"A.I." (Machine Learning Algorithms) To Generate Art

auxmux

@AudioGus said:

@echoopera said:
I just look at this AI generated work the same way i look at Samples and Synth patches.

It’s what you do with it that makes it something to enjoy.

That is exactly how I have been seeing and using this stuff. It is all sample fodder. The more skills and ambition an artist has to work with it the more they bring to the work.

BTW, here is the Stable Diffusion audio side. "Generate your own custom infinite sound libraries"

Have not tried it and no idea how good it is but I did hear a couple examples and do want in.

https://www.harmonai.org/

Interesting, signed up, curious to see how it works, thanks.

echoopera

I finally had some time to sit down and process how i truly intend to use Midjourney for some visual work. The following image is a composite of several elements from the way i work to build visuals. I’ve been doing this forever and have a clear idea of how i can leverage the AI pipeline for my benefit. If Rauschenberg were alive today, i would love to see his take on exploiting the tech. I for one am enjoying the process of having access to true visual serendipity to create visual narratives the way i want to 🤪

meta.morph | echo opera

Krupa

I’m rapidly warming to the thought of partnering with one of these organisations to realise a short film I’ve been working on. Might be an interesting experiment for them, as well as a way to actually get this damn thing made 😅

david_2017

Thanks man, thoughts the same.
I really like the fact that
Midjourney.com/app site collects all your stuff.

Is there already a take on different services what you get for our money vs. quality of the AI engine?

I am beyond impressed b midjourney so far

@auxmux said:

@david_2017 said:
wow this mid journey thing is absolutely magnificent. to anyone who is running the 10$ plan. I think its 200 pieces per month... is this more than enough? we were thinking of getting the first Tier and have a lot of fun, but I have seconds thoughts about 200 requests

special thanks to @echoopera who crossposted his creations on twitter, leading me to this thread haha thanks mate!

I ran out of 200 pretty fast. I upgraded to Standard for now, while I'm having fun with this.

auxmux

@david_2017 said:
Thanks man, thoughts the same.
I really like the fact that
Midjourney.com/app site collects all your stuff.

Is there already a take on different services what you get for our money vs. quality of the AI engine?

I am beyond impressed b midjourney so far

@auxmux said:

@david_2017 said:
wow this mid journey thing is absolutely magnificent. to anyone who is running the 10$ plan. I think its 200 pieces per month... is this more than enough? we were thinking of getting the first Tier and have a lot of fun, but I have seconds thoughts about 200 requests

special thanks to @echoopera who crossposted his creations on twitter, leading me to this thread haha thanks mate!

I ran out of 200 pretty fast. I upgraded to Standard for now, while I'm having fun with this.

So far, Dalle and Midjourney have different benefits and uses. Midjourney's database seems more centered around art/illustrations, so it's not good at generating photographs. But you can upscale images to a higher resolution than Dalle. Given's Midjourney's database, you have to try to push the algorithm so the output doesn't match the aesthetic of what everyone is generating. There's definitely a "look" to midjourney, so I'm trying not to replicate that.

Dalle is better at higher level conceptualizing, like language interpretation. It generates a mix of photography and art. Some of the art is kind of meh but I'm enjoying the photography.
There isn't a defined aesthetic, the results are pretty broad.

Combining the two is also fun. Dalle also doesn't let you generate faces.

auxmux

Latest exploration using midjourney

auxmux

Regarding cost, midjourney is more expensive since it's a monthly subscription. And, because you have to pay for upscaling in /fast mode but you have "unlimited" access in /relax mode.

Since, there's no upscaling in Dalle and it's credit based, its cost is dependent on use so turns out a bit cheaper.

echoopera

Tuning the workflow over here some more this morning:

Krupa

@echoopera said:
Tuning the workflow over here some more this morning:

👌

echoopera

Time to turn things up a bit around here:

GeoTony

Your producing some lovely stuff here @echoopera .. reminds me a bit of some of Brian Frouds work… this is one of my favourites of his…

AudioGus

@echoopera said:
Tuning the workflow over here some more this morning:

AudioGus

Alrighty kids you can download and run Stable Diffusion for free on a local GPU now...

Nightcafe also implemented it. https://nightcafe.studio/

echoopera

@AudioGus said:
Alrighty kids you can download and run Stable Diffusion for free on a local GPU now...

Nightcafe also implemented it. https://nightcafe.studio/

Cool. Will this run on a Mac? Watching video now.

AudioGus

@echoopera said:

@AudioGus said:
Alrighty kids you can download and run Stable Diffusion for free on a local GPU now...

Cool. Will this run on a Mac? Watching video now.

Windows only. They have talked about M1 and M2 non GPU support but it would be way slower from seven seconds to a minute on M2 and a couple minutes on M1.

echoopera

echoopera

oops...double post

AudioGus

@echoopera

Looks like Midjourney added a new --beta model. it is awesome for some things!

NeuM

@echoopera said:

That looks as good as a Bill Sienkiewicz illustration.

echoopera

@AudioGus said:
@echoopera

Looks like Midjourney added a new --beta model. it is awesome for some things!

Yeah. I was finding it too smooth for what i want. I like the noise and grit.

It’s good for some things though. Good to have this as a rendering option.

AudioGus

@echoopera said:

@AudioGus said:
@echoopera

Looks like Midjourney added a new --beta model. it is awesome for some things!

Yeah. I was finding it too smooth for what i want. I like the noise and grit.

It’s good for some things though. Good to have this as a rendering option.

I have to bounce all over the place now between different AIs and settings. Sigh, it's like bloody iOS music apps again.

echoopera

@AudioGus said:

@echoopera said:

@AudioGus said:
@echoopera

Looks like Midjourney added a new --beta model. it is awesome for some things!

Yeah. I was finding it too smooth for what i want. I like the noise and grit.

It’s good for some things though. Good to have this as a rendering option.

I have to bounce all over the place now between different AIs and settings. Sigh, it's like bloody iOS music apps again.

I’m just sticking with MJ for now to see how far it be pushed. I really like the stylistic choices it makes at times. It always looks nice.

Even when it misfires i get these bizarre worlds. Here i asked it to render an
Astronaut in a Turkish Bazaar in the style of some artists i like, and it was a total fail but an interesting fail nonetheless. It always tries to create something aesthetically pleasing to the eye.

Have a look at these render plates:

[Deleted User]

@GeoTony said:
reminds me a bit of some of Brian Frouds work… this is one of my favourites of his…

>

I've seen a number of his paintings in 'real life', prints really don't do them justice - it's phenominal work - makes the hairs on the back of your neck stand up and pay attention.

It's a million miles from this AI stuff - pretty though, as it is. He's honed his skill for 60 years, and lives and breathes the magic. Funnily enough we're off to the village he lives in next week - a mystical, ancient place, where you can feel the energy directly from the beings he paints.

NeuM

@echoopera said:

@AudioGus said:

@echoopera said:

@AudioGus said:
@echoopera

Looks like Midjourney added a new --beta model. it is awesome for some things!

Yeah. I was finding it too smooth for what i want. I like the noise and grit.

It’s good for some things though. Good to have this as a rendering option.

I have to bounce all over the place now between different AIs and settings. Sigh, it's like bloody iOS music apps again.

I’m just sticking with MJ for now to see how far it be pushed. I really like the stylistic choices it makes at times. It always looks nice.

Even when it misfires i get these bizarre worlds. Here i asked it to render an
Astronaut in a Turkish Bazaar in the style of some artists i like, and it was a total fail but an interesting fail nonetheless. It always tries to create something aesthetically pleasing to the eye.

Have a look at these render plates:

Holy cow, these are gorgeous.

AudioGus

@echoopera said:

@AudioGus said:

@echoopera said:

@AudioGus said:
@echoopera

Looks like Midjourney added a new --beta model. it is awesome for some things!

Yeah. I was finding it too smooth for what i want. I like the noise and grit.

It’s good for some things though. Good to have this as a rendering option.

I have to bounce all over the place now between different AIs and settings. Sigh, it's like bloody iOS music apps again.

I’m just sticking with MJ for now to see how far it be pushed. I really like the stylistic choices it makes at times. It always looks nice.

Even when it misfires i get these bizarre worlds. Here i asked it to render an
Astronaut in a Turkish Bazaar in the style of some artists i like, and it was a total fail but an interesting fail nonetheless. It always tries to create something aesthetically pleasing to the eye.

Have a look at these render plates:

Yah that is lovely. Just wish I could do that sort of stuff at work. Mostly the more coherent less creative stuff works for games.

i have been doing similar tribal robot musician monks in catheral bazaar music studios.

AudioGus

@NeuM said:

@echoopera said:
Blotters Bots Sheet:

I can see the appeal and I predict the entire illustrated art market will implode because of this. It's the advent of automation in the creative arts field and it certainly won't be going back from here. Art which used to cost hundreds or thousands will cost pennies or dollars.

And I’ll make a prediction on machine learning systems in music: I predict we will see at least one “top 10” hit written and performed by one of these systems within the next 5-6 years on the music charts. Possibly sooner.

https://www.fox5ny.com/news/ai-powered-rapper-what-to-know-as-tiktok-famous-virtual-artist-gets-record-label-deal

NeuM

@AudioGus said:

@NeuM said:

@echoopera said:
Blotters Bots Sheet:

I can see the appeal and I predict the entire illustrated art market will implode because of this. It's the advent of automation in the creative arts field and it certainly won't be going back from here. Art which used to cost hundreds or thousands will cost pennies or dollars.

And I’ll make a prediction on machine learning systems in music: I predict we will see at least one “top 10” hit written and performed by one of these systems within the next 5-6 years on the music charts. Possibly sooner.

https://www.fox5ny.com/news/ai-powered-rapper-what-to-know-as-tiktok-famous-virtual-artist-gets-record-label-deal

Yeah, I heard Rick Beato mention this yesterday! Happened even sooner than I thought! Although to be fair, I was thinking of software which makes most of the decisions on its own, with minimal input. For example, if Nameless A&R Executive instructs it to “make me a hit in the style of the current UK Pop Charts” and it delivers something actually worth listening to and buying.

AudioGus

@NeuM said:

@AudioGus said:

@NeuM said:

@echoopera said:
Blotters Bots Sheet:

I can see the appeal and I predict the entire illustrated art market will implode because of this. It's the advent of automation in the creative arts field and it certainly won't be going back from here. Art which used to cost hundreds or thousands will cost pennies or dollars.

And I’ll make a prediction on machine learning systems in music: I predict we will see at least one “top 10” hit written and performed by one of these systems within the next 5-6 years on the music charts. Possibly sooner.

https://www.fox5ny.com/news/ai-powered-rapper-what-to-know-as-tiktok-famous-virtual-artist-gets-record-label-deal

Yeah, I heard Rick Beato mention this yesterday! Happened even sooner than I thought! Although to be fair, I was thinking of software which makes most of the decisions on its own, with minimal input. For example, if Nameless A&R Executive instructs it to “make me a hit in the style of the current UK Pop Charts” and it delivers something actually worth listening to and buying.

Heh, at that point even the nameless a&r executive will be out of a job as we can then just ask siri for it. The Stable Diffusion experience has shown me that these things will likely be relatively trivial in the big picture and end up just being free for all.

[Deleted User]

@AudioGus said:

@NeuM said:

@AudioGus said:

@NeuM said:

@echoopera said:
Blotters Bots Sheet:

I can see the appeal and I predict the entire illustrated art market will implode because of this. It's the advent of automation in the creative arts field and it certainly won't be going back from here. Art which used to cost hundreds or thousands will cost pennies or dollars.

And I’ll make a prediction on machine learning systems in music: I predict we will see at least one “top 10” hit written and performed by one of these systems within the next 5-6 years on the music charts. Possibly sooner.

https://www.fox5ny.com/news/ai-powered-rapper-what-to-know-as-tiktok-famous-virtual-artist-gets-record-label-deal

Yeah, I heard Rick Beato mention this yesterday! Happened even sooner than I thought! Although to be fair, I was thinking of software which makes most of the decisions on its own, with minimal input. For example, if Nameless A&R Executive instructs it to “make me a hit in the style of the current UK Pop Charts” and it delivers something actually worth listening to and buying.

Heh, at that point even the nameless a&r executive will be out of a job as we can then just ask siri for it. The Stable Diffusion experience has shown me that these things will likely be relatively trivial in the big picture and end up just being free for all.

And when all the traditional artists, writers and musicians have given up - battered down by a market flooded with work that takes a 1000th of the time to create by anyone with a mobile device and a $3 app, they'll stop producing the stuff we used to love and enjoy, and all we'll have left are the Dead Dreams of the Cold War Kids.

Welcome to Utopia.

knewspeak

Has an AI critique been created yet> @monz0id said:

@AudioGus said:

@NeuM said:

@AudioGus said:

@NeuM said:

@echoopera said:
Blotters Bots Sheet:

I can see the appeal and I predict the entire illustrated art market will implode because of this. It's the advent of automation in the creative arts field and it certainly won't be going back from here. Art which used to cost hundreds or thousands will cost pennies or dollars.

And I’ll make a prediction on machine learning systems in music: I predict we will see at least one “top 10” hit written and performed by one of these systems within the next 5-6 years on the music charts. Possibly sooner.

https://www.fox5ny.com/news/ai-powered-rapper-what-to-know-as-tiktok-famous-virtual-artist-gets-record-label-deal

Yeah, I heard Rick Beato mention this yesterday! Happened even sooner than I thought! Although to be fair, I was thinking of software which makes most of the decisions on its own, with minimal input. For example, if Nameless A&R Executive instructs it to “make me a hit in the style of the current UK Pop Charts” and it delivers something actually worth listening to and buying.

Heh, at that point even the nameless a&r executive will be out of a job as we can then just ask siri for it. The Stable Diffusion experience has shown me that these things will likely be relatively trivial in the big picture and end up just being free for all.

And when all the traditional artists, writers and musicians have given up - battered down by a market flooded with work that takes a 1000th of the time to create by anyone with a mobile device and a $3 app, they'll stop producing the stuff we used to love and enjoy, and all we'll have left are the Dead Dreams of the Cold War Kids.

Welcome to Utopia.

That utopia may yet belong to AI and not ourselves.