I "primed" these responses by referencing certain things earlier in the "conversation", but this is pretty dang cool - it translated the poem from the Greek text. "A conversation with Mark Twain's Ghost (via ChatGPT) about Hymn 26 to Dionysus" (pic via MidJourney Ai):
Well now, let me tell you about this here poem, "To Dionysus #26". It goes like this:
1/ While I intend to get back to multiplayer
#toolsforthought wiki designs, the conversation around
#AiArt this week has gotten heated. I think we have to be very ontologically precise when we talk about these new tools. The Stable Diffusion rendering process doesn't simply generate one image one pixel at a time, top to bottom "like a photocopier" (or store and retrieve image data as a string from a database). SD generates an image full of noise, then iterates upon it <continued>...
5/ Bellona::3 style of Jem and the Holograms:: style of Sophie Campbell:: --ar 3:2 --chaos 23 --v 4
4/ Bellona, full shot::2 Style of Olivia De Berardinis:: --ar 2:3 --chaos 5 --v 4
5/ For more on these ideas, and the scope of the reactions to the tools in various communities here is Jamie Curcio's article: “Literal War Crimes” & A Walkthrough Of My Current Illustration Process
https://mythos-media.medium.com/literal-war-crimes-f8f454105659
(A landcape in the style of Patrick Nagel, by #MidJourney and @joshuafontany)
“Literal War Crimes” & A Walkthrough Of My Current Illustration Process
Several people have asked me if I would share more how I’ve been integrating AI with my existing illustration process. The truth is, I’m not entirely convinced that anyone actually reads process…
4/ This early "exploration of the psycho-geography of the idea" builds an "idea space". This gives me a set of "starter prompts" that I keep on my multi-clipboard & paste into the Discord interface. Once I find areas of the idea space that really match my underlying Intent, I find that I upscale 2-3 of every base 4-up. For now I have not been remixing a single upscale more than 1-2
more times (beta upscale, remix, light upscale, etc). Then I'll add all the great ones to a collection or album.
3/ I usually start with an idea like "x y z, a b, style of Artist" and then experiment with prompt drifting until I feel I have something worth rerolling a lot. I will drift that prompt a few times, rerolling the base to get 4-ups. By "drift", I mean in the sense of the surrealists' Derive/drift idea of allowing your attention to guide your travel (in a city space). I change the order of terms, add --chaos and other flags, change single words, drop the artist, add more than one artist, etc.
1/ Someone in an Ai Art group asked about the proccess of these conversations. Sart with the core idea that these tools are not sentient/aware in the way that "AIs (capital I)" are portrayed to be in fiction. These new tools cannot generate Intent, but an agent doesn't have to be capable of that in order to participate in a feedback loop. The rider on the motorcycle+road system, or the surfer on the wave+board system, are in a communication feedback loop with their tools-in-action.
#MidJourney1/ I love seeing other artists "talk" to the machine-learning art tools like
#MidJourney. Generating these images is an iterative exploration of imaginary spaces, each of these images could have been the outcome of a conversation that lasted seconds or tens of minutes long.
A few more Wizards, Warlocks, Sorcerers.
MidJourney profile:
https://www.midjourney.com/app/users/328660019392282626/