I chose Melody’s task 11 post as my fourth linking assignment. The prompt Melody chose for the predictive text is “This is not my idea of…”. She used Instagram as the platform to create her predictive text. Melody has not used the Instagram account for a year or possible two. This is where Melody and I have a difference in the experiences we had.
The initial difference between melody and I is the prompt we chose. Melody’s prompt is very open and could have gone in multiple directions. When referring to what is not someone’s idea of something, this could go to not their idea of fun, of work, of a good time, etc. The prompt I chose is “Education is not about…”. This prompt limited what my algorithm could have predicted and generally what I could have chosen. Even though our prompts differed and could have gone in very different directions, Melody’s algorithm actually brought her towards education as well. Melody’s algorithm led her to text going back to school, after commenting on the weather. Mine went towards money in education.
The next difference was the applications we used. I used the Notes application on my phone rather than Instagram. I am not sure if the application used will make a difference in the information that the algorithm has to use. The use of Instagram as opposed to the Notes app will then become a factor. For example, if the application does matter, the texts predicted by the algorithm for Melody’s Instagram account will relate to what she has previously posted. Typically, people are very selective with what they post on their social media accounts. My use of the notes app is more liberal than her use of Instagram, since the Notes app is just for personal use. This could effect the nature of the texts predicted.
Since Melody has not used her Instagram account for a long time, her algorithm was going off of old text information. My Notes app is something I use every few days, so it is constantly updated for the text I frequently enter. It would have been interesting to see if Melody compared the algorithms of this less often used app with a feature on her phone she uses more often. It would show if there has been a change in what she frequently texts. In general, that would be an interesting element to add to the task, seeing how the algorithm naturally changes along with changes in text.
One similarity Melody and I found was that the algorithm did well to sound like each of us, but was very flat. The voice the algorithm presents is very basic whereas if Melody and I could have chosen what we were typing, it would have added more personality and context to the message.