April 20, 2017

Training Your Replacement

Machine Readable Signage, 2017

Hello, Justin Manor here. I am one of the founding partners of Sosolimited, and I manage the San Diego studio and act as the Chief Innovation Officer. 

April 8th marked the opening of my first solo art show, at Basile IE Gallery in San Diego. Many of us at Soso have put personal artwork into group shows, but I have never had the opportunity to fill a whole room with my own pieces. My friends Paul & Chris manage the space, and they kindly asked last year if I would be interested in showing ‘my work’ there. I try to make a habit of agreeing to challenges that are far in the future, so that the difficulties involved are distant abstractions and easily discounted.

The gallery is located in the vibrant Barrio Logan neighborhood of San Diego. The artist showing in the adjacent gallery, Guillermo Valenzuela, invited traditional Mexican folk dancers to perform at his opening. Down the block at the VFW there was a Lucha Libre match the same night.

I am fascinated by the concept of machines and algorithms taking work away from people. We appreciate being relieved of boring tasks by handy gadgets and apps, but lots of people lose their jobs to advancing technology as well. Hence the name of the show “Training Your Replacement”, an exploration of our complicated relationship with technology.

Blasts from the Past

I included pre-existing work that I had created in the show, all the way back to my grad school days at the MIT Media Lab, in John Maeda’s Aesthetics + Computation Group. Using computer vision and GPU programming, I created a system for performing and recording 3D manipulations of HD footage. This print is a single frame of video captured from the system. The early video transformation systems I built were precursors to ‘filters’, like those found in apps such as Snapchat and Instagram.

Cinema Fabriqué, 2003

The image below was generated by a filter that replaced video pixels with TV logos. The input video was ESPN, and a sportscaster in grey coat with a red tie can be seen if you squint.

Station Pigment, 2005

There is one purely non-digital piece in the show, made out of sliced up skateboard decks, which have been arranged roughly by hue.

Decktrum, 2014

How Do the Machines See Us?

Sorting things has always been an OCD passion of mine, and I have been looking for excuses to get into open source Machine Learning software. I was curious what advanced categorization algorithms would come up with when fed portraits of our federal government. Using Gene Kogan’s ofxTSNE plugin for openFrameworks, I fed in the official portrait images found here.  Examine how the algorithms respond to the boundaries of gender, race, age, and outfit. How does your conception of physical identity compare?

T-SNE Gerrymandering, 2017

Similarly, I wanted to explore the level of similarity and diversity of people who appear on TV. I recorded a month of cable TV, frequently switching stations and shows. All off the faces detected in the footage were again fed into the same algorithms, which clusters them into regions of similarity. Due to the wide variety of lighting conditions and backgrounds across shows, the final result has a lot of macro-scale texture.

No One is Ever Really Alone, 2017

Digital Neons 

Numbers are the basis of all communication within electrical circuits. The two neon pieces here are explorations of the expressive quality of simple digits. Segmented numerical displays like the first one in the video have been in use for over a century. They are an efficient solution for encoding numbers as electrical signals to create a visual and legible output. How expressive can a single digit be?

In the near future, we will be surrounded by autonomous machines that transport us, deliver our goods, and do the tasks we no longer see fit to perform. Those wishing to communicate with these new members of our environment in an inconspicuous way might use signs such as these.

To control the separate neon tubes in these artworks, I created on/off animations in Arduino, whos digital I/O pins were tied to solid state relays that were toggling regular outlets on and off. There is something a little sketchy about switching the inductive loads of the neon transformers, so in the future I will definitely use transformers that can take logic-level inputs on their own to switch on and off.


Spiral Arm  & Lesser of Two Infinities, 2017

Lightpost 2.0

To round out the show I built three lightboxes - a black box, a white disk, and a translucent cube each with roughly 50 LED pixels. I set a goal to make the simplest lighting pieces I could. At Soso we have a disorder which tends to maximize complexity in our projects (we are seeking treatment), so I personally challenged myself to make the three pieces in a short timeframe with minimal technological infrastructure. The animations are inspired by neuronal activity, galaxy evolution, and competitive natural forces.

Cascade, 2016

The show will be running until June 3rd, 2017. If you are in the neighborhood, please check it out. You can get the gallery hours here.