State of AI

 

AI has increasingly become a polemic between people who have been studying it for a long time and people who are just getting into it. The latter group of people are singing its praises, just as people were back when the PCs came out. If you look back into that period in the late 70s and into the 80s there were groups of people who were also on opposite ends of that technology—people who were very excited about it and people who were cynical. It was that way with the internet as well.

I’m reminded of the book Silicon Snake Oil which came out in 1995 that predicted that the internet was going to go nowhere and then around 2012 there were a lot of books that came out that said that social media was going nowhere. And now we’re in that same situation with AI where there are some people saying that it’s not going anywhere or that it’s not very useful and I’m in that group. I can’t find any way to use it that is different from what I’ve used before that accomplishes the same thing. Maybe I’m looking at it the wrong way but I have no problem coming up with ideas. If I need some kind of a tool then perhaps I’ll use it. But let’s take that back to the 1980s: Say I never had a computer (and I didn’t until 1993). I was doing fine without one and, in fact, a lot of the things I’m doing now are a revisitation of that period where I’m doing things more manually and acoustically or “unplugged”. If I incorporated all the new technologies at once (computers, the internet, social media, and now AI) would that really change my creative MO? No. As a creator, you have to get to work at some point, and working now involves being involved on all those levels. But where are the core ideas in that? All the artificial ideas that I’ve seen emerge through AI are not really that interesting to me.

I suppose we’ll have to decide what’s interesting on a longer axis of people using AI exclusively, people not using AI, and people that are “centrist” or “moderate”.

6/8/2023

[6/8/2024: I’ve yet to use AI to any large degree. I see image generation as being an extension of what people used to do with fractal art, although natural-language commands are interesting. In music I want to see syllable mapping with lyrics, but that doesn’t take long for me to begin with and is the fun part. AI in many ways takes the fun out of creativity. It uses generative procedures that have nothing to do with generative music as it had been defined. It's selling point is that it can generate a song in 10 seconds. I like that they sometimes took 10 years and were finished at certain historical juncture as a part of an artistic period. I also see it another form of postmodernism without even knowing what it is. Generative music is really modernist. It wasn't that the people that were doing it were somehow saying that it was going to be the future that everyone was going to embrace--or even find interesting. Also, all these AI generators have some kind of upfront cost or fee, when you could make generative music yourself by setting up a system. But the old Koan software from the 90s wasn't free, and probably no longer exists. But it wasn't about "songs". It was a part of the avant-garde in a way, and AI music has none of that in it. Moreover, I'm never clear if in fact it uses AI, or is just piggybacking on it as a form of marketing--sort of like the "low fat" label on foods that don't have much fat in them to begin with].