That is to say, if there is some text to image generation program, and the human generates 1 image based on the first phrase they thought of, that image they generated is probably not going to be very interesting, at least based on the image generation stuff I've played around with so far. It often takes time and several attempts / variations on the phrasing to get something which is interesting and novel.
People have been finding tricks like appending the words "Unreal Engine" onto the phrase in order to produce certain types of results. Some people actually go into the code to tweak things to get desired results.
I do wonder how true this will be going forward. It feels like this media synthesis stuff is progressing at the speed of light, and perhaps it'll only take 1 iteration / attempt to craft the interesting and novel stuff that we're looking for.
So at least at the moment it's maybe not so different from anything else. I think of it as exploring the possibilities in a given space. I work on video game development, and it's the same story there. If you're creating a game prototype, you might have an initial idea which sparks development, but once the initial seed is implemented the game often takes on and is guided by a life of it's own, shaped by the possibilities in that given space which you are narrowing down into a subset which seem to work and gel together. That's how I view it anyway.
Perhaps art, going forward, is simply described as the process, not the result.
I don't doubt that a computer could (or will) produce a perfectly convincing Charlie Parker solo or Richard M. Powers painting. There really was less there than met the eye (or ear) and most of human output consists of cliches in any case.
My take is to become increasingly archaic and self-sufficient in interests. Greek red figure pottery looks like a good place to end up after the walls are filled with amateur art, but a kiln looks pricey.
I guess it's no different than the high value given to a lot of 20th C. art, it's just another form of marketing.
Maybe this bodes well for antique prices.
It very much depends on the type of DJ. But a good DJ performing live, is indeed generating music, as he feels the vibes of the crowd and mixes in the exact tracks fitting to that moment.
Not predefined track after track, but they blend in and over on each other.
If you do this with parts of a song from here and then there, with a small snippet from something else - it is indeed generating new music in my book.
And I think alien dreams and ai generated art alike ... can move into that direction. Generating art, by knowing the various tools and combining them.
Like is done here with the generated art, it probably does not make much sense for one person to claim he "created" and owns this image. Would be hard to judge, how much credit goes to the programmers, the basic AI researchers, all the people who created the input on the AI to learn ...
OpenAI already did this with Jukebox! I'm just surprised more people have not used it as an "instrument" vs as a novelty.
shameless plug: https://soundcloud.com/songshtr/sets/2106-ep
btw this is true if you upsample all 3 levels. I have found in practise that you are fine upsampling 2 levels and then using a DAW to "clean up"/"remaster" [0]... again will work for certain sounds better than others.
The last upsampling step is by far the most expensive. so cutting that down cuts total time by 75%
[0] simple hacks such as using a low-pass filter
there are various models that do this [0]... just doesn't have the same power to generate waveform music that jukebox does.
Jukebox effectively is a 2 step model a) a 3 layer VQ-VAE compresses the music (so think of this as the MIDI equivalent) and b) a transfomer then learns/generates sequences.
The compression is the expensive part of the model.
[0] https://towardsdatascience.com/how-to-generate-music-using-a...
Divination has been around throughout human history, so it makes sense that it would emerge again through the use of machine learning algorithms.
The spiritualist movement and similar late 19th / early 20th century occultism can easily be re-interpreted as oblique science fiction... or vice versa.
The whole industrial revolution is at least culturally and historically inseparable from the occult revival in the sense that all that "natural philosophy" stuff was the primordial soup of ideas from which modern science emerged. Methodological scientific empiricism is the gold that was left in the crucible, and it yields magic that actually works.
... now back to writing obscure runes in an alien language to instruct the daemons in the ether ...
In fact you could imagine a musical version of this where the operator is describing feelings or styles and the AI is generating music to match where the operator wants to go.