Vic 

Music and AI 356

Etude One: ChAI poetry!



This assignment was to create two poems using ChAI! 

Prompt and details listed on this site: ( https://ccrma.stanford.edu/wiki/356-winter-2023/etude1 )

 

My Poems Code FIles and Video Links:

Project one “Roses”:

 

Project two “Song Madlibs”:






Reflection:

This project was really fun! It was not easy, but it was interesting and had a lot of moments that were surprising and delightful. Of course there were obstacles but not nearly as many as I faced in 256a. This was the first time I have every used ChatGPT for more than a few minutes and its safe to say it became my assistant for this assignment. I had no idea it could be so helpful! It didn’t always have all the answers but the easy completion of the grunt work alone was incredible. I used it for turning sentences into arrays, asked for its advice on grammar, music information (midi note values), chuck code, and even the rhyming aspects from time to time. The most visible use was the outline for my madlibs song project, which was created in a matter of minutes. The rest was all me, but the outline was really helpful, and really interesting. It allowed me to enjoy the randomness of my words a lot more since that meant all of the lyrics had some kind of machine make it. 

 

My first project, “Roses” is a play on the well known poem roses are red, to see what kind of functionality I could get, and to see how I could manipulate something familiar to be new or more interesting. I particularly thought the section on when roses aren’t red was my favourite part there, since it seemed to spiral into such darkness, like the idea that something that is so well established being gone destroys even the machines idea of joy, almost like it became the person who said the poem in the first place. 

My second project, “Song Madlibs” was my more focused and entertaining project. There are many more features I wanted to add, but I ran out of time. The biggest entertaining aspect is the madlibs part, which is seeing how already random words, could become even more confusing. Madlibs typically has an order, or an expectation, you know the words coming because you provided them even though you don’t know the context. But here there is no context, and no expectation. You only have the beginner words and the hope there will be one you recognize. Trying to chase the patterns in the sensical non-sense makes it equally entertaining and thought provoking to me. 

 

Some credits should be given to chuck example code, as well as ChatGPT( for the reasons listed in my reflection).