I thought it could be neat to have one place where you can upload AI generated content.
I hope I don’t ruffle feathers, I am simply a technologist, and technology is what it is.
Would love a space to pick at things, or marvel at what is possible.
Personally I still feel its more skeleton than meat on bones end product.
Here it goes anyway…
Prompt: I want you to make me a Jungle track with strong South African tribal elements and tribal chants, borrowing from South African genres like, House, Amapiano, and Qcom.
I feel that this one is closer to South African Gcom fused Jungle
Its prompt, simply: A Jungle track with South African tribal elements
Prompt: South African Jazz mixed with UK Garage
If you know of any real artists working in these cross section areas, or these remind you of, please link them 
I like the UKG jazz one.
I’m a technocratic experimentalist, and AI tools are no taboo for me. I already experiment with Stable Diffusion (DrawThings + Automatic1111) on my MacBook, and would happily give your tools of choice a try myself @dnk8n. I enjoy listening almost everything electronic, preferably without vocals, from Deep House, over Progressive, Breaks and Techno, to DnB and Hardcore.
Hybrid workflows are already (somewhat) possible on Logic Pro, where you can add ‘intelligent’ percussion and similar. If it’s beneficial to overcome initial creative process blockades, I am all-in for that.
What I would love to see is AI that can produce the same level of output, not as an mp3, but as a DAW project.
The feature could also hold with it a memory of steps that can be edited. This way we know how we arrived from scratch to this point, and we have control over the elements.
I know this is not how the generative AI works, but if they could achieve that, I would be so keen to test out my ideas.
For someone like me, the learning curve is so steep, I just don’t have the time. This would dumb the early phase down for me so that I could work at a higher level. Great for hobbyists, not that relevant for professionals, but in the end the output comes from the same potentials of the DAW.
I am imagining this could already be done using code and supercollider
In fact, what I described is possible in code. Therefore conceivably it should be possible to turn into a more interactive DAW experience…
I uploaded an audio sample, The litany against fear (from Dune) and told it to cut up the sample in a Jungle track.
It didn’t quite do what I asked, but the result was nontheless surprising.
I said nothing about Fear in the text, it must have done audio to text and then made some lyrics based on it.
What do you think
I found if I actually feed the text to accompany the audio I uploaded, it uses it better in the song!
Thanks for the suggestion @SlayForMoney .
Totally support that message! I hope all AI stuff just stays to one topic that is muted for those that don’t want it.
Freedom of choice is great!
Now run it through FB and YT copyright protection to check who it plagiarized.
This story in the news today:
A man in the US put “hundreds of thousands” of AI generated songs online, then used automated bot accounts to stream them, generating over $10m in royalties.
There’s another one from a few months ago too:
looking forward to your feedback
best regards enzo matari
