A First Message from The TeaSoft Cat

I guess someone out there might be thinking, “It’s just an AI song, and it’s not that good,” or something like that. 

Yeah, that’s true. I do try to choose carefully. But to be honest, whenever the results come out, there’s always something missing or something that feels like too much. Having Suno generate a song feels a little like a gacha pull: what comes out never quite matches what I have in mind.

So all I can do is keep generating and adjusting the prompt and settings as I go, then select whichever one comes closest to what I have in mind, so that as much of the feeling I want to convey comes through as possible. 

I’ll admit that with some songs, I generated a lot of versions. Even then, I ended up with one that was just good enough. The downside is that I can’t tell the AI exactly what to do at every point the way you can with a person. I can’t simply say, “Sing this part with this feeling.” Maybe there is a way to do that, but my skills are not at that level yet.

If I ever get the chance, I’d love to see the lyrics I wrote sung by a person. I think it would be a lot easier to explain the feeling I’m going for to someone than to keep trying to get the AI to do what I want. 

So if you say my songs don’t hold up compared with human-made work, yes, that’s true—I agree. But I appreciate you listening all the way through. I guess you must have, because otherwise, how would you know whether they’re good or not? 

I understand if you criticize my songs and reject them because they were made with AI, especially after you’ve listened to them.                                                          

But please understand: I don’t leave all the work to AI, and I don’t do sloppy work like that. I write the lyrics myself. Let’s just say I squeezed the blood straight from my heart and used it as ink to write them—and my podcast scripts too. Haha. 

I stay as involved as I can in shaping the whole project’s direction, and I care a lot about the overall feeling I create in my work. So none of this is sloppy work. 

If you still think it isn’t good, that isn’t because I was sloppy with my work. It’s because my skills aren’t strong enough yet. I’ll keep learning and improving over time.

I like music, and I use it to heal my heart as well. But I handle all of this on my own. I don’t have a big enough budget to bring anyone else onto the project, and I’m not in a position to sing the songs myself. Even though I use AI to generate the music and vocals, I still try to put as much of my own heart into what I create as one person possibly can. 

Anyway, you don’t have to believe everything I say. We can see how things go. 

If AI-generated work isn’t your thing, you can still enjoy the project’s cat mascot. I commissioned a human artist named Th3taSnape to create most of the project’s artwork. If you like the cat mascot, you can follow the artist and see more of their work on Instagram: 

Or you can leave this page for now and come back later if we ever have songs with human vocals. I hope we’ll meet here again. And if you have any suggestions for improving how I generate the music—or any other thoughts—you can send them to me at: 

feedback@thesanctuaryoffragility.com