Semester 2, Week 7
Finishing Dissertation, Pixel Oracle 2
Finishing up that dissertation...
(The above picture is a Real documentation picture of my writing process)
This week started with me taking full advantage of the last minute magic for my dissertation. To be
honest, I wasn't able to think about the making as much as I wanted to, as most of the work went into my
dissertation .
Pixel Oracle 2
I spent a little time figuring out how I could make my pixel oracle into a physical installation. I
thought about the chart I drew
earlier and thought it would paint an absurd picture to have a machine making meaning out of
randomness that it created itself. Something about that was quite beautiful to me – a machine trained
and forced to make sense of its inherent chaos. So I sketched out a setup that could work.
1. Creating the Live Generator
I pulled out p5.js and modified the generator I created for the noise images, and made it a live feed that constantly generates noise at random.
2. Creating the Live Oracle
I then created a live camera oracle that assigned the "yes, no, maybe"s live based on the model that I had previously trained before.
Enacting the Live Oracle
I used my iPad and iPhone for both screens and placed them beside each other.
At this point it seemed like the oracle was not exactly equally predicting Yes, No and Maybe — which brought me back to my dilemma when I first created it. Is the point really the randomess? Should I aim for equal representation of yes-no-maybe to show its true randomness? Or just embrace the fact that randomness does not exist? Should I leave these questions for open interpretation of the audience? Is it the form of it that makes the impact less?