Friction is the point



LLMs and generative AI may be easy, but the hard stuff is the point because the friction of art is where the magic is. Art is not a product to be instantly spat out. I've been thinking about this since I updated my site to include an AI disclaimer and before AI was as widespread as it is now. Any artist will tell you that the process of making art is where the real magic happens.

There's an idea about making things that goes like this; An artist has an idea, executes that idea perfectly, then puts it out into the world. The art communicates everything the artist wanted to say. In this model, why not reduce friction and skip to the actual making? Go from idea to end product instantly. What this model skips out is the process, the countless revisions, the small decisions at every stage, the discoveries and the questions each new piece of work creates. I've lost track of the times I've spent time working on a poem and only towards the end have figured out what it means to me. Just look at Sam J Grudgings' revisions of a single poem to understand the changes and thinking needed at each stage. The process is the art of thinking made manifest. To skip this step is to miss the whole point of creativity - the exploration.

Friction, therefore, is essential. It is this that allows us to make these tiny decisions that change the direction of the whole. I wrote this blogpost out by hand first, not because it is easier but precisely because it is more difficult. The paper allowed me to organise my thoughts and slow down to think. When I then typed it up a couple of months later, I made changes to sentences and paragraphs, cutting whole parts and moving other parts around. I do this with poems all the time, making changes and edits as I go. Each of them alters the overall piece until it is different. But the friction of writing by hand, then typing it up, then editing allow me to think about what I am writing. Thinking exists in the friction between what I want to create and how I do it. When I can generate something that is close enough to what I want then thinking does not exist.

I see this same issue replicated across other disciplines. I'm interested in notetaking and keep a zettlekasten, which is marketed as an effortless second brain. People attempt to set one up and then realise it is a lot of work and wonder if they can automate links between notes. But the work is the point of writing down your thoughts. Updating notes, linking then together, putting ideas into your own word, making them concise and clarifying your point - this is the point of writing ideas down, so you can work with then and think about them deeply. Again, the friction of writing is the entire point.

This friction is what generative AI is missing. If you remove the stage where you consider what you are doing, thinking atrophies. If you ask Chat GPT for a blog post or poem you are not taking the time to think about it yourself. If you rely on it for quick answers then know the machine is predictive, not accurate. The making of art and writing of words is slow and sometimes difficult, but that is the entire reason. Why make anything if you don't enjoy the process? The only way out is through.



If you would like to recieve these posts in your inbox, why not sign up to my mailing list



Next: Tempestuous | Previous: Clearings