1897 posts
Random
Sunday, September 1st, 2024 at 7:06 AM
Ted Chiang: Why AI isn't going to make art

More response writing! Here we go. Same rules, don't summarize, pick out what's interesting to me even if it leaves a bunch of other stuff out.

First his definition:

art is something that results from making a lot of choices.

In the case of prompting to generate text or images, he argues, you are making very few choices.

What I'm saying is that art requires making choices at every scale; the countless small-scale choices made during implementation are just as iportant to the final product as the few large-scale choices made during the conception. [...] the interrelationship between the large scale and the small scale is where the artistry lies.

This is part of why I like making prototypes, going through the process of implementation, lots of small choices show up to be made, some of them suggest rippling, larger choices.

I doubt you could replace every sentence in a thriller with one that is semantically equivalent and have the resulting novel be as entertaining.

This actually seems like a doable generative experiment...

Another big distinction, AI is not a language user because:

Language [...] requires an intention to communicate.

Made clear by the example of ChatGPT being unable to mean it when it says "I'm happy to see you".

Ending is stirring but not completely rigorous.

I like it best as meta-question of why do we create art why do we use language? There he puts the focus on intention.

That seems right to me but also limited, if we were looking for sincere intention in art, art would quickly get repetitive. We are looking for novelty too, or a response to what's came before that appears as novelty, or originality. Truthfulness is part of it, I wonder what the rest is?