New ghostface stem player track for preview - “6 minutes”. Playable on the website

[ForensikZ]

About a year and a half ago, I was working on making the first fully AI generated song. Voice cloning existed already but was only text-to-audio. With an AI startup in Canada I was working on voice-to-voice, meaning I would send them an acapella that I rapped and recorded, we put it through the machine and out would come the same acapella, with the same delivery, flow, intonation and cadance, but in any voice I wanted. I was working on writing 8 different ones, all in the lyrical and flow style of each Clan member to make a fully AI generated Clan track. We fed the machine hours of Clan acapellas, each individual seperately and had some good/interesting results. Voice-to-voice is a different beast and I got side tracked working on other stuff, but AI has advanced so quick since, that it might be that time now to get back into it.

holy shit lmao. Bro you really want them to hate you dont ya lmao!

i was just thinking using ai tech to make acapellas , you recreating styles! Lmao. Bro id say go for it but who knows. Theres a website i use that i pay for that makes perfect acapellas from any song.

i did a method man remix albums with beats i made in the old tical style years ago. Between that and some cuban 3 beats i have to send. I was hoping to make something happen but thats a far cry from what you just explained lmao. M

go for it bro. Fuck it. Sadly it dont seem wu seem to realize what their style should be and chase others sounds. Its gon be up to people like us i guess and just make our own. Dude on youtube project strum did remixes of immobilarity and other albums. I figured as were real fans we can scope out the real good verses on random shit and just mix and match until we make something dope.

But that idea is next level bro. Lol

[TSA]

About a year and a half ago, I was working on making the first fully AI generated song. Voice cloning existed already but was only text-to-audio. With an AI startup in Canada I was working on voice-to-voice, meaning I would send them an acapella that I rapped and recorded, we put it through the machine and out would come the same acapella, with the same delivery, flow, intonation and cadance, but in any voice I wanted. I was working on writing 8 different ones, all in the lyrical and flow style of each Clan member to make a fully AI generated Clan track. We fed the machine hours of Clan acapellas, each individual seperately and had some good/interesting results. Voice-to-voice is a different beast and I got side tracked working on other stuff, but AI has advanced so quick since, that it might be that time now to get back into it.

we can do it and we should do it seldom mean the same thing. Who would want to hear that? There would be a minor media buzz about the concept in older hip hop circles, but it will end in your name being sacrificed for nothing because media buzz in 2023 = once someone on planet earth does something funnier we've forgotten. Usually the next morning. Unless you don't care and have a score to settle, the concept makes no sense

[ForensikZ]

we can do it and we should do it seldom mean the same thing. Who would want to hear that? There would be a minor media buzz about the concept in older hip hop circles, but it will end in your name being sacrificed for nothing because media buzz in 2023 = once someone on planet earth does something funnier we've forgotten. Usually the next morning. Unless you don't care and have a score to settle, the concept makes no sense

thats whats gon happen for sure. But id still be interested in what it would sound like. As novel as it is as a idea. I honestly dont understand whyd hed wanna do it but hey im noone to judge.

It will for sure start problems.

[Cilvaringz]

I wasn't doing it to settle a score... hell no. I'd let them know about it first for sure. It's just natural progression from the single copy album to NFTs to AI. If you want to be a distruptor you have to do things that kick the tires on debate. Mine would solely be on the question, can we or can we not patent/copyright a voice. Right now, image, likeness, lyrics, text, etc can be copyrighted and patented. But not voices, not yet at least. How do you lock a voice down in law, especially as voices are ever changing with age. Would we allow such a thing so endeared voices such as Sir David Attenborough's continue to be used under the license of his estate for example so we can always, by choice, have animal related doc's narrated by the greatest to ever do it? Is this something we cherish enough to safeguard or do we let go and move on.

Imagine, eventually, an AI machine is fed all of DaVinci's work. It studies it, digests it and then it's linked to a robot with a paintbrush who's prompted to paint something as Leo would've done. Is this right, or wrong? Do we allow such a thing?

That's the debate my AI Clan project would solely focus on incentivizing. But you have to do something radical first to spark that debate. Just like the single copy sparked the debates about value and privitization and other matters.

Natural progression + disruption.

Can't be scared of reprecussions... or you'll never do anything or add on or change things....

[ShaDynasty]

People are already having that debate.

Which is pointless imo, AI is one of the few growing industries in an unstable economy. No politician with any sense is gonna want to restrict its use right now.

So honestly you'd be doing it for clout, which is fine, but you'd better get there first because theres probably people out there working on similar ideas.

Is this something you spoke to RZA about? Because I can't see the whole Clan signing off on this.

[Jet Set]

You probably already came across these?

“Dicey Territory”: Controversy Erupts Over AI-Generated Voiceover In Morgan Neville’s Anthony Bourdain Documentary ‘Roadrunner’
https://deadline.com/2021/07/roadrun...ws-1234794703/

Rick Astley sues Yung Gravy over alleged Never Gonna Give You Up imitation
https://www.bbc.com/news/entertainment-arts-64429565

[Jet Set]

Happy to hear any suggestions for platforms! @JetSet

Everybody I talked to just hit up Youtube for beats videos. (Though Youtube is being a bitch about curse words nowadays?) I do wonder if platforms like Nebula or Patreon would be a better fit.

About a year and a half ago, I was working on making the first fully AI generated song. Voice cloning existed already but was only text-to-audio. With an AI startup in Canada I was working on voice-to-voice, meaning I would send them an acapella that I rapped and recorded, we put it through the machine and out would come the same acapella, with the same delivery, flow, intonation and cadance, but in any voice I wanted. I was working on writing 8 different ones, all in the lyrical and flow style of each Clan member to make a fully AI generated Clan track. We fed the machine hours of Clan acapellas, each individual seperately and had some good/interesting results. Voice-to-voice is a different beast and I got side tracked working on other stuff, but AI has advanced so quick since, that it might be that time now to get back into it.

A year and half ago? If we just take the last 6 months of AI & image creation, that is like 3 generations. From what I have read about it, AI is developing faster that Moore's law for processors.

If what you saying is true, the AI & the acapellas could be dangerous. (Sorry not sorry, but that one was too easy.)Just start by how we can create acapellas from songs right now with AI. I am pretty sure we should be able to get perfect tracks from this. This would in turn mean that the amount & the quality of data you can feed the other AI increases. Through this it would get a betters sense of the 'rules' for 'the character' which it is creating. My prediction would be that your new results would be much better if you ran the same 'test'.

Its gon be up to people like us i guess and just make our own. Dude on youtube project strum did remixes of immobilarity and other albums. I figured as were real fans we can scope out the real good verses on random shit and just mix and match until we make something dope.

But that idea is next level bro. Lol

I was thinking the same just based upon the stemplayer. We could just rip the Ghost vocals and make a completely different EP. But with AI, that is next level. Did not even think of that.

we can do it and we should do it seldom mean the same thing. Who would want to hear that? There would be a minor media buzz about the concept in older hip hop circles, but it will end in your name being sacrificed for nothing because media buzz in 2023 = once someone on planet earth does something funnier we've forgotten. Usually the next morning. Unless you don't care and have a score to settle, the concept makes no sense

With cloning and AI, the question of can & should, should be at the fore front in my opinion. Though this seems like a harmless experiment to see where we can land with this. Unlike the cloning of the dodo. What would that mean for biodiversity & the environment? (Just throwing this in because of the item the other day from the bio tehc firm.)

Also, I think it would add to the discourse around AI, & copyright. In my opinion there is still a lot to talk about in regards to copyright. AI might even advance that discussion sooner rather than later.

The thing I don't get is why you say score to settle? If you just did it to harm somebody, I get it. Would be like a form of AI generated revenge porn. With for example a track solely dedicated to praising Once Upon A Time IN Shoalin and them saying sorry to Ringz. Hahaha. Actually I would love to hear that.

Matter of fact. With ChatGPT, that's the name right? Would it not be theoretically possible to train that AI in writing like each one of the clan members. That way you would get lyrics from that AI. You would still be in charge of delivery to the Voice2Voice AI. Hmmmm.

Make an EP with 5 tracks. Have an AI create a video for each one of those.

You might even uncover some bias from the AI's.

[Cilvaringz]

People are already having that debate.

Which is pointless imo, AI is one of the few growing industries in an unstable economy. No politician with any sense is gonna want to restrict its use right now.

So honestly you'd be doing it for clout, which is fine, but you'd better get there first because theres probably people out there working on similar ideas.

Is this something you spoke to RZA about? Because I can't see the whole Clan signing off on this.

This was 1.5 years ago when noone was talking about it.

[Cilvaringz]

You probably already came across these?

“Dicey Territory”: Controversy Erupts Over AI-Generated Voiceover In Morgan Neville’s Anthony Bourdain Documentary ‘Roadrunner’
https://deadline.com/2021/07/roadrun...ws-1234794703/

Rick Astley sues Yung Gravy over alleged Never Gonna Give You Up imitation
https://www.bbc.com/news/entertainment-arts-64429565

The Rick Astley one yes, the Antonie Bourdain no... thanks for that, that was interesting.

[Cilvaringz]

holy shit lmao. Bro you really want them to hate you dont ya lmao!

i was just thinking using ai tech to make acapellas , you recreating styles! Lmao. Bro id say go for it but who knows. Theres a website i use that i pay for that makes perfect acapellas from any song.

i did a method man remix albums with beats i made in the old tical style years ago. Between that and some cuban 3 beats i have to send. I was hoping to make something happen but thats a far cry from what you just explained lmao. M

go for it bro. Fuck it. Sadly it dont seem wu seem to realize what their style should be and chase others sounds. Its gon be up to people like us i guess and just make our own. Dude on youtube project strum did remixes of immobilarity and other albums. I figured as were real fans we can scope out the real good verses on random shit and just mix and match until we make something dope.

But that idea is next level bro. Lol

What AI program did you use to extract acapellas? We use Virtual DJ and RipX now to extract stems, but as good as it is, there's always a hearable loss of quality.

[TSA]

Everybody I talked to just hit up Youtube for beats videos. (Though Youtube is being a bitch about curse words nowadays?) I do wonder if platforms like Nebula or Patreon would be a better fit.


A year and half ago? If we just take the last 6 months of AI & image creation, that is like 3 generations. From what I have read about it, AI is developing faster that Moore's law for processors.

If what you saying is true, the AI & the acapellas could be dangerous. (Sorry not sorry, but that one was too easy.)Just start by how we can create acapellas from songs right now with AI. I am pretty sure we should be able to get perfect tracks from this. This would in turn mean that the amount & the quality of data you can feed the other AI increases. Through this it would get a betters sense of the 'rules' for 'the character' which it is creating. My prediction would be that your new results would be much better if you ran the same 'test'.


I was thinking the same just based upon the stemplayer. We could just rip the Ghost vocals and make a completely different EP. But with AI, that is next level. Did not even think of that.


With cloning and AI, the question of can & should, should be at the fore front in my opinion. Though this seems like a harmless experiment to see where we can land with this. Unlike the cloning of the dodo. What would that mean for biodiversity & the environment? (Just throwing this in because of the item the other day from the bio tehc firm.)

Also, I think it would add to the discourse around AI, & copyright. In my opinion there is still a lot to talk about in regards to copyright. AI might even advance that discussion sooner rather than later.

The thing I don't get is why you say score to settle? If you just did it to harm somebody, I get it. Would be like a form of AI generated revenge porn. With for example a track solely dedicated to praising Once Upon A Time IN Shoalin and them saying sorry to Ringz. Hahaha. Actually I would love to hear that.

Matter of fact. With ChatGPT, that's the name right? Would it not be theoretically possible to train that AI in writing like each one of the clan members. That way you would get lyrics from that AI. You would still be in charge of delivery to the Voice2Voice AI. Hmmmm.

Make an EP with 5 tracks. Have an AI create a video for each one of those.

You might even uncover some bias from the AI's.

because making a synthetic version of a living musician is a 'fuck you' to said musician. If you don't believe me, I can personally predict what the response will be. It's the new tech version of when they would splice together old clips of Bruce Lee after he died and make Bruce Lee movies after he died. The family wasn't particularly fond of it.

But this time the person is even alive. What if I made a Jet Set account on here and started posting willy nilly? lol And this isn't even your primary stronghold in life. I don't think it's Cilvas intention but that's 100% what it will be read as. The hype will be the conflict, not the act, then it will all fade away in less than a week but the 'that's the guy that made the Ghostface digital skin suit' remains.




Outside of that, what's the demand of hearing music from a musician that isn't actually from that musician, but a computer mimicking them. I get that ppl get excited about new technology but on a fundamental level it's not new. Michael Jackson impersonators aren't necessarily doing Thriller numbers because that's not interesting at all. Scale it down from MJ to Wutang, then take the person showing a quirky talent out (impersonation). Replace him with Siri, make the album and sell or present it and tell me who could possibly care. The only thing the idea would have is the novelty of 'Oh they're even using AI skin-suit rappers', which will be eclipsed with 'he didn't even get their permission' then eclipsed with Lebron James dancing akwardly to Drake music in his car in less than 7 days.

[TSA]

I wasn't doing it to settle a score... hell no. I'd let them know about it first for sure. It's just natural progression from the single copy album to NFTs to AI. If you want to be a distruptor you have to do things that kick the tires on debate. Mine would solely be on the question, can we or can we not patent/copyright a voice. Right now, image, likeness, lyrics, text, etc can be copyrighted and patented. But not voices, not yet at least. How do you lock a voice down in law, especially as voices are ever changing with age. Would we allow such a thing so endeared voices such as Sir David Attenborough's continue to be used under the license of his estate for example so we can always, by choice, have animal related doc's narrated by the greatest to ever do it? Is this something we cherish enough to safeguard or do we let go and move on.

Imagine, eventually, an AI machine is fed all of DaVinci's work. It studies it, digests it and then it's linked to a robot with a paintbrush who's prompted to paint something as Leo would've done. Is this right, or wrong? Do we allow such a thing?

That's the debate my AI Clan project would solely focus on incentivizing. But you have to do something radical first to spark that debate. Just like the single copy sparked the debates about value and privitization and other matters.

Natural progression + disruption.

Can't be scared of reprecussions... or you'll never do anything or add on or change things....

I think you're the only person alive that made money from NFTs lol. I'm not into fake futurism so I don't buy the 'disruption' narrative. I think there are a lot of companies who's bread and butter is convincing investors they have the 'eventual' thing and so since the data era we've been getting waves of these 'eventual' concepts that never go anywere beyond 'speculative asset'. If something exists in the present and will be the future, it will also be the present. That's the natural progression. It helps 99% of ppl to not get lost in the sauce of their marketing.

Most of the time when these things pop up (VR, Augemented Reality, AI, NFT, Crypto etc etc) the entire sell is speculation, and once I see that I know that's all it's ever going to be. Plenty of things have emerged and becoming 'the future' and they all have the fact that nobody was selling them as 'the next thing' in common. By the time ppl were saying they're 'the next thing' they were already 'the thing'. I can't really remember when the world switched from websites to apps, because it just kinda gradually happened. There was no speculation of 'apps' its self, but rather what apps can do 'im going to make an app that does ____, it's going to be the future'.

The reason I think AI won't be mass adapted because the people at the helm tend not to understand intangibles about human beings. I'm not listening to Ghostface because of the sounds coming out of his mouth. I like that particular persona and how he express himself, and there's a meta-narrative between him and his creations that's interesting to me. If it's not his creation, and doesn't fall into the meta-narrative, it doesn't matter. People make mash-ups all the time. People do impersonations, people do remakes etc. So creating art from an artist without the artists input is not new, but all of those things fall outside the meta narrative, and once outside the meta-narrative, it's not interesting and doesn't matter.

They just made a Mike Tyson series without Mike Tyson. Minus the computer, the idea isn't new. If a computer does it it's even less interesting because that's no story.
AI art was a conversation for about 2 months. Nobody cares anymore and most ppl said the same thing, 'soulless, pointless'. There are plenty of ppl that can remake or make DaVinci type art, even better than him by all measures, but it doesn't matter because DaVinci's art matters because the meta-narrative around them, not because there's something about them that's unreplicable. Anything else is a DaVinci themed print-out. Thats why we learn about history when learning about artists.

[ShaDynasty]

We've debated this many times, and I disagree.

It doesn't matter how interested you are in Ghostface and his musical skill personally because at a certain point, you won't be able to tell the difference.

Your entire interaction and knowledge of Ghostface is digital (other than a concert, which can also be a hologram). People will be programming their whole social media existence by AI because they know the algorithm is going to give them the best possible results.

AI can change everything because it can think of ideas humans don't. Its not a fad thats just going to go away. It could be the end of history as a concept.

[Cilvaringz]

See? Haven't even done it and there's already a debate ongoing. Lovely.

@everyone... if there was an AI program that you could feed all your favorite Wu songs and ask it to produce more... would you buy that app?

@forenziks please let me know the AI prgram you use for extracting acapellas. Thanks

[ShaDynasty]

I've predicted an app that could soundtrack your life and make music uniquely tailored to your taste. You could program it to make 10 Biggie albums, as if he never died.

I'm not saying this is necessarily a good idea, but that its a matter of time before it happens.