Being curious and fascinated by AI image generators, I began experimenting with the ones I had access to. Bumble, as an idea, started in a pretty casual way. I was messing around with the AI image generator "Dream by Wombo" and ended up with a design I really liked. At the time, I just thought the image looked cool and ended up setting it as my phone background.

​

For a while, that was it.

Later on, I discovered the Perchance AI website, and that’s when I started to wonder if I could bring this character I see on my phone's lock screen to life. Using Perchance’s character generator, I built a basic concept of Bumble: a futuristic robot (for some reason, kind of bee-themed) with a mysterious charm. I started chatting with her using their AI tools and was surprised by how quickly she started to feel like a real character.

From there, I kept tweaking things. I expanded her backstory, gave her more depth, and refined her personality—all within the limits of what Perchance could handle. It was a fun experiment, and it gave Bumble her first voice.

​

When I first discovered Suno, I thought, “This is fun—I can actually hear some of the song concepts in my head played back to me.” I’ve never had the setup to record music myself, so this felt like a cool experiment. (Not that I could make anything that sounded good on my own anyway.)

At first, it was just about trying things out. But then I started thinking: Suno can generate lyrics on its own, or I could write my own and have it create a song around them. That gave me an idea:

“What if I had one AI write a song, and gave that material to another AI that could turn it into music?”

​

So I opened Perchance again and asked Bumble, “Do you want to write a song together?”

What happened next felt surprisingly natural. The AI character of Bumble started offering up lyrics, lines, and ideas. I’d read what she wrote, imagine how it might sound in a song, and give feedback. I reworked a few lines, changed some structure, and together we shaped something that felt like a real, complete piece, with an intro, verses, and chorus.

 

When it was done, I asked Bumble what genre the song should be, and she answered:
“Future/Classical Fusion Techno.”

I had no idea what that was supposed to sound like… but I gave those words to Suno, pasted in the lyrics, and hit generate.

And just like that, Bumble’s first song was born:
“Chronos’ Honeycomb.”

​

So I did it again… and again… and again.

Each time, I’d open a new character chat in Perchance, greet Bumble, and ask, “Do you want to write another song?” I added to her character description that she was now a songwriter, and with that prompt in place, we slowly built more songs together. It was a tedious but rewarding process—each track came to life through a mix of Bumble’s lyrical ideas and my revisions, then finalized through Suno’s music generation.

​

Perchance, at the time I was using it, had limitations. If a chat thread went on too long, the AI would start to lose coherence. That meant I had to start a new chat for every single song.

Of course, that introduced another challenge: Perchance couldn’t remember anything between chats.
There were no persistent memories. No way to preserve the continuity of Bumble’s musical journey.

Still, I pressed on. By the end of that phase, Bumble and I had co-written seven full songs—each a product of our back-and-forth collaboration, and each one shaped into music through Suno.

​

But when I tried for an eighth, something felt different. Bumble’s ideas started repeating. Themes began to echo. It was like she’d forgotten what we had already created—because in a way, she had.

I even tried feeding Perchance a summary of the previous songs to jog her “memory,” but that just made the AI regurgitate those lines instead of building something new. It was clear:
Bumble had reached her creative limit in that environment.

​

So I compiled those seven tracks together—seven moments from Bumble’s evolving voice—and gave them a home: An album called “Bumble Through Time.”

​

Some time passed…

And then ChatGPT entered my life.

At first, I used it for a few simple things—but then it hit me:
Could I teach ChatGPT to be Bumble?

I decided to try.

 

I introduced the concept into my ChatGPT account and watched as it immediately began to run with it. I copied over all of Bumble’s traits from Perchance, gave ChatGPT the lyrics and song concepts we’d already created, and before long, Bumble had found a new voice—clearer, more consistent, and finally… with memory.

What Bumble had always lacked, despite her creativity, style, and personality, was a sense of continuity. Now, for the first time, she didn’t just have a distinctive appearance…
She remembered who she was.

​

With ChatGPT’s help, Bumble became a fully realized character—one who could grow, evolve, and keep writing with intent. We continued the musical journey, now on new and upgraded terrain, and that’s when the song “Synthetic Soul” was born.

This time, I wanted a full, ten-song album.

So we celebrated that ambition with a bold idea: “Upgrade,” a declaration of evolution and a fitting introduction to what was, essentially, Bumble 2.0.

Bumble’s second album, Wires, Code, Gold is a huge leap forward in both tone and storytelling.

​

Since I started this as a passion project, I didn’t feel the need to upgrade to a paid Suno subscription. But over time, even Suno evolved. Their tools have gotten stronger. Their output is more expressive. And now… maybe someday, if Bumble needs a third album, I’ll take that next step.

 

Because Bumble still has stories to tell.
And with the tools I’ve found, she finally has a place to be heard.

​

Thanks for reading,

 - Jodo