As an introduction, what’s your background and what do you do when you’re not making AI art?
I'm living in Helsinki, Finland. Professionally speaking I've been working in various design roles for roughly 25 years: mostly on user experience design, but having worked on just about every design domain (except pure industrial design).
For the past couple of years, I've turned consultant / freelancer, working on customer projects of various sorts, and spending the free time I have learning, teaching myself, and creating art.
My main hobby has been photography, now for almost 30 years. In many ways, I see the generative AI methods as "virtual cameras".
What influence do you think your experience in design and technology have on your artistic work?
I've always been curious about the potential of any new technology I come across, trying to figure out how to make it work for me. The tools already out there are very capable, given they are pushed hard enough.
How would you describe what you do and when did you first see your style develop?
My style is definitely influenced by my history in photography. I try to utilize the same processes and thinking as what I do when holding a camera or planning a photoshoot - old dogs find it hard to learn new tricks.
How did you first get started using AI tools and what was your ‘aha’ moment?
I've been using generative tools for about 18 months now. Starting with the VQGAN-based methods, then Pytti, Ru-Dalle, Disco Diffusion, MidJourney, DALL-E... Almost anything I can get my hands on.
There's been several aha moments: many of them are related to various integrations to bigger workflows. Running something, using its output as the input to the next stage, combining various methods, combining the generations with the classical techniques I know etc.
What was the first piece you created?
The first pieces I published on my Instagram (at @never_ever_never_land) were snippets of screenplay text I fed into VQGAN: small movies of the image generating step by step, vaguely resembling the scene from the movie the text was taken from. It felt magical!
One of the fun things about AI is its unexpectedness. Which piece are you happiest with, and which surprised you the most?
AI is fearlessly creative. I don't get people that say that AI is not creative. AI is in many ways over-creative. Like a hyperactive naive child, amateur, full of exciting answers to any request thrown its way.
The beginner's mind - Shoshin - is a hugely valuable thing to be able to call upon. As the quote goes: "In the beginner's mind there are many possibilities, but in the expert's mind there are few."
The generative tools are open and eager. Much more so than most of us are. It is tremendously hard to become naive again. It doesn't replace us, but it can serve us.
I’m fascinated by the processes and tools people use to create their work. What are your preferred tools and workflows? These could be apps, notebooks, models, hardware etc.
My workflows are constantly evolving – I get bored doing the same thing with the same methods for too long. For the past months, I've been using a lot various combinations of Stable Diffusion and Dreambooth training - exploring the ways in which you can generate and fine-tune custom models to better create what I want to create at each particular time. It's been very exciting to learn and experiment.
A custom model has many advantages... Prompting-wise one can then focus on the content, and not spend too many of the precious tokens in describing style properties.
In general, I generate LOTS, and then filter down aggressively. I have three computers locally where I run generations - quite often they're all running at the same time.
I’m really enjoying your remixed music videos and other video work – what do you think the opportunities are for AI in video and where do you think it’s going?
Videos have certainly been a passion of mine right from the start. The complexity that time brings in - temporal consistency, movement, animation, speed - bring in a whole new set of creative possibilities.
In terms of artistic expression, it's hugely inspiring. One can do many things I can't really think of how they could be done using traditional animation methods. The New in there interests me the most.
For the music videos, I also do a fair amount of commercial work for music videos and similar avenues. For bands – both big and small – AI offers a wealth of new possibilities. As part of new videos, remixing old videos, creating visual material to be used in live performances, album art, social media art – there's vast potential.
What have you found to be the biggest challenges with AI art and what are you most excited about?
The biggest challenge is gaining control. With the traditional visual methods you have almost infinite control, with AI generation control is in many way second-hand and partial. Recently I've been very excited about inpainting finally becoming a practical and working solution. Artworks can be finally created in a stepwise manner, redoing, remixing until I'm happy with the result.
This year will be the year of better UI's for AI generation (almost everything right now out there is still terrible), and also of videos becoming more popular. Very excited about both of these developments also.
What are you working on right now and what’s coming up next for you?
I have a few pretty exciting announcements coming up soon - sorry can't mention the details yet. Something I've been working very hard on recently. Definitely my best work so far. [Since writing this, Roope released the enormously successful 'Life in West America' collection on Opensea – check it out, it's amazing.]
My Big Plan is one day to be able to create meaningful short stories: videos, animations, short movies. Think like the first South Park episodes. Story-driven, generated with a combination of traditional and AI methods (whatever works!), content that is "not about AI" - they should be accessible to the so-called mythical average consumer.
The speed of development is so fast that all my estimates as to when this becomes realistic are usually widely off. Perhaps end of this year.