The work of art exists precisely to get beyond representation, to presence, even if that presence is itself composed of words, as it is in poetry. If this were not so, a lot of effort could have been spared, as it could all have been better stated in prose. The work of art does not hide, represent, or body forth something else, that must therefore be decoded: it is precisely what it is. And yet neither is it opaque, as though we were stopped at its frontiers. It is semi-transparent, translucent: we see it all right, and yet see through it to something beyond.
The week before last, Adriana and I shot some footage of the current SELF ABSORBED prototype. We used the giant video wall at Concordia University.
I’ve been seeing images like this for two years now. I’ve spent so long with them that I’ve become desensitised to their weirdness, particularly these distortions of my face. Not that they don’t move me. But I have to remember it’s much weirder for other people. I’m coasting on my memory of my first impression.
AI is moving faster than any field I’ve experienced before. Developing work with it is exhilerating but can be stressful. Images that are mindblowing in June look dated before the year is out. If your work is riding on that novelty factor, then the rug may quickly be pulled from beneath you. I’m a slow thinker and slow creator. While the buzz may draw me in, once I’m there my only sane response is to stop, sink my feet in and immerse myself as it moves on.
For me, and I think many, immersion into a medium or theme is both the pleasure and essence of creating art. With immersion, it stops being about what’s new and is instead about what’s powerful. Things that seem obvious to me are anything but obvious to others. I believe this to be the place where it becomes possible to create art that, as McGilchrist describes above, gets “beyond representation, to presence.”
In these two years, I’ve spent many hours controlling this single AI model with my moving body, not really knowing how it works but releasing into the embodied intuitions that emerge. What I’m left with is a sense of the topology of the model. An image that looks like reality is revealed to be a coincidence of wiggly lines. The AI is trained by fiddling with the wiggles to make more and more of these lifelike coincidences emerge.
It’s hard to describe. Maybe that’s the whole point! But until you’re able to experience it yourself, I have to try.
The end of the weekly newsletter experiment
This letter concludes my February experiment in short weekly newsletters. Please let me know how you’ve found it. I’m going to reflect on how it's gone and I'll be back in touch next week with a decision on the future format.
Tim
Montreal, 24 Feb 2023