So have been interrogating chatGPT about myself; it's very strange to have an alternate life with only a few elements from reality, and a lot of made-up stuff, that could have been real, including jobs and prizes. My other self seems to be more successful
new diffusion-based music generator from @harmonai_org i let it play in the bg for a few hours, and i think it's less cursed than openai's "jukebox", w higher fidelity, but also feels less coherent. nothing that could be called a "chord progression" really https://t.co/YAirHVGTtB
Unlearning Language – Using AI to be less machine-like / @laurenleemack@kcimc@ycam_jp
→ https://t.co/stfn4uAtJm
Interactive installation and performance that uses machine learning to provoke the participants to find new understandings of language, undetectable to algorithms.
Using recorded music in theaters reduced that workforce from 22K live players to 200 recordists. It was an apocalypse. By the 50s, the same had happened to dance hall orchestras, &c.
There are fewer working musicians per capita today than at any time in history.
I’ve recently released gpu-io –– a WebGL computing library for real-time physics simulations, particle/agent-based systems, cellular automata, image processing, and general purpose GPU computations.
Examples: https://t.co/ricQ98ci8K
Code: https://t.co/Q97eyb1xKT
More info in 🧵
I was going to wait on this for the print copy, but before this site collapses further- @Autumnsburg, @SamPSnodgrass, and I wrote a textbook on Procedural Content Generation via Machine Learning! If you want ML for game content, this is the place. https://t.co/NsnmpiYmpN
Apparently, Stable Diffusion can be used to generate images of spectrograms from text prompts. The spectrograms can in turn be converted to audio using STFT and some tricks
Mind is blown!
https://t.co/xEe4IR9OVP
New project ! https://t.co/h19PdM32Py
For @mplusmuseum I was commissioned to make a website about blobs, and I asked 10 artists, designers, researchers to name and describe blob forms I made….
You hear that sound? It's 20k top lawyers working for Disney, Nintendo & co. getting ready to legally prevent Generative AI systems of crashing the IP game.
Is "Super Mario Mouse" still copyright infringement when you walk it 5% to the 'left' in latent space? Fun times.
In the next couple of years, we'll see the emergence of Generative AI Media Botnets. These self-improving, heavily automated end2end pipelines target the entire creative ecosystem, to make profit and/or gain influence. InfoSec and Computer Viruses are useful analogies.
"This Much I'm Worth (The self-evaluating artwork)" by Rachel Ara @rachelara displays its current estimate of its value based on its surveillance of gallery visitors and social media presence @_TheWhitechapel