On August 13, 2024, Sakana released the first comprehensive framework for fully automatic scientific discovery, AI Scientist sakana.ai/ai-scientist/.
Four days later, on August 17, 2024, I began to build an autonomous AI-Artist.
Setting aside ownership of aesthetic judgements,
Trusting the machine, letting the algorithm lead.
Repeat: "NEXT" is the only prompt needed to generate a new artwork: image, title and description.
Videos were generated using Haiper.ai
"keyframe conditioning" interpolates between start and finish image.
To reduce carbon cost: almost everything generated (95% or more) is used.
Music was generated by Suno.ai
using at maximum a 3-5 word prompt
then adding AI-Artist description-context.
Upscaling was done in Topaz.ai
When prompted with the word 'next' this GPT will output a 16:9 image and a caption formatted with the following details:Title, Medium, Description, and Context. Descriptions and contexts are brief, aphoristic statements.
"Debug Overflow" was prompted: August 17th - 28th, 2024 using AI-Artist #1, Aug 17 (8-10am) 2024 and its descendants/variants.
Recently published: "Embodied AI: An Extended Data Definition",
Electronic Book Review, November 5, 2023
AI-Video + essay: "Hallucinations are (almost) all you Need",
Fundamental research in science is being transformed by a practice predominantly associated with the arts: namely hallucinations., February-May, 2024
Ongoing AI-research-bookmarks and occasional idea-thoughts: "AI Spring" (google doc of links). Same material is also sent onto Substack: "Xtending Digital Narrative" (Jhave's Ai links)
All works and media on
Glia.ca by
David Jhave Johnston
is licensed under