The image is a machine.

2019

 

The image is a machine. The image drives the machines that produce the image´.

The image may be more of a complex ecosystem, more like a pond than a mechanism. The image may be mechanical, electrical, chemical, biological.

The image may be automated or autopoietic. The image´ is latent in the instructions for its performance.

The image is not reducible to source code. The image has the potential for variability of expression.

The image´ need not be built. Non-expression is a potential expression.

The image is a database. It takes in information and spits out electromagnetic waves.

The image may or may not be visual. The image´ may be instantiated in other forms, such as sound. Sound-image´.

Too much concern is lavished on the image´, the face. Of greater consequence is the commodification of the image, human capital.

The image forgoes scarcity in favour of the fecund circulation of images´. The image may populate the world with innumerable images´.

The image´ has no inherent value. The image is a producer of value.

The image´ is derived from the distillation of societal value systems. The image is fat with the intellectual, creative and labour value it has consumed.

The image may or may not look back. The image may be used as a mode of interpretation of images´.

The image constrains what images´ may be produced from it. The design of images directly conditions the production of images´.

The image is concerned with method. The image is processual, procedural, a practice.

The image is a machine.

 

Something of a manifesto of the algorithmic image, The image is a machine gives an explorative overview of the major themes of my PhD research. The work does not attempt to deliver a final word on defining the attributes of images. Instead it seeks to grasp the significant characteristics of ML-generated images, poetically, allowing ambiguities in its distinctions. This text-based image is roughly divided in pairs, which complement or correspond to one another. The prime symbol ( ́ ) is used to denote instantiations of algorithmically-informed images, in comparison with generalisations about images. Rather than being opposites, distinguishing between the image ́ and the image explores how algorithmic qualities of current media may fit within or relate to existing notions of the image. Defining the image in this way helps us better understand certain aspects at work there- in, but it also underscores the murkiness of this investigation. While it may elucidate some attributes, it also appears to obscure others in the process, which speaks to the nebulousness of the topic. This piece provides several different entryways from which to proceed onward.

Leave a Reply