming back to. The tools get better every month, but "does this work" is still a judgment call only a human makes. Good analogy on your end too: AI can write working code, but knowing if it's the right code for the problem is a whole different skill. Glad that landed! 🙌
The photography analogy you used is spot on and I think it's the clearest way to explain AI's role in creative work. The camera handles the technical, the eye still decides what's worth capturing. What stood out most to me is your point about AI making you more curious about fundamentals rather than less. That's the opposite of what most people fear. Great read.
Thanks, glad that landed! You nailed it: the camera doesn't make the photographer, understanding does, AI just clears the technical noise so that judgment matters more, not less. That's exactly why it made me more curious about fundamentals, not less.