How I Think About Making Things With Ai
If you like to make things on the internet, what’s happened over the past year with artificial intelligence can be a polarizing topic. There’s rightfully deep and passionate concern over ownership, the hand of the author, fears over the loss of jobs, and the genuine anxiety which comes from something outpacing our understanding. Everything feels as if it’s moving too fast, and that any degree of legislation or oversight is probably years away. The train left the station long ago, and some of us are on it, some of us refuse to be on it, and some of us are just watching it disappear over the horizon. As we recover from a global pandemic, the last thing we need is something else to worry about. The fatigue for new is real.
For those of us on the train, who’ve dived in and are along for the ride, it’s been exhilarating. I started about a year ago making things with Wombo, and transitioned to MidJourney v2 in the early summer of 2022. I sat on the beach looking between the ocean and what was coming back from the bot on my phone, and like the fish in the southern New Jersey sea, was hooked. From there I’ve continued to experiment with platforms such as ChatGPT, Jasper and Descript.
Everything that’s happened along the way is documented here on my site somewhere. I’m really happy to have built a sprawling non-linear narrative at Eurylae, been part of the 2022 LinkedIn Creator Accelerator Program, built a sizable and fast-growing social following on Instagram, and to continue to experiment every day on my own site here. I feel as if the work has momentum, engagement, and is growing at an exciting clip, in pace with the technology itself. My work is landing with a community of the curious, and I’ve even begun to experiment with prompt purchases, voice models, subscriptions and other low-impact means of monetization.
I’ve always made things, and the spirit of this only accelerated during the pandemic with all the hours I got back from not commuting, but the past year has only served to accelerate everything now that I can actually visualize whatever is in my head. I was doing it anyway. As Arthur C. Clarke once said, “Any sufficiently advanced technology is indistinguishable from magic”, and I think that’s certainly true here. For me, this really was just a technology which came along at the right time, in the right place, for the right person.
But with any growth, comes discomfort. A sense of going just that little bit too far out into the ocean where your feet no longer touch the floor, and there’s greater risk of not waving, but drowning. These risks come not just from putting something new out into the world that’s never been done before, but also the risk and anxiety which comes with the embrace of any new technology. Especially a technology which is not widely understood, even by those enthusiastically using it. I get it, because I hear it.
And while many of the comment trolling which tends to show up for me on Instagram is safely able to be filtered out at the source, the genuine lived experience of someone’s sense of threat is not something to discount. Just because the technology is facilitating the most incredible means of visualizing anything we can think of, does not discount those who choose to make things by other means. Especially if there’s the perception of theft in the training of the artificial models themselves. Many are choosing to fight the fight. But there’s only so much that the slacktivism of changing your avatar can do before you realize you’re just screaming into the abyss. And the impotent accusations of what is or isn’t art will never go anywhere, even independent of what’s happening with technology today. For those choosing to engage, you need to read your Duchamp again.
I’ve lived through many digital ups and downs, but I can’t help but feel this isn’t a fad, and that the genie just isn’t going back into the bottle. Attempts at legislation or some form of oversight will be years late and not enough, much as they continue to be with digital identity, privacy and social engagement. That doesn’t mean they shouldn’t happen, and I’m supportive of the necessary protections which will inevitably need to come with the democratization of these means of production, especially as they relate to forthcoming elections and the processes of representation. But images created with artificial intelligence are now so endemic to our experience of the world, that they can be produced by anyone, and at scale.
It’s also the case that the technology, as advanced as it may appear today, is still in its infancy. Even in the past year as a practitioner I’ve seen enormous improvements in what it can do, especially in terms of believability and perceived truthfulness. With every release MidJourney gets better, but I get better too. This is obviously just the beginning. As the platforms mature, what they can produce will also mature. Right now it’s prompts to images, or prompts to text in the case of GPT. But I suspect that prompts to video will be here at scale before the 2024 election next November. If we thought fake news had characterized the previous two American elections, I can’t help but feel as if we’ve not seen anything yet.
So if the ethics of this are unstable and uncomfortable, as a digital creator how does one navigate it in a way which is sensitive to other practitioners and their means of making, while continuing to experiment and move forward with a new technology? I think we’re already seeing a two tier approach to the use of generative Ai tools like MidJourney. There’s a lot of folks who know which prompts to use, essentially which buttons to press, and are highly proficient in getting the bot to create the incredible. But there’s a more interesting cohort which are using the technology to tell different stories in new ways, and to visualize ideas in ways never possible before.
I’m much more interested in the latter, although of course it’s highly informed by the craft of the former. Being able to bring new worlds and ideas to life, to boldly go where no one has gone before, is still dependent upon the craft of knowing how to drive the bus just as much as knowing where to go. So ultimately it’s a healthy blend of both, but this is often where things get conflated. Prompt crafting is just that, a craft. It’s true that ‘anyone can prompt a bot’, but not everyone can produce incredible storytelling with artificial intelligence. Not all prompting is created equal, and the accusation that ‘anyone can tell the toaster what to do’ is only as smart as the person doing the telling.
I believe the best and most interesting work in creating with tools like MidJourney is informed by a deep knowledge of the humanities. Not just how to write a killer prompt. A deep empathy and sympathy with what’s gone before. Not as a means of replicating, but as a means of extending and evolving. Of iterating and remixing. Of blending together to create something new in the world. And it’s this sense of taste which separates the technicians from the storytellers. The appreciation of type, layout, colors and composition. The understanding of lighting and framing. Knowing which references to include and why. If prompting is the craft, then language is everything. That language comes from a deep understanding of art and literature.
As the old saying goes, ‘with great power comes great responsibility’, and that’s certainly true when it comes to creating with artificial intelligence. But I believe that with any technology, there are those who will embrace it, and those who will resist it. I continue to resist reading books digitally, and hate to read books online for school. I love going to the movie theater instead of watching something on my phone. I am barely on social media. Despite working in news media, I still love to get the newspaper on a Sunday morning. I still print out boarding tickets for planes. I find the idea of self-driving cars frightening.
But it wasn’t that long ago that the idea of summoning a stranger’s car and even paying a premium to get into it and be taken somewhere was terrifying. My mom always told me not to get into cars with strangers. Now I do it all the time and pay for the privilege. I willingly entrust my food to couriers. I entrust my ability to get somewhere safely to a device I know nothing about. I keep in touch with friends thousands of miles away by sending them pictures of cats being stupid.
Yet artificial intelligence is already so pervasive, and so widespread, that any attempt at assuming the moral high ground is always going to swiftly erode. You can either go with it, or not. But the choice is yours. In time the models will get smarter, and I’m sure if you’re a creator and don’t want your work included in the model, this will become possible in time too. But this isn’t something existing Ai creators can, or should be responsible for.
I firmly believe that artificial intelligence is a means of supercharging what you already do. If you’re an illustrator, it will supercharge your craft. If you work in pre-visualization for movies, it will enable you to go further and faster than ever before. It won’t take jobs, but it will change jobs. This is nothing new. The internet changed jobs, I argue for the better. I owe my entire career to the internet. And again, the choice is ours as to the extent in which we want to embrace it. I believe that if you’re already a very talented creator, then Ai offers much. A lot more than to folks who are technicians.
I tend to think about the infinite choice offered by Ai tools as creating its own unique set of modern problems. I grew up watching four channels on television. Then as time passed four became four hundred. And now with streaming, reels, stories, and a multiverse of choice, four hundred is more like four hundred thousand things I could potentially spend my time on of a rainy Sunday afternoon. But what’s the consistent problem between all these experiences over time? There’s never anything to watch.
The same is true of Ai. The tool is only as interesting as what it means to you. If it can visualize things you’ve only ever been able to dream about, that’s one outcome. If it taps into your darkest dystopian fears about the end of the world, that’s another. Everyone will (and does) have their own take on it. Everyone should have their own take on it. But like so much of modern life, your take does not mean others have to share it.
Over the course of my career, I’ve been fortunate enough to have been there during many digital firsts. I was there when it became possible to make your own website for the first time. There in the early days of interactive television, building apps, online video and the rise of social media. Many of these have become mainstream but lost much of their initial momentum. The Twitter of 2010 is a very different place from the Twitter of today. And rightfully so.
Much of what we take for granted technologically still feels like science fiction to me. I don’t understand how wi-fi works. All I know is that it does work. I don’t know how Midjourney works, but I know that it does work. And in that work comes faith. A faith in the ‘indistinguishable magic’ Clarke originally spoke about in the sixties. And with that faith comes attention, and engagement. I will willingly give my time and energy to something which inspires awe and wonder in me. I am completely seduced by what has become possible over the past year, and could not be more excited about the year to come. I don’t see it as something to fear. I see it as something to inspire curiosity. It has completely transformed my life as a creator, which had always been there, but is now supercharged beyond my wildest dreams.
