Midjourney: in the middle of the universe and multiverse of AI imagination

Midjourney! The most discussed and perhaps the most controversial AI programme of 2022. A 21^st century digital wormhole for the creative human mind. A growing fascination with artificial intelligence has seen mankind finding a place for AI in every scenario. In such a process of blurring the lines between design, art and technology, the inception of Midjourney came as a breath of fresh air which soon became a whirlwind. Even though many other AI programmes generate images from text prompts, such as Dall E, Imagen and Stable Diffusion, Midjourney took a slightly different path. The programme was launched on Discord, a social platform. Thereby giving the creators a possibility to not only create but share and progress as a community. Balancing between a thin line of competitiveness and togetherness, Midjourney found its niche among other AI programmes.

So, what is Midjourney?

Midjourney is an artificial intelligence programme created by an independent research lab led by David Holz, co-founder of LeapMotion. Under a freemium business model, the programme is currently in open beta and can be found on Discord.

How do you create on Midjourney?

On Discord, join a channel in the Midjourney server and type "/imagine" followed by your prompt. Enter the text prompt of what you want, run it and Midjourney gives you four different images generated from the prompt. The image selected can then be upscaled or used to create variations.

For the world which is familiar with "a picture speaks a thousand words", Midjourney gives you an image in very few words. Along with the discussion of its exemplary outputs, there are also debates in the art and design world on its intentions, much like the debates of artists and painters when photography was invented. But here we are in a world where art and photography co-exist, in fact aiding assistance to each other. However, it's a matter of time before we know what AI can bring to and take away from the industry. Amid the discussions of whether it’ll replace artists and designers, how will it shape the industry and what’s more to it, STIR talked to a few designers and architects. While exploring the AI programme, finding and creating new things for the world that is here, a world that may never be and for the utopian-dystopian world of the future, we asked them about their journey on Midjourney, first prompt, thoughts on AI in design, and what’s NEXT.

Behnaz Farahi

For a designer who works at the intersection of fashion, architecture and interactive design, Iranian-American architect Behnaz Farahi began her journey with Midjourney by re-imagining the human body. "I really enjoy changing from the intimate scale of fashion to the large scale of architecture. I am exploring ideas from public art to architecture to furniture to fashion through the lens of AI,” says Farahi about her curious exploration of the programme. Expanding the realm and beyond of imagination, the artist has experimented with many scenarios of 'what if it was real'. She adds, "I have collaborated with AI Midjourney to envision some space station occupied by industrial robotic arms autonomously feeding pods of bio-fluids and algae, transformable spaces changing scales and material properties from soft to hard using soft robotics, self-regulating transparent cocoons providing nutrition and air for plants, self-adjusting inflatable suits for astronauts allowing maximum flexibility and yet resistant to extreme solar radiation, furniture in the shape of killer whales or inflatable furry sofa." While addressing the architect in her, Farahi has also extended her search to architectural possibilities. She shares, "I am also interested in the use of such technology for re-imagining and restoring historic buildings. In particular, I have done some Iranian architecture through the lens of AI Midjourney. I am curious about how the AI algorithm manages to pull images from its source data, and apply various artistic techniques to the resulting images. Interestingly, there are very few references to Iranian moqarnasand mirror work. I am very curious about the question of the dataset and where are the weak spots in such platforms, especially when historically certain cultures have been less well documented than others. Could such models exacerbate inequality or will we ever be able to fill the gap?”

I am fascinated by how language affects creativity, especially as someone who is bilingual. Are we thinking through language when we create? Is being creative complicit with being literate? – Behnaz Farahi

What’s NEXT for you in Midjourney?

Behnaz: "I will keep exploring ideas across the scale and applications using Midjourney. After all, I truly believe design is a large umbrella and we need to embrace interdisciplinary thinking one that brings many aspects of design together; urban, architecture, fashion, jewellery, furniture and industrial design. I am also working on bringing Midjourney to 3D and I am excited about this.”

Igor Pantic

Following the curiosity surrounding Midjourney, his initial attempt was to generate a series of Mechas. Igor Pantic is a teaching fellow at Bartlett School of Architecture UCL London, former lead designer at Zaha Hadid Architects, and an enthusiastic designer exploring the connection between advanced digital tools, and both low-cost and robotic fabrication techniques. While stating his initial experience with the AI programme, Pantic shares, “Primarily I have been using Midjourney to explore concepts which are in line with my personal architectural research.” The designer who focuses on generative design methodologies and material systems informed by algorithmic logic then moved into the exploration of materiality perceived by AI. “Initially, I started with a series of timber and stone/rock sculptures, exploring the ways in which Midjourney understands material and formal properties. After this, I moved to architectural studies, which continue this line of exploration, with a series of villas or houses located in remote areas, imagining structures which could be built using locally sourced materials and building techniques. These were mostly imagined as timber structures, by prompting different types of timber and ways of forming it (driftwood, steam-bent hardwood, voxelized vs organic etc…),” he adds. In his inquisitive quest of understanding the potential of Midjourney, Pantic is creating many interesting and unreal renders of his imagination. However, while dismantling the functioning of Midjourney, he is hoping to find ways in which AI can be incorporated into design to realise more complex works and representations. 

What’s NEXT for you in Midjourney?

Igor: "I am very interested in continuing the exploration of the way Midjourney can be integrated into the everyday design process, as a way to quickly iterate through a series of ideas and expand our imagination and repertoire. Also, as part of the research in Research Cluster 9, a unit which I direct at Bartlett B-Pro course, with a few of our student teams, we speculated on the way in which diffusion models can be used as ways to generate digital overlays over the built environment and its digital twins as part of our research into the application of mixed reality within the built environment.”

Nuru Karim

Making an identifiable mark in India with his parametric architecture and computational design interventions, Nuru Karim of Mumbai-based NUDES has opened a new perspective for the young architects of the country. Placing himself in the exploration of Midjourney, Karim has recently been experimenting with AI’s representation of his architectural philosophies connected to sustainable design, natural building materials and the relationship between humans and nature. “The response to the platform has been tremendous worldwide across several industries. It has caught the imagination of artists, architects, engineers, students and even curious onlookers who had an idea or a concept but didn't have the tools to produce that first sketch. I have been using AI to explore our studio investigations at the crossroads of bodies, nature and architectural space”, shares Karim. While investigating further on machine-human collaborative models, his Midjourney prompts revolve around the curious contemplation of buildings taking birth from the earth, cities where nature and technology co-exist, leaf structures and post-pandemic chimaeras, to name a few. Defining the potential of AI in the design industry, he further mentions, “We are also exploring the possibility of reverse-engineering these sketches and bringing some of these ideas into the realm of three-dimensional models and prototypes.”

What’s NEXT for you in Midjourney?

Nuru: "Midjourney is constantly evolving, the updated beta version allows for greater control and offers unique human-machine collaboration unprecedented in the history of design "thinking" and "making." The speed of production where AI-driven neural networks are concerned is incredible. The cornerstone is to constantly innovate and push the limits of design thinking irrespective of the tools at your disposal.”

Ronald Rael

Beyond the title of architect and author, Ronald Rael is known for his enthusiastic role as a design activist. The co-founder of Rael San Fratello, an American architecture studio that designed the Teeter-totter at the Mexican border, Rael has forever been a keen participant in the discussion of architecture’s responsibility for the betterment of society. Extending his research are interests that relate to indigenous and traditional material practices in contemporary technologies. Finding new niches of addressing the essential, Rael has been experimenting with Midjourney which he states, “I am using Midjourney as a creative tool for ideation, akin to sketching or sketch modelling. Right now it is mostly an exploration, attempting to learn how to coax or evoke the images via prompts and refining the language.” Furthermore, he mentions how technology can helpful in the architectural process of deriving and representing ideas. "For years I made sketches that were renders and animations in 3D modelling, which I called #rasquachetecture — a term that means doing the best with the least. Midjourney is a way of continuing this line of quick sketches, although I have also used it for a recent commission. Perhaps the most fun I have had with the program is exploring the possibility of making piñata houses — piñatatecture!”

Midjourney is a way of continuing the line of quick sketches. – Ronald Rael

The architect whose works have been a curious introspection into the topics of additive manufacturing, border wall studies, and earthen architecture “teases out ways of generating earthen buildings with the tool.” He shares, "Because Midjourney draws from databases of images, earthen architecture doesn’t seem to be well represented in the database of images and it has been difficult but I have had some interesting outcomes.” Co-founder of Emerging Objects, while staying true to the activist and conventional concepts of architecture, explores Midjourney as an interesting tool to enhance the process of creating and indulge in the unrealistic sarcasm of AI architecture. Rael’s version of Midjourney paints a picture of AI as an intelligent kid, who has curious ideas about a thing, helpful and fun yet incomplete and learning. Here AI presents itself as a tool for a creator and not a creator itself.

What’s NEXT for you in Midjourney?

Ronald: "I look to combining images with 3D modelling and texture mapping to create new possibilities for being both in control, and collaborating with AI in image production and design. I also hope to develop a short graphic novel using the tool.”

While the sceptical discourses of Midjourney’s pros and cons juggle from ruckus to whispers, the platform is raising eyebrows and questions. Anything new is both exciting and scary! Especially when you don’t know what it is or what it is capable of. Midjourney isn’t the first or won’t be the last AI programme to generate images from text. Dall E has been around for years and Apple’s Gaudi hopes to generate 3D scenes from text prompts. However, to bring the majority of this generation - a generation that is in a constant quest for newness - onto a platform and encourage them to imagine the unimaginable or defy reality is worth a story. Though the image is generated by AI, there have been claims that it imagines what you enter. The prompt and the line of thought belong to the creator and AI only becomes a tool. While we enter prompts and AI gives us the imagery of it, we seem to be a step closer to “at some day machines would be able to read minds”. However, for humanity that still seems to dive deeper into imagining a world taken over by technology, the line seems thin. In such a context, the question seems valid, is artificial intelligence reading our thoughts or is AI influencing and shaping the way we must think?