We tossed around a few ideas when we first arrived. I considered making an application to help people organize group hiking trips. Dieter was not super interested in this idea. We considered going separate ways, but I decided I had more conviction to continue working with Dieter than I did to pursue this specific idea. We then found an idea that was at the intersection of our skill sets and interests: ai-generated anime.
The idea spawned because of the way that Dieter and I became friends 5 years ago. We became friends through recording a series inspired by Sword Art Online, my favorite show at the time. We made the series in our media tech class, and after our media tech class was cancelled due to covid, we worked together at our local library trying to make our own anime. We ultimately failed because we realized it was incredibly time-consuming to draw out each frame-by-frame. The idea for this project was that it would allow people like Dieter & I to make our own anime with the power of ai-generated anime images & voices.
Like most programmers, we've tinkered with the open ai api in past projects, but we wondered if we could do a self-hosted solution so we would have more control over the output and not have to pay for OpenAI. I'll get more into the MVP that we built in the next story.
The foundation of the project was that we found the tech super interesting. We loved working with Llama to try to get it to generate great stories & for it to make image prompts that result in great images. It was a really fun process, again I'll get more into it next story. I think it was at this time when we made the biggest mistake of our time in Japan (yes... at the foundation... a fatal mistake...). We focused on a new & interesting technology that we liked instead of focusing on a real problem that we can solve.