The Interview | Melissa Wiederrecht

Universally recognized as one of the pre-eminent generative artists working today, Melissa Wiederrecht’s practice long predates art on the blockchain. Having produced her first algorithms with pencil and paper, by the age of twelve she was translating these projects into generative art using digital tools. In the decades since, she has refined her style into a recognizable yet adaptable signature whose range is evident in the different collections she has already dropped this summer on fxhash, Feral File, and Highlight.

Ahead of her participation in “GEN/GEN: Generative Generations” at Gazelli Art House this September, the artist caught up with Whitney Hart to share more about her own personal journey and the worlds it has generated. 

‍Whitney Hart: What are some of your earliest memories of making art and working with computation? How did these experiences lead you to become a generative artist?‍

Melissa Wiederrecht: I have always loved art of all kinds, even since I was a tiny kid. I remember being asked in second grade what I wanted to be when I grew up, and I responded: “I want to be a famous artist,” which seems to me an odd memory because I don’t actually remember any specific art I made that early or have any memories of making it. 

During middle and high school I invented a system of drawing mandala-like motifs with coins. I would choose a single coin at random and trace either a three-, four-, or six-way rotation, and then choose another coin and trace it all the way around, touching two other traced circles exactly, and then I would repeat over and over. 

I made maybe hundreds of variations using that drawing system and every one was unique and different from the rest. Looking back, this was clearly my first generative algorithm.

Around the age of 12, I taught myself to code a computer. I went to the local library and found a book on HTML, which is how I learned to code. I went through the entire thing from start to finish and then convinced my parents to get me books online. I devoured a book on Visual Basic and then a few on Perl. On the day I turned 14 and was legally allowed to work, I was hired as a web designer and I kept that job for four years until I went to college. 

It was there that I learned Macromedia Flash and ActionScript, discovering a book called Flash Math Creativity that gave examples of work by different artists who coded in Flash. Finding that book was one of those moments in my life when everything seemed to come together perfectly and I was 100% obsessed, determined, and passionate, thinking: “the world better watch out because this is what I want to do.” 

I spent a lot of time playing with Photoshop during middle and high school, familiarizing myself with blending modes, brushes, layers, masks, and gradients, and becoming more than a little obsessed with the idea of coding my own Photoshop filters. In 2011, I received a BS in Computer Science and Math, followed by an MS in Computer Science in 2014. As soon as I had graduated, I picked up generative art again, learned Processing, and have been making generative art non-stop ever since.

‍WH: I read that you were working as a surface pattern designer prior to minting your first NFTs. How has that shaped your technical approach and your subsequent career as an artist?‍

MW: I wouldn’t say I was “working” as a surface pattern designer so much as hacking. It all started when I put up some classes about generative art on Skillshare, which was my first attempt to see if I could turn generative art into a career. I would watch other people’s pattern design classes to discover what made seamless patterns work. I then learned that you could sell surface patterns on Shutterstock but to be successful you needed a portfolio of 50,000 designs or more. 

I promptly turned to code, developing algorithms that would make seamless patterns, outputting anywhere between 20 and 200 variations of a particular style, and then upload them to Shutterstock. It turns out that coding by hand was a little too time consuming as well, so I started using Adobe Substance Designer to design algorithms that I would then shove through a Python script to output however many random variations I wanted. That way, I could design, output, and upload a collection within a few hours.

However, I got bored of adding keywords to all of my designs for Shutterstock, so I trained a deep neural network to take in seamless surface patterns and output a set of keywords to describe them (which worked really well). I ended up compiling a portfolio of over 33,000 surface pattern designs in hundreds of different collections and styles. I was grappling with the problem of variety within a long-form collection as well as the question of stylistic consistency across different collections before I ever learned of NFTs or read Tyler Hobbs’s essay on the subject. 

After making so much darn work over the course of six years, by 2020 I was gravitating toward a number of specific elements that would become core to my approach — certain colors as well as a certain texture that I would always overlay on top. I still use that texture, recoded with shaders, in all of my work and I consider it fundamental to my style.

But the process of uploading a ton of work to microstock websites was odd. There was literally zero feedback on the work, and zero way to know who had bought it, what had happened to it, what it was used for, or even if the buyer liked it. 

I also had no real boundaries — no one saying “generative art is pure and cannot include images, textures, photographs, paintings, drawings, etc.” I used all of it. I would also regularly design new sets by inputting old sets of my own patterns and reworking them into something entirely different. [This might involve] jumbling them up, colorizing them in a crazy new way, blurring here and there, stylizing like this or that, popping a few textures on top, and voila: a brand new thing based on an old thing. It was very creatively freeing.

But there was a moment I remember clearly when I had made something that I loved so much but which nonetheless made me seriously sad because I didn’t have an outlet where people would really appreciate it for what it was: kick-ass generative art. I started wishing for an alternative way to sell my work. Six months later, I discovered the world of NFTs and I have focused on that ever since. It sounds clichéd, but it is life-changing to me that generative art now has an audience with collectors, and that I can participate in this community and sell my art without ever leaving my desk in my apartment in Mecca.

‍WH: It’s incredible that you started defining fundamental elements of your visual language so early in your career and they continue to play a fundamental role in your work today. Would you say you have a recognizable style? If so, how would you define it? ‍

MW: I think I do have a recognizable style — or so I have been told — and that style comes down to process. The pieces come out looking like I made them, because I use my tools and my process to do so. One of my favorite tools is a variable blur, which just means a blur that is more blurry in some areas than others. I have used this tool in nearly all of my projects since entering Web3. I always rely strongly on color: turquoise, pinks, and reds, as well as mixes of neutrals, off-white, and off-black for contrast. I almost always top off my pieces with the same or similar set of textures. 

I think my style comes from reworking and recombining a few favorite tools in wildly different ways. I also think that my particular set of tools and processes are unique, and I would attribute this to my experience making tons of work for surface pattern design.

‍WH: In a recent conversation with Jeff Davis, you said: “I think in terms of noises, layers, levels, filters, blending modes, variable blurs, and warps rather than languages, functions, algorithms, libraries, and other ‘programmer tools’...” You speak so eloquently about your practice but it occurs to me that, for a layperson, it might be hard to comprehend the creative labor that goes into your art. Is there a way of translating the language of programming for the public in a way that doesn’t diminish the technical sophistication of generative art?‍

MW: I sometimes think “art engineering” might be a good way to describe what generative artists do. But that doesn’t really make it much easier to understand. I guess the easiest approach would be to say we write out a set of steps that the computer should follow to make an artwork in a language that both programmers and the computer understand. Translated into plain English, the steps might look like this:

1. Choose a random line thickness between 1/100th and 3/100ths of the width of the canvas and save it with the name “lineThickness.”

2. Choose a random baseline height between 0.33 and 0.67 times the height of the canvas and save it as “baselineHeight.”

3. Choose a bunch of random points that a line should draw through and save them as “points.”

4. Start drawing a calligraphic line from the left side of the screen at “baselineHeight” with a thickness of “lineThickness,” smoothly hit all the points in “points” and come off the screen on the right at “baselineHeight.”

That is the principal idea behind the element of my project, Sudfah (2022), that draws a calligraphic line. It is, of course, written to the computer in JavaScript in just the same way that a software engineer would write code to make a computer do another task. But it’s in that tiny snippet of the algorithm that one can see both the intentionality behind the art and the randomness incorporated throughout. Every time it is run, there will be a line, with a baseline of between 0.33 and 0.67 times the height and a thickness of between 0.01 and 0.03 times the width, with some random wandering but always beginning and ending at the baseline. The process is extremely intentional and the computer will not invent anything outside those bounds. 

However, there may be rare occasions when the random points chosen by the computer seem to spell out a word. In all of the many thousands of iterations generated when working on Sudfah, on one occasion it appeared to spell out “the” in cursive.

Given enough random variables, the program will produce a unique artwork that not even the artist has seen before. Generative artists choose thousands of random variables — bgColor, lineThickness, baselineHeight — in every algorithm. But very few of those random variables are ultimately exposed as features for the public to see.

The core job of a generative artist is to write the program in such a way that beautifully balances intentionality and intentional randomness to produce pleasing outputs every single time the program is run, and to make sure that the outputs are continually interesting and surprising. 

What that looks like in practice is months of iteration on the same small set of code, refreshing the browser window numerous times, and adjusting numbers by tiny fractions over and over again, trying to weed out and prevent outputs that are not quite what you wanted in order to shape the output space toward what you intended. One can output sets of thousands at a time, checking for bad outputs and trying to prevent them, with the aim of discovering pleasantly surprising things that make one excited. This process may be repeated until one is entirely satisfied, or until the deadline arrives.

‍WH: Speaking of your collection, Sudfah (Arabic for “happy accident”), you carry so many different cultures with you — born in the United States, living in Saudi Arabia as a mother of five, married to a Sudanese man, while being fully immersed in the digital ecosystem of Web3. How does your “in-between” experience shape your work as an artist? Do you regard your algorithms and their outputs as visualizations of a hybrid identity?‍

MW: I do indeed live and work and think in a strange space that is both part of and not part of all of the cultures you have mentioned (and more!). Pieces of every one of those have had an effect on me and who I am and how I think and what I choose to make. I am sure that it is actually really confusing to others on the outside trying to look in and understand where I am coming from, because I am coming from about ten different places.

Who I am and what I am trying to say with my work and what sort of hints about all that happen to surface all change over time in a strange, blurry, blobby sort of beauty. 

I am solidly only me at my core, but I think it may take 100,000 pieces of art to try to express the different angles from which one could view my identity. The layers are deep, just like the clothes I choose to cover myself with head to toe as a Muslim woman. Sudfah draws inspiration from Arabic calligraphy, which could be said to relate to my identity in the country in which I live (Saudi Arabia), my husband’s culture in which I am immersed on a daily basis (Sudan), my religion (Islam), the language I speak reasonably well (Arabic). Yet not one of these is an identity I was born with. 

That piece also draws inspiration from the messiness that I’ve learned to appreciate as a mother of five kids. It utilizes my identity as a computer scientist and a mathematician and an artist. It lives in Web3 as NFTs. It draws on my years spent trying to emulate a water spill in code while I was working on surface pattern design. It even visually and conceptually connects to the point I am trying to make in answer to this question — that a solid line has become less than clear and a little bit messy and blurry and not so easily definable. Maybe I am that line, or all 400 of them.