Overtooling

Column test for Johnson Wax Building--Frank Lloyd Wright

Have the idea first, then do the engineering.

"Watch what's going on and capture the accident." --Josef Albers

In order for something to be truly innovative, there has to be faith in the idea as truly novel. It can't sound like everything else or derive something new from something that already exists and produced so it sounds exactly like it. Of course, this is a part of the equation in coming up with anything new--including synthesis of tools and original ideas. Either artificial intelligence or human intelligence in combination with AI will make that synthesis. 

In my view, the best ideas come randomly in the form of happy accidents. If artificial intelligence is to be effective in creativity it will be in its capacity to "unwittingly" cause happy accidents to happen, and for us to be aware of them and affected by them. In reality, this kind of thing is always floating through consciousness regardless of the technology. In fact, if we ever get to the point where artificial intelligence is in the flow of creativity, it could very well be that you spent hundreds of thousands of dollars on systems, only one day realize that the best ideas that you came up with had nothing to do with the technology; You could have even used an old cassette tape machine. Very often the happy accidents will appear in serendipitous cuts or elisions--the tape splice made in heaven. AI will be making these and it will be up to us to make the catch. But what if we're too busy thinking about the code to make that happen? What if we're too busy solving other software and/or hardware problems? There will be black swans there as well--a glitch making some kind of sound that gives you an idea. Ideas are "naked", "clothed" by things we choose. But if AI is the "store", then that could be an interesting metaphor: we don't really have to make things--we simply shop for them. And shopping will be the primary activity.

I think we should admit that artificial intelligence won't produce emotion well. Take for example the feeling of pathos: How would you construct a system that would make you feel it?  

How we partially or artificially construct pathos is in the selection of audio and visual elements without randomness. This is a part of production in which humans have to be almost completely in the loop.  As I've said before, artificial intelligence is simply a tool. Tools are useful for different aspects of the project. Once you're done with the excavation, heavy equipment, cranes, backhoes--you don't use them when you are doing the interior decoration. It's a silly analogy, but I think it makes the point.  Likewise, the operator of the heavy equipment isn't thinking about the shade of pink in the bathroom while pounding the caissons into the bedrock. Imagine that experience of power and the focus it brings!

This is exactly the experience I feel when I'm actually playing music or practicing. Eddie Van Halen often talked about the power of amps to move the hair on your forearms and to feel it in your chest. If we think about artificial intelligence systems doing this, it's the same analogy of thinking about the color of the bathroom walls and how they contrast with the tile colors and textures.

AI is likely to be a distraction from creativity--as most tools can be. Work with the naked idea first, not buy all the clothes and have a baby so you can use them.