“We’re trying to create a playful experience between human and machine”
An interview with the Poetry Camera creators and ICV Creative Residents Kelin Carolyn Zhang and Ryan Mather
Welcome to the latest installment of our Q&A series with our talented Creative Residents! As part of the ICV Creative Residency, we provide brainstorming buddies, creative support, and a stipend to some of the most inspiring entrepreneurs, designers, and engineers — and wait to see what they do with it.
In the latest edition (check out the previous story here), we talk with the creators and collaborators behind Poetry Camera, Kelin Carolyn Zhang and Ryan Mather. A unique merging of AI and art, Poetry Camera captures a visual and translates it into poetry. Instead of printing a photo, it prints a poem. Kelin and Ryan are using their time in the residency to expand on the camera’s various modes that influence the camera’s output, making them more customizable to the user. Keep reading to learn more about their unique collaboration, how they think about the lines between AI and art, and their words of advice for other creators.
“I like to say that Poetry Camera is not quite poetry, or photography, but this third new exciting thing.”—Ryan
Hi! Can you share a brief introduction — who are you both and what do you do?
Kelin: I’m Kelin Carolyn Zhang, a creative generalist who thrives in the space between traditional disciplines — but for legibility’s sake, I’m an independent designer, programmer, and educator. I make products that celebrate community and creativity, both for clients or for myself. For the past year and a half, I’ve focused on designing novel software and hardware interfaces for generative AI models. Before my independent practice, I designed software in more conventional settings — at agencies, startups, and tech companies.
Ryan: I’m Ryan Mather. I started my career as a toy designer at littleBits. During and after that, I kickstarted a role-playing game called FlipTales. Then I worked at Google where I designed futuristic AR prototypes to influence the device roadmap. When I started playing with GPT-3, I realized it was going to be a big deal and joined Sudowrite, which makes AI-first creative writing software.
How did you meet?
K & R: We met through a Slack community of NYC designers back in 2016.
What are you currently working on as part of the Residency?
K & R: We’re working on the next version of Poetry Camera, a camera that prints poems of what it sees. We started it as an art project so we could have fun with technology again. Now, so many people have expressed excitement about getting a camera of their own — and we continue to enjoy working on it — that we’re working towards producing a few more units in a limited drop.
What inspired the original idea of Poetry Camera?
K: The day ChatGPT came out, I stayed up until 5am playing with it. I felt like a kid again, realizing that computers can talk back to me now. All I wanted to do was to share this feeling with friends.
R: I immediately started playing Dungeons and Dragons with ChatGPT. It was so magical playing an AI-powered game that lets you do whatever you want.
K: I had just left my job designing consumer social products, but I was still in that headspace, looking for ways to use generative models to bring people together and help them have fun. How could I turn the fun of playing with AI into something that helps us talk to each other, rather than just talk to a machine? Could I design a social experience centered around collective reactions to the machine?
R: As an industrial designer, I had always wanted to design a camera of some kind, so I started to think of ideas for how a camera could harness this magic capability.
K: When Ryan mentioned the idea of a camera that prints poems, I immediately imagined using the camera as an icebreaker to talk with strangers. So I started building a cardboard prototype to do just that.
R: And so the idea for Poetry Camera was born. Some other inspirations for the project were reacting to the sensory overload of modern social media and longing for a quieter way to capture memories, and experiences of reading a book and seeing the movie version only to realize the imagination from the book version was way better than any movie could ever be.
Is the AI the poet in Poetry Camera? Or is it the programmer? Or is it still the human looking through the viewfinder? How do you see the relationship between human and machine when it comes to something like Poetry Camera?
K: I deliberately avoid putting emphasis on the poems or the photos because it traps us in old ways of thinking — we humans are used to thinking of poems (or paintings) as the end result of a laborious creative process. Generative models can create infinitely many words and images in an instant, which gives us the opportunity to use poems (or images) as intermediaries in a new type of creative process. What then becomes the end result? What are we putting our effort into creating?
What we’re trying to create is a playful experience between human and machine. We also want people to feel like they remain in control of the experience. There’s a knob on the camera to let you switch between different “modes” that influence the output — as part of the work during this residency, we plan on making these modes customizable by the end user, so they can explore even more creative options.
“We humans are used to thinking of poems (or paintings) as the end result of a laborious creative process.”—Kelin
R: I like to say that Poetry Camera is not quite poetry, or photography, but this third new exciting thing. I think the person using Poetry Camera is also in that new third state. People’s inner child and imagination naturally comes out when they use Poetry Camera, and I think that’s the right mindset to have going into this new world where AI shakes up traditional disciplinary boundaries. Poetry Camera is kind of like a playground where people get to explore these new hybrid spaces.
If type designers don’t manually set kerning anymore, are they still type designers? If accountants don’t balance the books manually anymore, are they still accountants? I think the answer to all these questions should be: “Who cares!” I think that ultimately we’re all people trying to make a dent in the universe with our own unique values and skills, and we should bring an open mind to how that unfolds.
Tell us more about the intersection of art and machine. In what other ways do you hope to see AI evolve in its role in creating inspired outputs like poetry?
R: Personally, I think AI should have every role people want it to have in their process. Technology has always been involved in art-making and writing. Even words are a kind of technology. Modern LLMs are new and scary and weird, but I don’t think they have any intrinsic traits that fundamentally change how they might fit into a creative process. Humans will always take a technology and flip it on its head, subvert expectations, and push boundaries.
K: Ultimately it’s about the human artist and their process — AI could serve as a tool, collaborator, or muse. I hope to see more artists exploring that.
Where do humans fit within this brave new world? What does adoption of technology as an artistic medium look like from where you’re standing?
K: I know so many independent artists and creatives working with various forms of technology as their medium. I see generative models as yet another tool in a creator’s toolkit — as a material artists may choose to work with.
R: I’m so happy to exist in the age of modern AI. I think it has made it so much more accessible to learn and create new things. Since starting Poetry Camera, I’ve probably read a few hundred AI generated poems, which is a lot more than the amount of poems I was reading before I started working on Poetry Camera. I recently sat down and wrote some of my own poems from scratch for the first time in a long time, and noticed how I have so much more confidence in my word choice and I can recognize my own poetic voice so much more distinctly than before. Previously I might have feared I have nothing unique to say, but my experience experimenting with AI has shown me that I do. At the same time, AI doesn’t need to be involved in every creative process, and many artists should avoid it if it doesn’t inspire them.
“Previously I might have feared I have nothing unique to say, but my experience experimenting with AI has shown me that I do.”—Ryan
Do you think AI and machines can create things that are truly beautiful?
K: Of course. Beauty can come from the surface-level aesthetics, or from a deeper meaning, or from what it makes you think about. As the saying goes, beauty is in the eye of the beholder.
R: One common critique people make of computer art is that it lacks “soul.” Personally, I find “soul” to be a very slippery term. Sometimes people will say it lacks a “human element.” But humans are so different from each other! I think if pushed, people could come up with a more specific description of what communicates “soul” or “the human element” in their own aesthetic sensibilities. Maybe it’s the humorous selection of focal point; maybe it’s the surprising focus on mundane things; maybe it’s a certain kind of warm or irregular mark-making. If we’re specific enough, I think many of us can find machine-made art that speaks to us. Some of my favorite creators of soulful digital work are Zach Lieberman, Sputniko!, RiversHaveWings, and Xin Liu.
“What we’re trying to create is a playful experience between human and machine.”—Kelin
How do you find inspiration for what you work on? Where do you collect the things that particularly spark your interest?
R: Here’s my list of current inspirations: Hayao Miyazaki, Dieter Rams, Naoto Fukusawa, George Nakashima, Midori stationery, Peow2 comics, indie video games, Dungeons and Dragons, Legend of Zelda, Final Fantasy Tactics, Anji Play, Seymour Papert, Shoyan the Carpenter on Youtube, Kaweco pens, MUJI, trees, my dog, sunsets, and tea. I don’t intellectually pursue inspiration, but I think I am naturally drawn to seek out things that make me feel warm and fuzzy inside, and those are the kinds of things that inspire me.
K: I tend to seek inspiration for how to live my life, rather than inspiration for particular ideas (which just comes naturally when I clear out distractions). For example, I love listening to interviews with creatives who have to balance artistry with business. Women in pop music particularly inspire me — Beyoncé, Lady Gaga, Taylor Swift, Lorde, Charli XCX, Chappell Roan, Olivia Rodrigo, Billie Eilish, boygenius, and more. Also, women filmmakers like Greta Gerwig. It’s incredibly hard to make something with broad commercial appeal while developing your artistry and leading a creative team.
In design — I’m very lucky to have peers and friends whose work I deeply admire. Spending time with them motivates me to work harder and get smarter. I’m also guided by gut feelings — reactions to colors, phrases, ideas — and riffing with friends. I love taking familiar concepts and turning them on their heads. I don’t tend to save or review inspiration unless I’m more actively executing on a visual project, in which case I’d save stuff on a Pinterest board or something.
What drives your artistic practice, and how does it change over time? Do you see your work as being in conversation with, or reaction to, anything?
K: I create to be seen and understood. I create to leave something behind, to embody my best self, and create the world I want to see. I’m always asking myself, “If I died tomorrow, would I be happy with how I spent my time? How would I want to be described in the obituary? Would I be proud to stand behind this work?” It made me realize I didn’t want to be a nameless employee for a corporation — I needed to make a name of my own.
R: I think my work is largely driven by two opposing forces — a desire to create work which is unique to my values and my sensibilities, and (embarrassingly) a desire to impress as many people as possible.
Let’s talk about your process. How do you start approaching an idea? How do you find collaborators?
R: My ideas always start as sketches in my sketchbook. Then, I start talking about the ones I like to friends. The good ideas make me excited to talk about, and the bad ideas fall to the wayside. If it’s really worth doing, I’ll make a prototype. If I show that prototype to people and it makes them very excited, I keep working on it. In some cases, I create a ranking system to help me choose projects — other times I’ll let the ideas choose themselves and not force it too much.
K: I get lots of ideas from joking around with friends. A lot of new ideas start with just a name or a pun. To start executing on anything, I talk to ChatGPT now to riff on how to get started. If I’m really excited, I try to make an end-to-end prototype as quickly as possible, regardless of if I’ve got other stuff going on right now. It’s important to catch the wave of creative energy while I can. Then I’ll share it publicly to see if it attracts any interest. Sometimes the “prototype” is as simple as just tweeting out an idea or mockup.
Let’s talk about tools. What’s in your stack? Can you give us a brief tour of your work stations?
R: For 3D modeling, I use Rhino. I like how freeform it is and I feel very comfortable there. For my sketchbook, I prefer the Midori dot grid A5. For my pen, I like MUJI .38mm gel pens, usually in a green or orange color which feels approachable. Most of my work happens on a laptop, and I generally design in Figma. I use a makerbot Replicator 2 3D printer to print the Poetry Camera parts. My browser of choice is Arc, because it helpfully forgets extraneous tabs and lets you organize bookmarks by project.
I like to brew tea in a Pyrex. I drink coffee out of ball jars, and my external monitor sits on a shelf made out of scrap wood. My keyboard is a split mechanical keyboard (lily58). I use a logitech gaming mouse because I also use my PC for gaming and I’m not going to play Overwatch with a handicap. I use a standing desk because I get lethargic if I sit too long. My PC is custom built and fits into a Dan Case case from Germany. Don’t relive my mistakes. I use a desk organizer from MUJI and a little shelf from Yamazaki.
K: I’m quite unfussy with my work tools; anything fancier is just too much work. I write ideas and to-do lists on loose printer paper and post-it notes. From a writing perspective, I like the Uniball Jetstream pen, but any will do. I use Apple Notes for quick notes and Google Docs for longer ones. For coding, I use whatever basic IDE is needed — `nano` in the terminal for editing code in the Raspberry Pi, Arduino IDE for ESP32, Cursor for any other projects.
I work at an absolute mess of a dining table that’s usually piled high with electronics and cardboard boxes. I love the chaos but it’s quite embarrassing to show in pictures. Things on my desk currently include:
- Cardboard prototype of a Raspberry-Pi-based Poetry Camera
- Cardboard prototype of an ESP32-based Poetry Camera
- Some round screen components I ordered on a whim
- My cat
- A monitor I rarely use except to prop up a phone tripod
What do you find most exciting about emerging technologies like AI, ML, crypto, and connected devices? What do you find most worrying?
K: I’m still excited by the fact that with generative AI models, the computer can talk back to me now. It’s just such a fun material to explore as a designer. There’s so much potential in new hardware devices connected to AI models — of course, that’s why I’m working on Poetry Camera. 🙂 There’s such an interesting opportunity to reimagine familiar devices in new ways.
I’m worried about the (very predictable) way that corporations are using AI models to replace human labor with something cheaper, faster, and worse. Job loss due to AI is already happening, not because the AI is necessarily better than the human, but because the people in charge are happy with shoddy work as long as they’re cutting costs. I’m also concerned about the environmental impact of overusing AI, when a non-AI solution would work just fine. It feels as wasteful as single-use plastics. I don’t think regulation can keep up with how quickly things are changing.
R: I am very excited about the deflationary capabilities of AI. It’s now free to talk to a lawyer, doctor, or a consultant who’s an expert in whatever you need help with!
I think the main issues I’m worried about are political and economic. I’m worried that AI will concentrate power in the hands of the few and our governments will not catch up with policy until many, many people have died (whoops, that got dark quickly!). I also think that capitalism is not a good fit for the world we’re heading into so I’m always interested in new models, and of course, none of this matters if the planet is too volatile to live on.
Any trends you’re following, in particular?
R: I think the trend of software influencers is super interesting. As software gets easier to make, people will start using software because they like the person who created it.
What are the alternative futures you hope are possible? What’s getting in the way?
K: I hope for the revival of the American Dream. I hope for an end to the Palestinian genocide. I hope for universal healthcare, a vibrant democracy, and a strong social safety net. Authoritarianism and archaic institutions are getting in the way.
Any words of advice for other creators like yourselves
K: There has never been a better time to be a beginner, thanks to the internet and YouTube and ChatGPT. You can just do things. Start your project, change your look, move somewhere new. You can change up anything you want about your life.
R: Just do the thing! You already know everything you need to know and it’s never going to be perfect.
Stay tuned for more news from the Poetry Camera team as they finish out the residency. Interested in becoming our next creative resident? Apply here.
