Three Things #35: September 18, 2022
Three Things #35: September 18, 2022
The impact of AI art
Lately it feels like everyone is talking about the impact of AI art, but mostly in a very pessimistic, dystopian way (that seems to be the only way people are able to talk about AI). I’ve heard people discuss who will lose their jobs in which order, the biases present in the training data, how creepy AI-generated art often is and how it fails to cross the uncanny valley, etc.. But I feel that, if we look at AI art as a tool, like any other tool, it becomes obvious that it will have a very positive impact, and that we’re at the verge of that impact really beginning. I agree that a world where artists have been replaced by AI is pretty terrifying and dystopian, but that’s not at all the world that I expect will come to pass.
For the past few days I’ve been playing with AI art tools including DALL·E 2 and Stable Diffusion. They make me feel the way I felt the first time I got my hands on a computer, the first time I wrote a computer program, and the first time I sent a blockchain transaction: I feel an explosion of possibility and creativity. While the technology is still far from perfect and today it’s still mostly just a toy, I have a very strong feeling that this will change soon. We’re at an inflection point in terms of how fast this technology is developing, how cheap and accessible it’s become, and how useful it will become. I’ve been pondering what all of this means, so I thought I’d use this space this week to reflect on the question.
Here are three ways I think AI art will change the world.
(Throughout this issue I’ll include a few images I generated using the tools described, as well as links to some of my favorite pieces that others have shared.)
Thing #1: Expression
One of the wisest things ever said was, The medium is the message. Give people new tools, new mediums for communicating and expressing themselves, and they will adapt to those tools and change how they communicate (the style or tone of communication, the length or frequency, whom they communicate with, etc.).
We don’t often realize it or think much about it, but we’re in the midst of a revolution in the way humans communicate and in the ways we express ourselves. Up until the advent of the digital age and personal computing, these things changed only very slowly and infrequently. For millennia humans have been used to expressing themselves through mediums like body language, words, music, and art. That’s a pretty powerful set of tools and there’s almost nothing we can’t express with them. But we nevertheless decided those tools weren’t sufficient.
As our lives became increasingly digital, we took the mediums we were already familiar with and developed them into novel, digitally native mediums. First came emoticons (I’m old enough to remember that these existed before emoji) and Internet slang acronyms like LOL, AFK, and ROFL. Then came emoji, animated gifs, memes, memoji and, eventually, TikTok videos and NFTs. Those of us young and tech savvy enough to be digitally native have integrated each successive generation of communication tools into our daily communication repertoire, sometimes joyfully and sometimes with some degree of cringe, but always without really noticing the process unfolding. Nevertheless, it has happened, and it’s gotten to the point where, when I find myself without one of these tools at my disposal—whether because I’m speaking to someone face to face or typing on a system without emoji or gifs—I actually feel constrained! (I think this is the real test of whether a new paradigm of expression has really entered the lexicon.)
The commentary I’ve read and heard about AI art so far almost totally misses this important point. AI art represents a new paradigm in how we express ourselves and communicate. Emoji and animated gifs seemed fringe until they weren’t. Now they’re integrated into the keyboard and chat box on nearly every app and platform, and no modern app feels complete without them. It’s only a matter of time until AI art gets cheap, powerful, and accessible enough that it’s also ubiquitous. Rather than trying to find the perfect animated gif or meme to capture your mood, you’ll enter a prompt—or, more likely, the app you’re using will just figure it out for you and offer a bunch of totally novel AI art options: images and animations no person has ever seen before. Why reuse the same gif, meme, or Telegram sticker when you can always create and share a stream of new ones?
Where does this take us? Think Snapchat filters on steroids; or, the long-awaited realization of the Holodeck a la Star Trek. Forget lame avatars or Zoom backgrounds. Once hardware catches up—and it will, it always does—and we have cheaper, more comfortable VR, we’ll finally break free from a grid of faces and be able to meet, well, literally anywhere. The missing piece of the puzzle to realizing Ready Player One’s vision of the OASIS isn’t computational power, nor is it VR and haptics, it’s AI that can generate those massive and diverse worlds fast enough for millions or billions of us to construct and explore them.
Never forget that the medium is the message. Every discussion on Twitter quickly devolves into a battle. Internet-native communication tends towards the ironic and I don’t think it’s a coincidence that the generation that grew up communicating online leans this way too. When your primary communication medium is emoji and memes, you’re going to be snarky and ironic. What happens when you throw easy, free, fast generative art into the mix? The construction any version of reality you want on demand? I don’t know the answer to this question but we’re going to find out and it’s going to be awesome.
For more: For chrissake, if you haven’t already, try out AI art. The fastest, easiest way is using this free Stable Diffusion demo (DALL·E 2 lets you generate a certain number of images per month for free, but it has too many obnoxious content filters). Think of an idea, then try to express it using the tool. While you’re at it, browse this list for inspiration.
Thing #2: Accessibility
Venture capitalists like to divide business ideas into three categories: candy, vitamins, and painkillers. As the saying goes, “throw away candy, look at vitamins, invest in painkillers.”
I think there’s a fourth, missing category, which I’ll call wings (as in, gives you wings). Wings enable you to do something incredible that you simply couldn’t do before. The personal computer falls into this category, which is why Steve Jobs famously described an Apple computer as a “bicycle for the mind.” I remember feeling that way the first time I used a computer—and, subsequently, the first time I used a GUI, the first time I used Photoshop, the first time I created a web page, etc.. The best tools not only make a task easier (a painkiller), but remove so much friction that they enable large groups of people to do things they could not (or would not) otherwise have done.
If the advice is “invest in painkillers,” then the advice with respect to wings should be: stop everything you’re doing, sit up, and pay attention immediately. You need to understand what’s going on here.
Wings (a better name might be paradigm-shifting technologies) come along once every ten years or so, at best. The camera allowed millions of people who couldn’t or weren’t interested in painting or illustration to become artists. The elevator allowed us to build up, and the air conditioner allowed us to settle and build in places we couldn’t otherwise. The bicycle, automobile, and commercial aviation gave people the gift of greater, faster, longer, cheaper mobility. The personal computer ushered in the era of home publishing and, eventually, blogging. The smartphone enabled the sharing economy and gig work. Bitcoin allowed people around the world to send value as easily as they send email.
The impact of AI art will be just as big. In order to explain how and why, I’ll use myself as an example. I’ve never been artistically inclined, which means that entire categories of creative endeavor have always been off-limits to me. I would never have imagined being able to create an animated film, a video game, or even a children’s book (at least not without finding and developing relationships with illustrators and animators I trust, who understand and are able to execute on my vision, etc.). After spending a few hours playing with tools like DALL·E 2 and Stable Diffusion, these all suddenly feel within reach—in quite the same way that publishing a pamphlet or newspaper suddenly felt in reach with a PC and a printer in the nineties, creating and publishing image and video content felt like with Instagram and YouTube ten years ago, and running a newsletter feels like with Substack today.
I cannot overstate how important this accessibility effect will be. As AI art becomes easier to use, cheaper, and more widely available (which is inevitable at this point), millions of people around the world will create billions of works of art that simply wouldn’t exist otherwise. Most of them will be crap, but there’s bound to be a few Van Goghs, Matisses, and Annie Leibovitzs in the mix. In the same way that desktop publishing and blogging led to an explosion in the amount of written material, and tools like digital mixers, autotune and SoundCloud have made it cheaper and easier to create professional-sounding audio tracks, we’re about to see an explosion of animated films, video games, and all sorts of other visual art.
Overall the trend will be positive as a huge number of talented people will be discovered and we’ll all benefit from a flourishing of art. As we all know by now, however, the rise of blogging and social media, and the democratization of publishing also had the effect of driving a massive transformation in the print media, especially news, and drove most legacy newspapers and magazines out of business—and what arose in their place is very messy and chaotic. I think the same thing will happen to other categories of creators as the trend of anyone, anywhere being a content creator accelerates. Existing, enshrined creators will continue to lose their monopolies to upstarts. Netflix, Paramount, and Disney should be very afraid of what comes next.
For more: What categories of art or content do you now feel empowered to create that you couldn’t or simply wouldn’t have created otherwise?
Thing #3: Freedom
As a thought experiment, imagine giving a painter better paintbrushes, or better paint, e.g., a wider color palette with more vivid colors. What impact would it have on her work? At worst, it would have no impact, since she’d keep using the tools she’s already used to. But it would give her more creative options: more freedom. It might also save her time, freeing her to spend more time on each piece, or to produce more art.
If we follow the same thought experiment to its logical conclusion, how could giving artists access to AI art not make their art better?
The best metaphor I can think of is the emergence of photography in the 19th century. In the beginning, of course, artists and elites ignored photography because it was too clunky, expensive, and poor, not to mention that it had the effect of cheapening art. But like so many popular technologies it rapidly improved. As it became easier, cheaper, and more widespread, and as the quality increased, artists began to fear for their careers. Before photography, artists were focused on realistically depicting fixed subjects. After photography, they were no longer needed to achieve this.
Contrary to what many people at the time believed would happen, however, photography did not put artists out of work. Instead, it had two very powerful, positive long-term impacts on art.
First, photography had the effect of creating an entirely new class of art and artist: the photographer. Millions of people that would never have become painters became photographers instead, and of course the photography industry grew quite large and created all sorts of peripheral jobs. You could say that it didn’t reach full maturity until the very recent emergence of smartphones with powerful cameras and many features, and Instagram, putting great photography in the hands of billions of people.
The second and subtler impact was that it gave artists more freedom to choose what sort of art to produce. Photography and the camera freed painters and other artists from the need to realistically capture nature and gave them the freedom to pursue more abstract subjects and types of art, ultimately leading to movements like Expressionism and Surrealism.
I’m confident that, over the long term, the impact of AI art on art as we know it will be just as big. It’s hard to forecast at this stage, but it may be the case that artists will be freed from repetitive grunt work such as manually creating pixel perfect digital art, and frame by frame animation, and will instead be freed to pursue more creative endeavors such as concept art and art direction. It may be that artists, designers, and other creatives are just able to work more efficiently with much more powerful tools at their disposal. AI art will result in more, better art being created by and accessible to far more people than today.
For more: Consider the intersection of AI art with blockchain technology like smart contracts and NFTs. What ideas, products, or experiences could this intersection produce?