I made a KANYE SONG with AI - Music is FOREVER changed.

[@rpn]

Check my latest video for an update list of resources - everything you need to create your own AI-assisted music.

[@ashtonphoenyx]

So what's really going to be game changing, is being able to train and create your own voice models by merging other artists' voices. That way, you can create brand new artists to perform on your tracks.

[@larssonk22]

you should pay royalties for sampling voices just like you would with any other part of music production

[@sepia2227]

@@larssonk22 There will be ways to blur the lines and make it individual and unrecognisable it will be a free for all

[@afrosymphony8207]

@@larssonk22 yes you should but if they create an ai that can merge two voices to create something new that'd be a fucking game changer.

[@Youtubelongvideos]

@@afrosymphony8207 😁

[@anupjsebastian]

yeah eventually what will happen is there will be famous voices of people that don't exist, and aren't owned by anyone. Imagine having open source artist voices, that aren't really of anyone in particular, but just a mix of many artists to the point that they are unique and not really recognizable as any of the originals. Eventually we will have some voices that are recognizable by everyone and famous, but free for anyone to use, as they aren't real people, nor can they be traced to the originals, as they are recognizable different. That would be the nightmare of the music industry, but a huge win for music fans, and just people who are just concerned about art and not profit. Similar to how AI generated images of people that don't exist, cannot be claimed by anyone as their 'image rights'.

[@michuXYZ]

Please release a full version of this kanye AI song!!!! That just sounds sooo goood

[@danmoorewashere]

LMK what you think about my song

[@cryptoplaya69-gg2lx]

he cannot release it. It's kayne own property = own voice. and that's why. he would be sued if he did.

[@michuXYZ]

@@cryptoplaya69-gg2lx Um? No? Haven't heard of a law prohibiting using ai generated voice in songs So what if there was a person with voice identical to ye would he also be prohibited from making fake ye songs?

[@jahkobeprod]

@@cryptoplaya69-gg2lx there is no law like that. Your voice is not your property

[@p.z2389]

PLEASEEEEEEEEEEEE

[@woodybob01]

As a producer this is awesome but also scary. Though thankfully it seems like rather than AI "replacing" producers, it's more accurate to say it becomes a new tool that producers would be able to use EDIT: In the late non-forseeable future things will have changed so much that we, by it's very nature, will not be able to predict what happens. Things will be so different. People will ALWAYS want a human touch, it's creativity. The very source of our current fear stems from that same want. We're scared that we'll lose the human touch in our creative practices.

[@ihavetubes]

give it sometime.

[@domerame5913]

Yes, billions are being invested to give you a new tool, not to replace you and the need to pay you wages. Surely

[@denver-gi7ot]

what if you couldn't tell the difference?

[@Haarba1]

People will ALWAYS want what business wants them to want. As for creativity... nice joke. 99% of the people dont give a f-ck about creativity, look at the charts, look at the instagram photos or tik tok. Neural networks are already as creative as most people, NNs do exactly the same - they take previous works and mix them. What we need to be ready is to be f-cked by corps who will use this instrument, not sure how they will do it, but they will.

[@edwardmartin7090]

It will replace producers too. It is replacing vocalists as we speak, next people will be able to say, make a beat in the style of X or Y producer. Just wait.

[@Citizen5000]

After having tried it myself... huge props to you for your lyrical skills and also the delivery! You made it look easy. Its not.

[@hernebaybob]

…..yet

[@SirQuantization]

It will be soon

[@jrbone5662]

man them lyrics was ass. “ i got a fantasy that’s dark and twisted” ??? paying tribute to another man’s dead mom? wtf???

[@chief5774]

@@jrbone5662 you clearly didnt get the dark & twisted fantasy line. That was the name of one of Kanye’s Albums. & he’s obviously trying to mimic Kanye style.

[@Swissco9665]

@@chief5774 yeah but he said another bogus line.... he apologised and said he was wrong when in fact Ye was not wrong he was telling the truth and he would never apologise for that so lyrics was still AZZZZZ, I totally loved this flow and delivery and the mimic, Ye would not say that.

[@WondaRicky]

Mann fuck that can I get a full version of the sample song. Those verses had me jumpin

[@joshuadurham6349]

Why did it go so hard

[@denjisan8507]

Yeah even before kanye vocal i knew it was gon hard

[@BreGrateful]

On me!!!! I need this song that shit slaps 🔥🔥🔥

[@Stallzyx]

I swear I heard it before

[@cloud3013]

nah that shit sucks

[@TimidMC]

I want to see the tutorial on how they trained that voice model.

[@Daniel-dj7fh]

it probably just listened to kanyes songs, took word by word sample, and then did it's adjusting.

[@HIMRhustle]

You not gonna link us with the guy who gave you the tutorial? That’s a weak move imo

[@reubenginger]

Really weak move

[@gffg387]

W move

[@MrArtVein]

This dude take a hard L. Cuz he's selling a Kanye biography on Audible with his sh*tty Rodeo royalty free picture

[@LimboIsLucid]

@Mug Beer no he didn’t

[@rpn]

thats my bad, I shouted him out in the video, here it is: kzbin.info/www/bejne/oqDMmGdrqKyXmtE

[@SoraFan23]

Its gotten more advanded than before.

[@RichPrest]

Amazing! Well done and thanks for sharing this incredibly cool capability!

[@dashman2062]

I'm just thinking about how now we can consistently make new music from dead singers.

[@stalkmoonwowweee7196]

This is cool but very dangerous at the same this gives the music industry labels especially a way to take an artists voices without actually needing them alive. yes this is dope but tbh it’s also violating at the same for us artists & our voices.

[@Kidcrowley]

You’re right but also wrong, nowadays most mainstream artists sound like shit live anyway, their voices are so overproduced it’s still not them. This still sucks though.

[@wingsoffreedom3589]

I'm more excited for voice actors like Kevin Conroy being immortalized as Batman or James Earl jones as Vader which he has given his consent to.

[@dibbidydoo4318]

yo, who the hell is going to pay music industry when they can generate it themselves.

[@mikhailosiko]

That's good. They can sell their voices and quietly engage in their own lives and their own creativity. Earn money by doing nothing.

[@FromLaurel]

Yo this is fire and SO EASY TO USE! My little brother just turned 7 and I sang Happy Birthday as KANYE WEST for him xD

[@youture519]

Funny.. great idea for a gift. What Ai tool you' ve used?

[@FromLaurel]

I just followed the video :) the links are provided I think

[@jeremyreynolds2215]

Your MEGA file upload was for Kanye West. How do you generate and plug in different VOICE URLs of other artists to that field?

[@futurehistory2110]

I made a web series about this in 2017 (set in 2027) and now it's happening. My web series also predicts major cultural conflict between AI tech and creatives and even musicians protesting on the streets. Waiting to see how much of my predictions unfold... but as they say 'truth is stranger than fiction'.

[@lucaarrigo4986]

The kanye AI model link got deleted, how can I find it?

[@johnsaylor1583]

Updated URL for model or repository with others like it? The other has been removed

[@mr.dankster442]

Ai can't replace, in person performances. That's a whole different level of experience.

[@tghetto5]

What about holograms?

[@mchrisfields4773]

Tupac’s live performances post death have been seen by more people than when he was alive

[@jqhn316]

Which is why it will be used in a virtual world where it will feel like you are there.

[@badradish2116]

yet*

[@shamz_ai]

Just like artists use autotune on mics now, whats to say there won't be something similar where their voice is filtered through an Ai voice changer

[@idirmakhlaf1435]

Everything works until the last step. By clicking play on the final botton im getting an error. please help This WORKS! 1) On the top click +Code 2) put 'pip uninstall numpy' without commas press enter, and then enter Y a bit below where is asks you if you want to uninstall package. 3) put 'pip install numpy==1.23.5' and press enter 4) press 'RESTART RUNTIME' button below. 5) run last step one more time.

[@MrKrtek00]

thanks for sharing. just one remark: the collab file is not running on your machine, it is running on a Google server (you can use a slow machine, a fast machine, and the same result)

[@Mulapu1]

This is not good at all. And worse part is it's only going to advance and kill authenticity.

[@rpn]

Yep. AI represents a fundamental societal transformation. We must continue to have conversations regarding its possibilites and dangers.

[@Mulapu1]

@@rpn I'm afraid at it's best it will do a lot of damage than good.

[@rpn]

@@Mulapu1 I still continue to believe that human creativity will always prevail, but what you are proposing is a very real and dangerous possibility

[@xXtuscanator22Xx]

@@Mulapu1 AI has to happen and we’re going to have to leave what we’ve always known in the past, and make way for what the future holds.

[@internet_dude]

@@rpn so kanye west saying he loves hitler could have been A.i with an actor at the alex haley show

[@juanmoor7560]

imagine how many people have been framed with this tech that has certainly been around since auto-tune has been a thing..

[@alecfante8967]

got my grammy vote

[@fauxdauteur]

Fuck, is it too late to close this Pandora's Box? We better enjoy this last generation of music before AI takes over.

[@leedza]

It's not just music, media, regular jobs. This stuff could make us irrelevant in so many ways.

[@Strepite]

It’s the end of mainstream music thats mostly bullshit, the real power is in underground

[@Rome.Monroe]

This is amazing a songwriter because I write and produce so many songs and I kinda hate my voice. Using the actual singers voices on reference tracks is a game changer. And like you said... this is the WORST it will ever be.

[@Thesmashconnoisseur]

Oh my god this is actually the future!

[@bobackerman7430]

Is it possible to use gpt-4 to write a song and then use a celebrities vocals to cut it? Or do you have to physical sing it first?

[@rpn]

It certainly is, but it won't sound great without a reference track. This will of course change as the tech advances over the next months and years.

[@nickmork13]

Thanks for sharing the code! This is fascinating.

[@thewallstreetbusinesstrapp2761]

How do you use the code?

[@hernebaybob]

Dead singers and groups are coming back - The Beatles singing Oasis songs!

[@MamiMrl]

This is extremely mind-blowing!

[@HIPHOPCROWNNATIONTVTM]

This is insane but dope at the same time 🔥

[@cardonbmusic]

Is there any way to train our own models based on our own voice or another singer's voice?

[@Silenced11]

y'all better check my uploaded version of this before it gets removed. Uploaded a preview!

[@AlphaNxOmega]

Man can you release the full sample song. It was clean

[@jamestodd51]

Where can I find more voice models?

[@fromnothingalegend]

We need a whole song of this... 🔥🔥🔥

[@WXNBEATS]

yeah, kanye with that flow is crazy

[@adany4556]

This was made 3 weeks ago, Ghost writer made a song called Heart on my sleeve by AI Drake and AI The weeknd. Things are moving fast!!!.

[@paulnunez8033]

This is awesome idea I’m in on this. Only thing is now artist will start charging for using their voice and name, as any upcoming artis can say “ I have a song with Kanye “

[@guitarmeetsscience]

I saw Penguinz0's video which led me here. Awesome work, and thanks for the tutorial! How in the heck was that model created though?

[@rpn]

That is beyond my understanding … I’ve tried to track down its creator(s) to ask them

[@guitarmeetsscience]

@@rpn I've tried a few - especially on huggingface, but the clone you're using is at another level. My guess is that it was trained over a good period of time using Python libraries (and probably a really new GPU). If I find out I'll let you know. Meantime - much success to you!

[@rpn]

@@guitarmeetsscience Yeah man - I have tried them all on Huggingface, this Kanye one is by far the best one on the Internet at doing what it does. Haven't found any other celebrity models that are even close

[@3dge315]

@@rpn Actually here is a Michael Jackson on that I'd says around this level. kzbin.info/www/bejne/rJnZqWiGobOKaLc

[@rodericklan1824]

the short answer is calculus, a lot of samples, time, and a gpu farm

[@ItsTheLaw]

But how can we create a new voice? Like someone other than kanye?

[@byebye1i1849]

fax

[@xxxod]

You make a data set somehow

[@domerame5913]

You just train it on human voices

[@B-Nice]

Be careful, you might get Kanye to say some crazy shit with this tool..... oh wait 🤷🏽‍♀️

[@rpn]

😅😅

[@filipstevens6263]

Hi , your vids are awesome , can i change in a for example muse song the voice of another famous singer in ai , and what ai tool should i use

[@matrixfull]

I feel like singers will barely have any work in singing. They will have even more time to focus on presentation ( being the face of the songs and albums ) than they do now. Some may opt for being even lazier than now but some will use extra time to make catchy show.

[@abracadabruh1313]

Using Kanye's voice over your wack bars should be illegal.

[@rpn]

Lmao

[@paulb1953]

yeah we should use the tech to make Ye say something non hitler glazing instead

[@LilZoty]

do something better the ghost writer

[@furious1777]

This was even better than any of kanye's bars mug

[@EskyIsAlive]

This sounds fuckng crazy lol, kinda weird to mention his mum but ye aside from that my mind is blown

[@trap_time]

I did convert and nothing happened

[@lucasmaroney4041]

Same did u find out how to fix ut

[@Cooly-D]

I was just having some fun with it.. only the next day it stopped working, you have that same problem? getting that runtime error in the last process

[@samons4002]

having the same problem, did u you find a solution ?

[@noisylightning]

Musicfy is a good way to do this.

[@firework8140]

Just watched a critikal video on this. I wonder if artist or labels can now copyright their voice

[@Breadandbananas]

Hell, I've seen news and legislature around AI so the concern is definitely there

[@clownavenger0]

@@Breadandbananas there is money involved so the vultures will come

[@xxxod]

@@clownavenger0 you can’t copyright a voice

[@clownavenger0]

@@xxxod currently... People with money get to make up new laws in a capitalist country.

[@ScottThePisces]

@@xxxod you couldn't copyright anything at one point, laws change with the times

[@We_God]

This could be good for musicians, lots of people give them ish for lip syncing but doing concert after concert, tour after tour is going to be very stressful on the artists vocal cords. I could see them incorporating a hologram of themselves in concert while they are doing another concert at another location. This could also be very problematic; but not just for musicians. This is a turning point in our history.

[@chriswright7124]

Most of your modern artists just lip sync anyway or perform karaoke. They are paid that high to perform that service.

[@robo_t]

The society we live in now and the systems we have in place, it seems more that it will be a problematic thing. Also, the big part of going to a concert is to see that artist live in person, what difference would it be than watching it at home besides the fact that you’re in a different place with other people. Really, what would be the point in going when you know that they aren’t actually there and it’s all just prerecorded/automated

[@xoslasher-_--5786]

imagine paying top dollar to see your favourite artist and its a hologram how exciting

[@watermelon2223]

Charlie

[@clownavenger0]

If Kanye actually came back with these bars I think many people would forgive him. I understand losing a loved one while blaming yourself can destroy a persons mental state. I've experienced half of that with the loss of my younger brother growing up and it messed me up for 9 years. While he said some crappy things nobody had their life ended by him or anything. I wonder if he will ever redeem himself.

[@YEE941]

jew mad

[@xCokeMaanx]

I liked Kanye a lot but all that crap he’s talking about is really bad and I can’t support that. I really hope he will turn around and come clean again.

[@Flapjack3734]

Great video, Ye covers are about to go crazy

[@banji7612]

Isnt it all basically Vocaloid? I mean its pretty much the same and it has been around for over 10 years

[@eapinkerton17]

Whats the program

[@marcoshcd]

How did you create the Mega File with Kanye's voice? Is it from Eleven labs? Please tell us.

[@ryudious]

No, it’s Google Collab

[@Princivil]

This is scarily amazing.

[@IssaSissoko]

incredible

[@6lack5ushi]

home boy ran code with almost no idea what he was doing, thats the real game changer here!!!

[@nikkig.2040]

I like you a lot... hugs... I knew this would happen and your are just bringing it to light thank you

[@manavgx]

Roberto putting our the best content

[@rpn]

appreciate u brother!!

[@LozaAlexander]

links dont work anymore they already cracking down on it smh

[@regulusashcroft]

4:12 mine have no `aimodel` selection in it, do you got any clues?

[@pikasfed]

Damn, thanks for the tutorial. What about training new models? Dp you have any references or tools for that?

[@Benprogfuse]

That would be cool if they made one to model people who actually sing, instead of people who talk into microphones..

[@JavierEzpeleta]

today they took down the Kanye AI voice model file from mega 😭 do u know where/how to find other voice models?

[@ajsmets]

How do I get other artist models?

[@saifury2]

Dude finish the song, those lyrics were FIREE

[@swaydaygaming7571]

Yé about to use this and create 50 versions of himself and charge for each album lol

[@Maxenakamarco]

I'm waitin' on the new Kanye The AI Kanye The well-trained Kanye Straight from the prompt Kanye

[@ExpressAI]

The mega file has been taken down :( What other models can we use?

[@kevinondatrack7754]

brooo, wtffff, its incredible😮

[@martinvallee6609]

Thanks for your video, where can we find others models?

[@Tuumku]

Hello there I keep getting this error on the last step. Do you know by any chance that how to solve this problem (ImportError: numpy.core.multiarray failed to import)

[@tlegendbeats]

That was a dope bar

[@CabrioDriving]

The next thing is AI mixing famous voices of similar type into a new voice, so nobody can sue you. It is what people will do.

[@GamingDad]

Or use their own voice but have AI make it sound good.

[@Pawnsappsee]

​@@GamingDad I want my voice to sound good😅

[@3RDEYELOVE]

Ultra dope💯 you gona be linked with Ye very soon

[@tomblaze2]

Kanye 100% replaced over night -

[@alanbeckco]

This sounds absolutely sick man. Do you mind if I use the singing part in a video and then push people back to your channel to watch the tutorial? There's not many creators I recommend on KZbin in terms of "Ai" but this is amazing

[@jmiles.]

its not working when pt I megalink then last step I put in file to content folder.. when click conver get a go tohubert folder error or something no play button or anything

[@84unisol]

Whats crazy is this sound like the verse he would rap if he was apologizing.

[@SoraFan23]

AI is really changing everything. I am so concerned for the future.

[@marimoedmond]

No no no, those bars! That track goes hard!

[@aaronalbertson8670]

What's that song playing in the background before you went to the Kanye AI song?

[@vladimirvladimirovich5177]

Peace, man! Incredible work! But I can't access full tutorial. Kanye requested to delete?(

[@blurqeqoherds]

The tutorial video has been removed. Any other info on the method behind this?

[@ellalstrauss1994]

Impressionante!! Mas eu quero o resto da música agora!! Nossa, posso produzir minhas próprias músicas e fazer com que a IA cante!! É genial!! Melhor do que esperar retorno de alguma gravadora que dificilmente vai aceitar a letra, o tom... Enfim, obrigado!!

[@kryptonera719]

Bro, you mind sharing more the next step after downloading it🙏 how do I get to add the music 🙏

[@BeidRiskTM]

FULL VERSION!!!

[@alemejilla6769]

We are loking at the future right now.

[@meerubruh]

sick brother

[@Deform-2024]

That's a good amount of artists out of jobs in the music industry...

[@Growthlifestyle]

Your 8 bars are 🔥

[@chrispviews]

I got this working and its impressive and fun. what's the difference between this and Eleven Labs quality? and the parameters for customization? how can I use a dataset for Kevin Hart or Obama or Eddie Murphy vocals?

[@thefamousdjx]

U a real one thanks for sharing

[@infinitysrustplace9263]

Please drop the whole song, it's insane

[@John055TO]

I like this ... a lot. Lawyers losing out to AI for common law practice will be laughing to transition to working in cyber-copyrighting law. Does an artist really own their voice after it's been genned, twisted, synthetised, auto-tuned and regenerated again in multiple loops? Where's the proof that a voice is stolen?

[@Jack2511]

Kanye AI model link missing?

[@Lonersquare]

Dope gamechanging video! I just tried it and it does work! But how are model files of other famous voices are made and/or found?

[@amsterdamtalkrecords]

yeh i want Barbara Streisand rap like "Onyx - Slam" 🙂

[@aaagaming2023]

You have to train them on a GPU.

[@sixfeetshakur2634]

bro how did you record your voice for it i mean what app did you use for your vocals

[@toddmaek5436]

Damn. Yall remember that Roots song where it was featuring Big Daddy Kane and Kool G Rap with Blackthought, but in actuality it was ALL Black thought rapping?? Welp this is the automated version of that.