|
|
Post by andrewwerdna on Apr 6, 2024 11:42:26 GMT -5
Uh oh! It has arrived. I've been excited for something like this to come along, but considering how closed-minded people in the DS community are about AI visual art, I'm thinking we may not be ready for this. Here's the first thing I tried: "An atmospheric dungeon synth song about a clock maker" was the prompt I put in. The first few results were a little strange because they had lyrics and vocals, kind of cool and interesting sound though. But when I switched it to "instrumental," yeah it seems to definitely know what dungeon synth is: app.suno.ai/song/e6510021-d38d-43ae-9f3e-b9044de8eb3capp.suno.ai/song/0f962a96-2322-4678-8bc4-c6bf6aa860fa |
|
|
|
Post by Pilgrim's Shadow on Apr 6, 2024 14:09:05 GMT -5
|
|
|
|
Post by andrewwerdna on Apr 8, 2024 1:17:29 GMT -5
This is pretty compelling and going to be hard to ignore I think. Just from tinkering the past few days I've heard several instances of styles which seem coherent but are completely unfamiliar. Not dungeon synth, but maybe similar vibes, a darkwave song using lyrics from La Musique poem by Charles Pierre Baudelaire translated by Clark Ashton Smith: app.suno.ai/song/8effc5fa-8fbf-4f3c-befb-f8ef35bc783f |
|
vandus
Townfolk
Exploring Aquaaris...
Posts: 34
|
Post by vandus on Apr 9, 2024 13:55:18 GMT -5
Pretty neat. Seems like a fun thing to play with and it's quite impressive to see how far AI has come in this area. I just have a cynical view of AI overall that taints my excitement.
For work, it's fantastic. Hate work. Love that AI lets me work less. I do some social media stuff for a small company on top of everything else I was hired to do, so something like this would be awesome to generate little jingles for videos and such. Currently I just throw together a few loops, takes about 15 minutes to pump out a corp-core loop. If this gets even better, it could take seconds.
Gives me more time to get stuff done so I can work on music etc 
Thanks for sharing this neat tool
|
|
|
|
Post by andrewwerdna on Apr 9, 2024 22:03:13 GMT -5
I personally am interested in it mainly for what it can do differently than traditional music, same goes for AI visual art. I actually want it to look and sound like AI because the point for me is to discover what new aesthetics are possible with this technology, not to just have a shortcut to what's been done before.
Don't get me wrong, I don't think the DS scene should embrace AI wholeheartedly or whatever, but I think it is so profoundly powerful that all art will be affected by it going forward, even if in eternal opposition, so it's probably worth having an open enough mind to at least understand where it's at and what it's capable of.
But it seems like this thread has upset some people (from a Facebook thread). Sorry about that. I don't want to be causing conflict. I will shut up about AI now, and I think maybe it will be a good idea for me to take a break from posting for a while, at least until I stop being so fascinated with this stuff.
|
|
|
|
Post by doomed on Apr 10, 2024 3:20:40 GMT -5
People will always hate the new big thing when it starts becoming more accessible. AI is just another tool that will have huge impacts on many sectors. The first stages of this are always rocky, just think about when the internet went from being this weird thing you used to send emails on your home computer, to being pretty much a utility and necessity. Same will happen with AI when all the initial turbulence cools down. Personally I find it fascinating and I want to see how far it goes. I don't think it cheapens the process of making music, so long as one identifies it with a project. Obviously having AI music pass as human made music is equally as scummy as trying to trick people with a bad photoshop. You shouldn't need to apologize for being excited about AI music, it's fun, interesting and can be useful. Ultimately, AI is here to stay and it'll only keep getting bigger and wilder.
|
|
|
|
Post by Pilgrim's Shadow on Apr 12, 2024 7:32:49 GMT -5
I personally am interested in it mainly for what it can do differently than traditional music, same goes for AI visual art. I actually want it to look and sound like AI because the point for me is to discover what new aesthetics are possible with this technology, not to just have a shortcut to what's been done before. Don't get me wrong, I don't think the DS scene should embrace AI wholeheartedly or whatever, but I think it is so profoundly powerful that all art will be affected by it going forward, even if in eternal opposition, so it's probably worth having an open enough mind to at least understand where it's at and what it's capable of. But it seems like this thread has upset some people (from a Facebook thread). Sorry about that. I don't want to be causing conflict. I will shut up about AI now, and I think maybe it will be a good idea for me to take a break from posting for a while, at least until I stop being so fascinated with this stuff. I'm not on Facebook, but i think that it can also uplift some people, me included. I'd say it'll be great to read the community insights, so if you ask me, please DO talk about AI, especially while you are fascinated with it. I'm not judging anyone, but i am curious about how it might be so upsetting that one would simply expect to not hear about it at all? Even if one considers AI to be his "enemy", is it not important to discuss of the "enemy"? or learn from it or whatever? It is a big thing, but i find the drama around it quite odd. For work, it's fantastic. Hate work. Love that AI lets me work less. I do some social media stuff for a small company on top of everything else I was hired to do, so something like this would be awesome to generate little jingles for videos and such. Currently I just throw together a few loops, takes about 15 minutes to pump out a corp-core loop. If this gets even better, it could take seconds. Well I work as a composer, so as you can experience, I'm definitely screwed  |
|
vandus
Townfolk
Exploring Aquaaris...
Posts: 34
|
Post by vandus on Apr 18, 2024 6:10:08 GMT -5
I personally am interested in it mainly for what it can do differently than traditional music, same goes for AI visual art. I actually want it to look and sound like AI because the point for me is to discover what new aesthetics are possible with this technology, not to just have a shortcut to what's been done before. Don't get me wrong, I don't think the DS scene should embrace AI wholeheartedly or whatever, but I think it is so profoundly powerful that all art will be affected by it going forward, even if in eternal opposition, so it's probably worth having an open enough mind to at least understand where it's at and what it's capable of. But it seems like this thread has upset some people (from a Facebook thread). Sorry about that. I don't want to be causing conflict. I will shut up about AI now, and I think maybe it will be a good idea for me to take a break from posting for a while, at least until I stop being so fascinated with this stuff. I'm not on Facebook, but i think that it can also uplift some people, me included. I'd say it'll be great to read the community insights, so if you ask me, please DO talk about AI, especially while you are fascinated with it. I'm not judging anyone, but i am curious about how it might be so upsetting that one would simply expect to not hear about it at all? Even if one considers AI to be his "enemy", is it not important to discuss of the "enemy"? or learn from it or whatever? It is a big thing, but i find the drama around it quite odd. For work, it's fantastic. Hate work. Love that AI lets me work less. I do some social media stuff for a small company on top of everything else I was hired to do, so something like this would be awesome to generate little jingles for videos and such. Currently I just throw together a few loops, takes about 15 minutes to pump out a corp-core loop. If this gets even better, it could take seconds. Well I work as a composer, so as you can experience, I'm definitely screwed I like to be optimistic though, and say that it will help be a tool for generating new ideas at best. You also still need to have an ear for music to be able to hear if something sounds off. Plus, there seems to be a LOT of support for authentic art.
But this stuff is getting good and getting good really fast, and we can't just rely on good will to trust that everything being made from here on out is 100% human made. I can sympathize a little bit knowing that I spent hundreds of hours for a shitty rough draft of a novel, and still need to spend another hundred hours and hire an editor and figure out marketing to get this right...all while someone out there is just copy+pasting chatbot outputs and spamming Amazon with books and making money off of it. It sucks, so having your livelihood threatened extra sucks.
As you seem to indicate though, there is opportunity here. In your position, you have authority to say whether these new tools are good or bad. You can still use your skills in combination with the tool to elevate your craft in some cases.
It will just be interesting to see how all of this plays out. It's too late to stop it, so might as well do what we can to steer this ship away from the iceburgs.
|
|
|
|
Post by RangerRick on Apr 21, 2024 14:45:07 GMT -5
I just don't see the point. There are so many people actually doing things, with intention; more music than I could ever listen to. So many indie artists I could discover and boost and support. (This forum is full of them!)
What is the point of listening to something no one bothered to make?
As for "what it can do differently" -- literally the best these large model ("AI") tools can do is a mashup of existing things. It is a core of the design. It is by definition just a giant chunk of math that gives you the most likely response to a set of inputs, with some additional randomizing code for providing variation. Any wonder it might provide is taken from the people whose art was processed to train the model.
|
|
|
|
Post by andrewwerdna on Apr 22, 2024 1:47:32 GMT -5
If you grant that it can capture the wonder of the art that it has learned from, then any given generation could have the essence of wonder of hundreds or more different artists. Aren't you curious what random nouns and adjectives might spawn from that soup?
|
|
|
|
Post by RangerRick on Apr 22, 2024 13:23:33 GMT -5
I was summarizing based on the previous comments. Perhaps I should have said "subjective wonder"? Like, yeah, AI tools are fun toys to goof around with, but the curiosity lasts the 3 seconds I play with the thing and it leaves nothing behind. There's no "ooh, I want to explore this artist's catalog" or "I'd love to know what they were thinking when they did that thing, I should look for the liner notes or an interview." There's nothing there in the end. And if I'm not looking for an artist in my art, I might as well just put on any old random background music. Add to that the incredible cost of AI tools (compute, and by extension, environmental) and they are just wildly not worth it for almost everything we're throwing at them. For now, there's a gold-rush of people trying to find nails to hammer with LLMs so they are also being heavily subsidized, but the moment they start charging what it actually costs to run these models a lot of that is going to dry up and leave a ton of people hanging.
I'd just rather... support actual artists.
|
|
|
|
Post by andrewwerdna on Apr 23, 2024 1:07:33 GMT -5
Like, yeah, AI tools are fun toys to goof around with, but the curiosity lasts the 3 seconds I play with the thing and it leaves nothing behind. There's no "ooh, I want to explore this artist's catalog" or "I'd love to know what they were thinking when they did that thing, I should look for the liner notes or an interview." There's nothing there in the end. And if I'm not looking for an artist in my art, I might as well just put on any old random background music. The curiousity is lasting a lot longer than 3 seconds for me, and several of the songs I generated have been stuck in my head to annoying extent. Also, of course there's no wanting to explore artists' catalogs or interviews etc because those things don't exist yet. It literally just started and the medium itself is still being figured out. |
|
|
|
Post by arillius on Apr 28, 2024 14:01:13 GMT -5
If/when we become cyborgs, does that count as AI music? |
|
|
|
Post by andrewwerdna on May 2, 2024 9:45:52 GMT -5
|
|
|
|
Post by Mike Moth on May 6, 2024 9:53:57 GMT -5
"theosophical uncanny olde cursed nightmare" "Theosophical" is an interesting choice for a prompt. Been reading Blavatsky lately? |
|
