Pablo Delcan is one of the most exciting and prolific graphic designers working today. Originally from Spain, and now based in New York, some of Delcan’s many accolades include being chosen as an ADC Young Gun, Print New Visual Artist, Forbes 30 Under 30, and recognition from Alliance Graphique Internationale, the Society of Illustrators, D&AD Pencil, Art Directors Club and the Type Directors Club. Last month he tried something new and rather radical: he created the very first non-A.I. Generative art model, which he titled Prompt-Brush. The ask was simple and straightforward: Declan invited illustration prompts from his audience and then went about creating what was requested. The prompts were varied in their complexity, and included topics such as “a fire on a cold morning,” “a dream remaining distant,” “a sky filled with judgmental eyes,” and even one from me: “Pablo and his dad drawing together.”
The effort has gone viral and Pablo quickly built a website to house the burgeoning results. I had the opportunity to talk with him about how and why he started this initiative, his thoughts about artificial intelligence and why he’s not worried about the future of artificial art.
Debbie Millman: Pablo, I understand this project really started as a joke. What was the inspiration?
Pablo Delcan: I spend a lot of time drawing with my son, drawing things he wants to see, or things that make each other laugh. The day when I started this project, I was at the studio starting a fire, trying to get the studio to a decent temperature, while I waited to see if the logs were catching the flame, I had this idea that made me laugh. I thought it’d be funny if I turned myself into a non-AI artist that spits out simple drawings of any prompt it gets, I named it Prompt-Brush 1.0. So I put this on my social media, announced that I had created the very first non-AI art generator model and that I was taking in prompts to showcase it.
I had recently seen the latest announcement from Google Bard and I had been toying with the new GPTs feature from OpenAI, where you can train your own AI model. I’m in awe at how powerful and useful these tools are already.
The response was immediate and viral—why do you think people responded so vigorously?
There is a real tension about how AI is going to change everything. I think the excitement is overshadowed by the destructive potential it could also have. The idea of poking a little fun at this serious concern drew people in. It’s also really fun to be able to ask someone to draw something for you, see them interpret something you’ve thought of. With AI we’re all commissioning endless amounts of images and text, but it’s interesting how different it is when that same process is applied to one person instead.
Talk about the new website: what do you hope to accomplish?
I’ve done around 200 drawings that I’m slowly updating to the website every day, there is also an embarrassing queue of more than 200 prompts that I’m trying to catch up to. I think by the end of this month I will have chipped away at most of these prompts, I’m getting faster and better. It’s hard not being an AI.
I’m not sure where this project will lead but I’m really enjoying it as it is. I’m connecting with people in a new way, and I’m developing these drawings at the same time. I want the drawings to get looser as this progresses. As opposed to Midjourney and Dall-e, that keep getting better and better and highly detailed and hyper-realistic. I want these drawings to get rougher, still communicating something but in a more direct way, more child-like.
How can people participate?
I’d love for more people to participate. If anyone is interested, it’s very easy.
Visit Prompt Brush.com input your prompt in the text box on the main page, add your email so I can send you the drawing, and then be patient… The biggest difference right now between Prompt-Brush 1.0 and any other generative model is how slow it is. Prompt-Brush doesn’t operate on weekends and sometimes takes vacation days.
What have been some of the prompts that have surprised you?
I really enjoy the ones that become a mental puzzle. I’m trying to create something that speaks to the prompt, so the image and the text compliment each other but aren’t redundant. Those are more challenging. “5 second rule” was fun in that way. There are some that are more abstract, there is more for me to interpret. For example, a one word prompt I got was “Forgiveness”, it took some playing around to get to an image for that.
I want these drawings to get rougher, still communicating something but in a more direct way, more child-like.
Pablo Delcan
Any particularly difficult or challenging prompts?
The harder prompts are the ones where there is something that I don’t know how to draw. Yesterday I had to draw some frogs, different breeds of dogs, an ostrich and some horses. I don’t think I had ever drawn a frog before. So it’s a bit of a crash course on drawing anything well enough that someone might identify what it is.
What are your goals for the site?
I’m hoping the site keeps archiving these drawings for the rest of 2024. It’d be fun to see how these drawings change, or don’t change. I’m hoping to finish the year with hundreds of them and see what the next chapter, Prompt-Brush 2.0, might be. I’d also like to offer the option to have an original drawing of their prompt delivered. That might come next.
You’ve stated that that this is “a topic that has consumed so much of this year, and will likely pick up even more next year – especially for artists and illustrators.” How are you feeling about AI? Are you scared, worried, excited, all of the above? What do you think its power really lies?
Once AI is able to do the work that I do better, there won’t be a reason for me to continue doing it anymore. I don’t think I’m scared of that possibility. I think there is an existential edge to it too, would anyone want to create something that a robot can do faster and better? Once that happens the work will have to change, maybe my work will need to be more human, and I’ll have to figure out what that means. I’m mainly very excited to see all the changes that AI will bring. So far, at it’s current stage, it’s been an immensely helpful tool.
There is something AI can’t replace yet. The uniquely human need to connect with the maker.
Pablo Delcan
Have you used any AI prompts in your own work? If so, which ones and how did you feel about the results?
I use AI every day for my work. I created a GPT assistant to help me organize the prompts and keep a running queue for the Prompt-Brush project.
For my editorial images or the book covers I make, I frequently use Midjourney, Chat GPT and the new Photoshop generate tool for sketching. The results are great and get better and better each month. It’s fun and empowering.
You’ve stated that “there is something about being able to make people laugh or smile with what I’m doing that just seems so incredibly gratifying!” How do you try to do that in your practice with other projects?
I’m always trying to spark some kind of emotion in myself with the work that I do. Either making myself laugh or get that feeling of a deep breath of fresh air you sometimes get when you see something beautiful or profound. If I can get a bit of that during the process, hopefully that eventually carries through to someone else.
Anything else you’d like to share?
There is something AI can’t replace yet. The uniquely human need to connect with the maker. Consuming music, reading a book or watching a movie generated by AI is going to be an artificial connection, no matter how good it is. And I don’t think most of us will want to consume artificial art, I don’t think I would. I’ll want to feel a real connection and the real warmth of something made by another human.
