top of page

PAINT the creative potential of odd software


What is digital aesthetics today and do designers really have creative liberty when using the computer? Did programmers and creators of lazy and lame software such as Microsoft Paint ever think of their creative potential?

This project shows why Microsoft Paint is as valuable as AI-generated images - the latter demanding huge computing performance - and how both can be combined into an electric symbiosis of technoart, between techno humour and high-performance criticism, disputing today’s highly consuming, pricey and elitist design computation techniques. 


keywords: lazy software, machine learning, hallucinating digital painting, computational creativity, animals in art, aesthetics of technology, MS Paint as technocultural artifact, digital media.

Project Description

Social technology/software or software ethos

“The political ethos of the Net, its extreme libertarianism [...] whoever's the most technically able can do whatever they want. It's really not "everyone can do whatever they want"; it's that the more technically able you are, the more you should be able to do. And that's the way it is online to me. It is a kind of meritocracy in a very narrow sense.” (Ulman 1997)


The accessibility to design knowledge became a social issue when softwares could either be free, chargeable, or preinstalled. If Unreal was a preinstalled free software for all, would kids learn how to design games on the family’s computer out of boredom at a young age? I played Microsoft Pinball on my grandfather’s computer as a kid and it was a lot of fun. But now, Microsoft hasn’t included Space Cadet Pinball in any release since Windows XP with the change from 32-bit to 64-bit operating systems, and the internet community wasn't happy at all. What a loss. Thankfully, with a little internet search you can install Pinball on a newer Microsoft version again. The same happened to Microsoft Paint when in 2017 the company announced it would stop updating the painting software and that Paint 3D will be the new scribble application installed by default instead. Microsoft Paint would still be available on the Windows Store but it is unlikely that system administrators will allow employees to install apps from the Windows Store. How rude. Megan Saunders, Microsoft’s general manager of the 3D for Everyone initiative, said: “Today, we’ve seen an incredible outpouring of support and nostalgia around MS Paint. If there’s anything we learned, it’s that after 32 years, MS Paint has a lot of fans. It’s been amazing to see so much love for our trusty old app.”


In the ODD SOFTWARE series we use not the most capable software and make new media and the first of this series is using Microsoft Paint. Let’s unroll our motivations for this and read this with a little techno humour…

Aesthetics in the digital era and the automation of aesthetic decisions: to which degree?

What is digital aesthetics today? Today, the computer program often defines the aesthetic. A software wraps up a certain set of tools and the aesthetic outcome depends on them. Those tools are a result of a codification process that reduces the complexity of the task execution into a simple clickable button within the interface. This is what Andrew Witt calls “encapsulated knowledge” (Witt 2010). This reduces both the user’s design knowledge but also creative diversity within the same software. Yes, by looking at the images “generated” by Artificial Intelligence, or Houdini, Maya, Grasshopper 3D models and renders the software is recognizable!  Ellen Ullman, author of the cult classic “Close to the machine”  says that “imbedded in this technology is an implicit way of life. And that we programmers who are creating it are imbedding our way of being in it.”(Ulman 1997) You don't need to know exactly how the code behind the button works but you extract it in order to be creative. All softwares used in digital art and architecture like Houdini, Illustrator, Cinema 4D, Unreal, etc were developed to fulfill very specific tasks on a very narrow field of knowledge. Yet, only a tiny elite of knowledgeable artists can use them because you need time and skills to understand the interface and the program itself. 


In our opinion media is the output of software. It’s what the user was able to extract. Accept the central role of software! It is the software that decides on the form of the media output. Next.

Can we extract creative potential from a more basic software? This principle also works with easy and less elitist design software that is free, running on any modest machine, and approachable to non-technical people with a straightforward user interface - which means almost no skills needed beforehand. Even if you don't have the best graphics card, a pricey software subscription and a high-end fancy computer you can still use the “aesthetics-extraction-principle” to achieve a technology-driven aesthetic. This is why the artworks from the ODD SOFTWARE series have this tension between the layers with the AI-generated (high-end computation) art as foundation superposed with the Microsoft Paint (low-end computation) doodles on top. The peaceful multicolored doodles hide the performance needed to train the AI model that produces massive amounts of data or digital junk.


Overlay of AI art and Microsoft Paint doodles

The first layer in the paintings is from an AI algorithm such as Google's DeepDream or Pix2Pix that does heavy computing through a list of thousands of data inputs and morphs together what it learnt into a series of outputs. These foundational patterns made by ultra-performance intelligent algorithms are extremely confusing and nonsense in their original state. It is especially interesting in an aesthetic society (Manovich 2020) of today which is the practice of over-aestheticization of images, goods, interior design, services etc for the commercial functioning of a brand identity.  Here, we do not aspire for beautiful images nor aesthetic pleasure that fit into the global aesthetic culture but rather we are looking for a technological aesthetic - one that expresses differences, transparency, and layers. 

As a base we used for example DeepDream images from the internet which is successful in creating its own aesthetics. An aesthetics that is almost completely defined by the tool. New tool means New Aesthetics. Our selection of ODD SOFTWARES also have an embedded technological aesthetics. 

“These AI creations [DeepDream images] in my view are more successful than the automatically generated slideshows of user photos [...] in the case of DeepDream, AI can create artistically plausible images which do refer to the human world because we consider it “modern art” and expect big variability.” (Manovich 2018)

We selected FIVE images created by AI algorithms to produce the paintings that were reverse-engineered by redrawing them using Microsoft Paint as the second and top layer. As mentioned before we - as lazy as our software - took five AI images from the internet. This prevents designing the aesthetics before using Paint. The aesthetics should stay relatively bare and unmodified, that's why any artistic decisions are muted. This workflow with Paint as the second step in the creative process also decreases the aesthetic diversity of outputs (more on « aesthetic diversity » in Lev Manovich’s book AI Aesthetic, 2018) since the integrated tools decide on the outcome. Therefore the only rule is to just paint on top of the old image and try to match the colors with the original ones. Only Paint standard colors are allowed though and no fading brushes! We don't want to get sophisticated here but to suppress our intuitive designer-urge for complexity. 


There has been art using another very unlikely tool - Microsoft Excel - made by Tatsuo Horiuchi aka the Michelangelo of Microsoft Excel. He has been painting using this software usually used in informatics, management or business for numeric computation purposes. In his case, he used the Excel canvas and filled it with vector curves and color patches. Here, we used Paint’s canvas to draw on top of the AI-generated base art. The commands are very limited, but this is the point in this project: we decided to reduce the use of different parameters to a bare minimum in order to manifest the obvious creative restrictions of Paint. Let’s just agree that the carved stones developed by Homo Faber are as efficient today as Microsoft Paint.


In this case the process is divided into two layers but Harold Cohen, a pioneer of computer generated art from the 1970s with similar aesthetics has programmed an algorithm to draw “paintings” on its own every day and is known today as pioneering AI art. The computer trains itself and tells a machine to pick colorful brushes and paint, which is basically the same process shortened down to a single workflow that produces its very own aesthetics too. 

Paint is used to extract the main pattern of those AI images to draw a composition map according to their dynamics, eventually dissolving into small and large spots that cover the canvas or just stand out as details and outlines in the foreground. Made from hallucinating illustrations coated with some digital material paint, yet following the same colors and patterns, these works are overlays of both automated and lazy processes. True contrast is the opposition in the differentness of the same: red and green are identical in being colors. These true contrasting differences still return to unity because of their sameness - identical in being colors from a limited palette of Paint - forming harmony within the border. The internal contradiction lies between the layers of paint, between the AI jpg and the Paint doodle. We like pixelated outlines in primary colors! The doodles are subtle although expressive enough to indicate either furry friends, animals, abstract shapes, or whatever you see in them! 


And yes, like any other software, doodles made in Microsoft Paint are recognizable too. That’s because they’re basically factory-made from the same set of encapsulated tools. So why do we even do this artistic experiment when the aesthetics is already defined by the software? And why do we combine these two softwares? Probably because combining the two very different computational approaches, each with their own aesthetic embedded, could lead to a new aesthetic! Maybe that's what designers do right now. Maybe the work is more about curating the tools instead of actually using them perfectly. In a world of endless craving for experimentation, innovation, and crazier art than ever, designers actually tend to combine a few different softwares in order to differentiate their art from others. This project is a hybrid of two processes: mega computing and lame software. 



Andrew J.Witt. A Machine Epistemology. Candide No. 3. 2010

Ellen Ullman talks about what makes programmers tick. Salon. 1997  (last visited October 16, 2021)

Karl Rosenkranz. Aesthetics of Ugliness. Log 22. Written in 1853 and translated from German by Sarah Haubner. 2011

Lev Manovich. AI Aesthetics. Strelka Press. 2018

Lev Manovich. The Aesthetic Society. Data Publics, eds. Peter Mörtenboeck and Helge Mooshammer. Routledge. 2020

bottom of page