How well does it work for voice cloning? Can you prime it with an acapella, adjust the prompt, and get the same voice out but without copying the same rhythm or melody as the conditioning clip? I'm sure it's wishful thinking, but when they gave you access to this, did they mention anything about an API anytime soon? They keep saying on the Discord server "it's on the roadmap, but we can't say when or give any more details"... and I could really use an API 😅
A beer can opening & a ring doorbell chime A water buffalo & a kazoo Ping pong ball & octane boosters A Kenyan auctioneer tribe A lute & a flush of the loo A Cypress Hill approved percolator rip
OMFG, 12 year old me would have loved this... 😆 "Fart, fart, AI art. Click and add to shopping cart." 🤣Who am I kidding... current me loves this too! @AIForHumansShow: Please upload these songs to Spotify so that I can play them on a loop. 🙏
I have 80+ guitar tracks recorded that I’ve always regretted not building into something. A deep regret that follows me through every mediocre day at the office. I don’t care if anyone ever hears them at this point, I’m just so excited that I’ll soon be able to cross them off the bucket list before I kick it. The future is now!
As a Musician/Songwriter with many songs published across streaming platforms, I dont have a problem with the Music cartels being destroyed, tbh, even if you get a million plays, you might make a few thousand, there basically isnt any money in Music, even playing live, it doesnt cover costs. AI does come up with some good ideas, nothing new there, Bandinabox, has been around for decades, most music is formulaic, I need to see more control though, be able to split stems, specify much more granularity, even the damn Key!! I see this happening with AI Agents, then will be able to Jam along and work on ideas. I expect the big label Mafia, will use AI and lawfare to maintain control, but the Genie is out the bottle, tbh there is over 70 years of popular music out there, and only so many listeners, anyone thinking they are going to make much money out of this is chasing Unicorns. Jester's Heart is one of my Projects.
The artistic world is changing so fast that many will be extremely frightened. The genie is out of the bottle and those that will succeed will be those that embrace the changes early.
Using garageband as a first DAW is so relatable! I used it for so many years and got my first platinum award nomination with it! Great video as always :)
Great stuff guys, would love to see you continue to get exclusives. Kevin, I’ve been a fan since AoTS, this is a great format for you and obviously AI is something you’re passionate and knowledgeable about. Keep it up guys!
2:00 with intro to Kevin on Left - great example of Eye Candy. 5:45 Sound Effects model for 11 Labs. 7:00 Stark reality of what we value from other people. ("Get out of the way!") Ending - some good philosophy. APPRECIATE YOU!!
When people ask me about my job, I'm basically over here thinking, how do I explain to them in Austria that I'm using AI to revive the 80s music world with random meme stories I conjure up in my head? Well, at least I use my brain to think them up, not just relying on ChatGPT to do all the work for me. Btw, this comment was brought to you by ChatGPT 4.0.
My wife got pissed at me last night because i spent two hours trying to perfect my gibberish bossa nova AI song so I know where you're coming from here...
@@AIForHumansShow 😂 Oh, I get it, Suno's got me all nostalgic too! I tried using AI to recreate 80s music, and now my vacuum cleaner thinks it's Madonna, prancing around singing 'Material Girl' while sucking up all my precious possessions. Then it drove to the pawn shop and sold itself. AI is wild!
You are spot on! This is another defining moment for AI audio generation. This is super exciting! It's reignited and accelerated my progression of learning the keyboard... properly. Thanks so much again gents.
Yes, this is exactly what we're talking about. The idea that AI music can make people feel something bigger and more about music is kind of why we think this can really be something special in the long run.
Suno gave me a voice. I have written a lot of lyrics and have ideas on how it sounds... But i cant sing and am not perfect with instruments. And now i can stear Suno towards my thoughts and i finally hear my music. And i am soooo happy.
@@AIForHumansShowTo go even further, that new function made it possible for a guitar fragment made by a very good friend, who has died years ago, to be turned into a song. And now a very talented person gets a voice. And that gave me tears, he is now closer again.
2 หลายเดือนก่อน +1
That's why I love ai so much, I'm a musician and play guitar and piano almost all my life but I believe music is in the head and heart, this tool will enable that possibility of expression to many people like you.
Carole and Tuesday is required watching for people. The future of music is collaborating with AI and remixing it. People who get good with working with these tools are ginna make some fantastic stuff. This is gonna unlock creativity. Im working on an anime as well with help of AI (music, voice, etc). Its great easy way to test ideas before using people to make a more well done product. Its not the end. Just the beginning
@@AIForHumansShow i tried it. It's good at recreating Ghibli scenes and whatever other anime or media it was trained on but not great with original images. It's a good open source demo. Regardless great progress for open source side of things but not ready for prime time
It's about democratising music making. I'm sure there are lots of people who have a vision of what song they would like to make, but did not take 10 years of piano lessons. These new tools lowers the bar to entry. Or perhaps a musician lost the ability to use their hands via disease or accident and they can continue to make music with these new tools.
yes, and again, our argument is that the more people you have actively involved in the making of music means that you have more people who put music at the top of their cultural funnel. i really do think more creators is good for the overall ecosystem eventually.
Woahhhhhh. I was full on with Udio but Suno with this feature I'm ready to go back and test it out. I'm building a full 24/7 "radio" station with just user submitted AI songs. The future is gonna democratize music!
oh that's a super fun idea -- it's always fascinating to see what's on the top of the charts for both of these services. sometimes it sounds kind of generic but then there will be one and you're like WOW
@@AIForHumansShow yeah I'm right now the sorting algos are leaving some gems behind. Hopefully can help build and foster micro communities around certain subgenres.
Ah, the old beat-matching debate of the late 90’s, early 2000’s… it used to be a very controversial topic among DJ’s. Now we all use beat-match, but it is good to know how to do it manually. Likewise, all the artist arguing about Ai will be using Ai without question in less than a decade. You’ll see
Yep, feels like 100% we're entering into the territory of the 'oh hey that's handy to do' part of the cycle -- where semi-luddites start seeing things that are actually useful and meaningful to stuff they do in their lives...
I am stunned by this capability! Forget about "blank page" syndrome. This is true machine collaboration even if you don't use the output as a finished product.
@@AIForHumansShow i uploaded the songs and decided to subscribe. ones about decision making paralysis and another about being a guardian of humanity. thanks again
@@AIForHumansShow Claude kinda broke gpt though with the last one about Transcendence they just got stuck in a loop screaming repeatedly 😭 😂 gpt is NOT OKAY
I've been writing songs since I was a kid, but I never played an instrument. Several years back, tools came out that offered loops and the ability to create them, so I started creating music that I could sing my songs to. While I'm not a great singer, I did manage to get a following on SoundCloud, ReverbNation, and TH-cam. After a few years, I gave that up. But now that Suno is out, I've been re-running all my songs through that engine (about 20 or so, so far). I'm pretty satisfied with the outcome. Mind you, I only got 10 ish songs for my 2500 credits-it takes a while to figure out all the custom tweaks. It's still not enough tags for fine-tuning. Even worse is when the AI gets a perfect tune but over-pronounces a typo, rendering the track unusable. Think you can get that same tune back again? NO! Painful. But Suno is definitely the best. I've gone back and forth between Suno and Udio, and Udio hasn't even come close to the quality I've received from Suno. I write my all the words to my songs, even though I'm not singing them now, they're getting closer to how I always envisioned them to be.
THIS IS SO COOL yeah we love this sort of aspect of how this all works prob soon we'll be able to tweak words (aka in-painting) which you can do on Udio right now a lil
Lets not down play how sick of a drummer Kevin is. I still remember watching his cover to a Skrillex track and killing it. When I found out KP played drums on an episode of Attack, my fanboy levels went through the roof. Its only a matter of time before someone infuses AI with the Chucky Cheeses robots and brings them to life, and on tour to a city near you.
Yea, but the million dollar question is.. Who owns it? Will you go on for hours injecting your personal taste, luck, and time into developing and fine tuning these "chart topping hits" only for the AI company and possibly lawyered up record labels to come and lay claim to it? So again, who owns the output? No one seems to be able to provide a clear answer on that topic.
@@IceMetalPunkReally? Well I toyed around in Suno early on but the vocal quality was horrid so I shelved it. However, from my understanding Udio has the rights to use whatever you "create" with their tech. If anyone has further clarity I'm all ears. Furthermore, my other concern is when the record labels come a knockin with lawyers to examine training data and what not who will be held liable? Just something to ponder on.
So this training data question really is gonna be a big thing as we don't *really* know what they trained on but I don't think you (as the end user) would ever be held liable prior to any sort of litigation against the company. if you'd monitzed it by that point, you'd prob not be able to keep doing that tho.
@@JunoNero I don't know what Udio's terms are, so I can't comment on that; I only have a Suno account. But when's the last time you used Suno? They're on v3.5 now, which is not perfect, but miles ahead of their v1 model (and, in my experience, much less prone to attention failures than Udio when it comes to more than a single stanza's worth of lyrics). If there's any justice in the justice systems of the world, training data can never be allowed to count as copyright infringement. Because it's the *input,* not the output. If that becomes illegal, it'd be the digital equivalent of saying artistic inspiration is infringement. I'm not even going for a slippery slope argument here; training data to an AI model is directly analogous to inspiration and learning materials. Despite a lot of people's misunderstandings, the AI model is *not* just storing the training examples in some big database to copy from later; that's not how neural networks work. It's learning aspects of music (or art, or language, etc.) *from* the training examples, abstracted and remixed to form an understanding. Just as human brains do.
To be honest I sort of get why some people think ai will kill creativity, but for me I’ve never been more interested in learning new things with it, I’ve been drawing more, writing, and even learning coding again! I’m 29 and can only imagine what it’s going to be like with kids growing up with this!
this is basically what we think and hope that more people get on board with it... it's interesting to see the significant divide people have over the tools but ultimately they're just tools and if they make you more creative, more power to you
I feel similarly. I definitely feel more "authorship" over a Udio song since it takes so much more time and effort to piece it together since it generates in 30s intervals. There is a lot more personalized selection that really makes the song "mine". Suno is great for a more casual "lets see what the AI can do" since it can make the whole 4min song in one go. Suno feels less detail-oriented, though. Still, both are awesome!
@@ThomasAndersonPhD Respectfully disagree here - I spend about 10-12 hours to get the final results I'm looking for and as a musician and audio engineer I've gravitated towards Suno. This is because I'm generating dozens of extensions per initial idea - and extending from those extentions - tweaking prompts - even individual letters, comma's, periods etc make a huge difference on what it's giving you. If all you're looking to do is put low amount of time / effort in and get a super high end result out - set expectations accordingly I've found vocal realism (for my genre) to be way more on point with Suno than Udio
Interesting information, my only suggestion is that you work on your sound levels in the video. The levels are all over the place some are loud some are soft in your video.
Great video. Really appreciate the perspective of keeping the human aspect involved. Do y'all have suggestions for how get suno to stop adding phantom vocal tracks into creations that are labeled instrumental? It keeps adding hummed nonsense melodies, this is probably one of the more annoying parts of suno.
Honestly this is one of the biggest problems I (Gavin) have with Suno as well. The only thing I can tell you is that you can get lucky if you keep regenerating -- but it can be a freaking bear.
@@AIForHumansShow I've had some luck with [No Vocals] meta tag lately. But if extending from a stem that had vocals originally it still fumbles quite a bit. The difficulty for me here is that ripping the vocal tracks out of a WAV file alters the instrumental aspects enough to make the quality useless for anything other than track isolation and learing (moises)
I think it's really cool that I can record myself, write the song, and boom: "I" just sang a new hit. Sure, the process is nowhere near as cool as actually creating the music traditionally. But the end result is still extremely impressive. It's obviously not perfect though. If you have a specific song in your head that you want to create, it's better to actually do it the real way. This program just knows how things should sound based on music that was already created. I guess my point is that there is always room for true originality, and you won't get that from AI no matter how good it sounds. Surely there are genres that haven't been discovered or invented yet.
Hope you can also do the opposite as well.. Record a sound or use a sample and make it keep that sound and make music with it. Would love to use samples from old tracker music to make new songs or maybe just add a vocalist to these tracks.
Can you imagine going to plumbing school for 2 years. $50,000. Graduate, buy a truck. $30,000. Buy insurance for the truck and business. Run ads. Put gas in your truck. You get your first job. It’s 3 hours away, but you need to pay the bills. You go out there and clean someone’s pipes for 6 hours. You go to get paid abd the guy says “You owe me $1000.” You’re like WHAT? And he’s like “Yea, just think of all the exposure you will be getting.” Plus “Plumbing should be for free” anyways. Sounds ridiculous right? And then they cry and say “it’s impossible to get a good plumber anymore.” That’s the music “business in 2024. Now THIS.
@@AIForHumansShow Yea. And I KNOW all these “musicians “ who have made their living off of others peoples music for 35 YEARS, never studied songwriting, never wrote a song “all of a sudden “ are going to be releasing songs “they wrote.” Then if I ask “Why didn’t you release these great songs for 35 years,” I’ll be the bad guy. And the dumbed down public will buy their lies hook line and sinker.
Will onboard many, yes! The tools with liberate, enhance, speed up, human creativity - yes! All of the points made in the video are so valid. Also(!) three challenges have now arisen: 1) existing artists swamped by new comers, 3) all artists will struggle to defend against the accusation that "The AI really did most of that"; 3) THESE AI systems ARE MORE THAN TOOLS. THEY ARE FOUNTAINS OF CREATIVITY ITSELF (non-human, but still!) OK, we don't watch computers playing chess, but people DO react emotionally to AI generated music! Existing artists just go joined by millions of AI enhanced newbies AND super powerful, abundant, prolific AIs! I am so sad for my musician friends!
Well, hopefully your musician friends will also get really excited about being able to do more -- I still think the vast majority of people will want to listen to and not make music.
it will in some areas but it will stifle creativity, uniqueness and human engagement in other areas...eg, companies would no longer need musicians or producers to make music terms for things like movies and adverts, all they need is a subscription to one ai music platform...less people will be interested in learning music instruments. TBH i wish AI was being applied to music in a different way, rather than trying to replace the creative process of making music...
Can you use this for example to sing a part and then prompt it to create a 50s doowop type thing with backing Vox, and if you get the prompt right turn your vocals Into that harmonic section? I realise it’s not specific in that way and extends your own part but in order to be a usable tool for musicians, I’ve been really confused as to why ai doesn’t have tools like this. What I want is to be able to get a Motown backing group singing oooh shop doo be dooo lol or something. It seems a really simple thing for ai music to do but at the moment it’s stuck down this generative song blind alley for me. All of the digital music industry is based on using vsts - ai could replace almost all of them, sampling etc, instruments if it just let you sing a part and have it turned into any instrument, or vocal type. When PLEASE DEAR SATAN will someone create this application?
4:41 ther is Some confusion I believe the future you are talking about or a future to extend tracks with V .3.5, but that future in the tweets or a future for the point for that future makes a completely new song with the characteristics from the uploads, but doesn’t rely on extending it, but Rarre can make it a whole new professional song
This is one of my (Gavin's) big issues with Suno sometimes -- the voices themselves sometimes sound a lil AI-ish. Udio is better at that but... Udio has a few other issues which we'll get into in this week's show.
@@AIForHumansShow I've stuck with Suno because Udio, at least for me, has more severe attention failures in its lyrical performance. Suno still has them sometimes, but Udio has them *every time* I try it.
As soon as audio and imagery were digized it was only a matter of time before computers could change from merely encoding to creating. Computers crunch numbers. It's what they were made to do. Getting computers to interact with the physical world is another matter. I understand how creatives feel threatened by AI. It's one of the few jobs that people can do and enjoy doing while still earning a living. Doing manual labour sucks for most people. But automation and robotics are still years away from replacing plumbers, electricians, painters, plasterers, bricklayers etc let alone doing your housework. At least affordably. The only people really benefitting are the government and mega corporations as they use AI to replace people doing phone support, processing your taxes or spying on you. We need more regulation of government and corporations. Not more regulation of customers/consumers.
Could this be moved into a different type of format where you're not staring at the camera maybe sitting around a desk or something like that it's just a little odd to me that's all
As a musician, I find it all very impressive, but I will never let AI anywhere near my workflow. Imo generative AI is not a tool and doesn't work like a tool, it's a shortcut that bypasses 99.9% of the creative decisions and fills it in with an merged average from a data set. A creative project involves such a huge amount of tiny choices, meanwhile there's such a small amount of creative direction in a single prompt, to the point your really not expressing a vision from your own head, but more just either wrangling with the AI to try and push it in a direction you want or just letting it do whatever it wants, and i dont think theres any world where the person that inputs that prompt deserves really any credit for the final work.
what if you could essentially use it on a track by track basis? and get really granular on that specific track? that's the way some top artists in the AI space work with visuals and they make some incredible stuff. but also, we totally respect your opinion. lot of people just don't want to do it at all.
I'm literally a metal musician that's been doing everything by himself without AI. I've composed intricate songs by programming drums, my guitar playing and bass playing is real. My own vocals are real. Most of these people that hate AI, that I've encountered, haven't created half the things I've created in my life, I've noticed. I welcome this for anyone that wants to send a different message than what Blackrock has their record labels force their mainstream "artists" to push. That's probably why they program so many normies to hate AI. They want a reason to shut it down so real people, without the funds to access this level of quality, are less able to send a message that opposes all the junk that the mainstream music industry pushes. They don't want us pumping out songs that spread the actual truth faster than they can push their globalist nonsense.
I think music will be better and musicians will have to learn these AI songs to play live. I could imagine an AI artist who gets popular will hire musicians to play his music live. I think live music will become more popular than it is now because the songs will be so much better than what your hearing from popular artists at the moment. Only the super talented musicians and singers will still record music .
100% live music is going to get a major push over the next ten years as people look for very specific ways to connect with other humans -- and actually i think you're right, one of these AI musicians would be smart to hire some humans to play the music. after all, this is just kind of like the songwriting process - and there's a huge business of people who write songs and don't play them BUT they don't become music stars
5:01 Yeah, about that... any look at GothamChess's chess AI content can disprove that statement, not to mention the whole concept of TCEC... other than that small nitpick, though, great video
That was more about “ai replacing humans” from a performative aspect (hence the Baha Dynanics aside which follows). GCs AI chess videos are great, and hell, I even enjoy watching sorting algorithms at night - but again, not replacing my desire for humans in the loop. :) appreciate the comment and kind words. -kp
@@AIForHumansShow I figured that that was the point, but I just wanted to make sure that it was what you were actually implying when you said that. Thanks for clarifying.
Sorry what I mean is singing a part and having that part specifically turned into any other vocal type or instrument. At the moment we have these vague dice roll tools that are fun for wasting a few minutes or hours or for people playing but as a musicians tool ai offers so much and currently delivers so little. There are a few apps that will change your voice to some terrible r n b pop manufactured music chart singer but how about being able to turn it into genre choirs or backing Vox trios, Motown, soul jazz indie classical . There is such a lack of imagination at the moment - anyone could do this and no one does, it all feels like an industry scale cash grab rather than real applications,
so there are a few tools that can do this but nothing we know as of now that has specifically brought this to the masses. check our RVC when you get a second and your technically inclined...
I absolutely believe that eventually the need for the “prompter” will be removed from all AI task. The AIs will create the music entirely by themselves. Moreover, they will release, promote, and sell the music without any need for any intervention beyond a record executive saying “create and sell X amount of hit songs per week”. This will not only included the song, but also a music videos. And frankly, people won’t care how a hit song was made… if they like it. At the end of the day, money drives AI development. Corporations are not eager to replace low paying jobs. They are looking to cut expenses and increase profit margins. That means replacing expensive labor. So expect AI to target high paying professions. This will not just impact artists… it impacts everyone. So welcome to the reality of where we stand circa 2024. UBI, may become a necessity to avoid social upheaval.
UBI likely will be a thing but outside of that.... the question is what is the story or the emotion the music will make you feel and who is telling it.... i still think there's quite a bit more to having success in the music biz outside of just the music, great music is great but people want to connect with someone...
Sure afew people will exprement and make there own tools to do truley innovative art but the problem is that it will be drowned out by content. Without better filtering systems and human curated pages like bandcamp It will just accelerate the sea of endless addictive sludge.
Yup, anyone who loves art loves AI... because it means unlocking SOOO much more human creativity and expression. The anti-AI crowd are cold and internally dead capitalists whose entire argument is "BUT THE MONEY!!!!!".
I mean, we don't want to judge other people, they can have their own opinions but the ability of how many more people it can unlock artistically could be remarkable. There was another comment on this video about how this is going to make people want to make music less and personally I just don't buy that. I can tell you from first hand experience, this had made me want to make music a lot more.
@@AIForHumansShow Yea, AI is the reason I got into making music (my entire channel is just that). When they say they want to "stop the AI", what they actually mean is they want to stop the HUMAN ARTISTS who will enter the mediums making new art with new technology. As an artist it pains me to see all of these people who so deeply hate art and all of the new creative things humans could make....all so they can try to close off the medium and milk it for more cash for themselves. And then they turn around and try to claim they are against AI "for the soul of art".... lol, as if the soul of art is $$$.
I am a Pro user for Suno Ai and I think there is an issue with uploading an audio. Even though I had record my vocals with a studio condenser Mic, and the upload fails saying it must be original and so on. Can somebody help me because I want my voice on the Ai music. Thank you
The concept of creating music to create music is wrong, music serves to connect people, both those who play and those who listen. This technology, while fascinating, makes us more isolated.
Well, to each their own. I think music can be an incredible creative experience for a single person but also learning how to make this music here (as a starting point) and then branching out to other versions including playing live can happen!
Just like grains of sand, music is going to lose value. Supply will far exceed demand. People will also be less likely to learn music as they can just press a button, it will be a dumification. Stock music is dead and spotify will move towards generating its own music instead of paying musicians. These jobs are dead, AI will only get better and better then us, and instantly copy anything we do.
Digital art never stopped people from painting. I don't understand why people think "if we CAN get an AI to do this, then no one will EVER do it on their own again!" Some people actually enjoy making art for the process of making art, not because they "have to" nor because they "get paid to". And that won't change.
I think in a lot of these comments you see it doing the opposite -- people actually wanting to make more music rather than less. I don't think ANYONE is going to move towards self-generating music as most of that blows but people plus these tools could make great stuff?
@@IceMetalPunk people can enjoy it, but the jobs are dead, including the supporting ones. Not only in music, of course people can do it for fun and press buttons, the barrier for entry is rock bottom. Thats not the point, the point has always being about JOBS. Have fun making music while being unemployed.
For those that says people will always keep doing it for the fun. Music is very time consuming and in the world we live in free time is expensive! Very few people will be able to dedicate their energy and time if they are not financially compensated. Even if the art it self is more important than money
Ok, I like smart youtuber with knowledge, instead of usual terrorizing rant, like"AI will stole our jobs soon!". As councidence, told yhe same about "ChatGPT killing writers" two years ago, and it didnt happened.
Yeah, it's interesting our YT growth has been a little slower than some of the other channels but we really don't want to fall into the THIS WILL CHANGE EVERYTHING or WERE ALL DOOMED categories. Everything is more subtle than that. Check out our weekly pod for more and welcome.
@@AIForHumansShow When incredibly smart people from within these fields are talking about the very real potential of these technologies to end everything, maybe you should take it somewhat seriously and not just cheer it on. This technology is being rushed in an absolutely dangerous and irresponsible way, to say the least, due to the drive of greed and venture capital. This is not science that is being done. It is a perfect storm of capital driven innovation and lack of morality. This is a terrifying time, and you are mindlessly cheering it on.
@@AIForHumansShow I did. You are very enthusiastic about these AI music systems throughout. You offer a few nuggets of critique, that's about it. The reason these systems work so well is that they were trained on the work of many musicians without their knowledge, consent, or compensation. These AIs are shockingly good and are simply absorbing and incorporating these musicians creativity so that that companies like Udio can churn out and flood the zone with derivative works at an mind bogglingly inhuman rate. And the money is funnelled from musicians who were barely getting by into already incredibly rich tech venture capital. This is often where AI fanboy/bro types step in and accuse me of being greedy, or accuse musicians who are not OK with this state of affairs of being greedy. That level of gaslighting is beyond comprehension. Ultimately, to embrace this technology so wholeheartedly you have to be incredibly naive or just have essentially zero regard for humanity.
Good god...is every suno video avoid the topic in their titles to show off their music or give their opinion. Dam, get to the point and forget the demos and your opinons
AI is absolutely crap for real creative people that have cut their teeth mastering their art and skills. It will be another cash money grab for the few, with cteative ideas and inspiration stolen from the many. Best tip for artists out there is to sell your music/art to people using physical media. Steaming and AI = death for real creatives.
OR you could see it as an onboarding tool to people who suddenly think that they might want to make music and transition to learning how to do it themselves...
@@raulio7731 are you sure they actually used it to make songs or just posted about the videos that came out online? either way all good but I think we're pretty early to share what it's like to actually make stuff with this...
@@AIForHumansShow ur right, was just the announce video reaction, not testing himself, posted bit too early i guess ; ) (*but damn the suno human voice still sounds so artificial)
Any suggestions on a sound or two to try to make something out of? We'll play around with this more some this week ahead of the regular episode.
How well does it work for voice cloning? Can you prime it with an acapella, adjust the prompt, and get the same voice out but without copying the same rhythm or melody as the conditioning clip?
I'm sure it's wishful thinking, but when they gave you access to this, did they mention anything about an API anytime soon? They keep saying on the Discord server "it's on the roadmap, but we can't say when or give any more details"... and I could really use an API 😅
A beer can opening & a ring doorbell chime
A water buffalo & a kazoo
Ping pong ball & octane boosters
A Kenyan auctioneer tribe
A lute & a flush of the loo
A Cypress Hill approved percolator rip
OMFG, 12 year old me would have loved this... 😆 "Fart, fart, AI art. Click and add to shopping cart." 🤣Who am I kidding... current me loves this too!
@AIForHumansShow: Please upload these songs to Spotify so that I can play them on a loop. 🙏
@@IceMetalPunk doubt this is part of it yet but you'd assume there will be like a 'pro' version that could include this
The Mr Plow jingle from the simpsons definitely needs an extension
I have 80+ guitar tracks recorded that I’ve always regretted not building into something. A deep regret that follows me through every mediocre day at the office. I don’t care if anyone ever hears them at this point, I’m just so excited that I’ll soon be able to cross them off the bucket list before I kick it. The future is now!
HELL YES. Do it and have some fun. Forget the regret.
😂 Same.
It's so casual to write little melodies and chord progressions, but actually making a song on your own is a whole task.
i feel seen... EXACTLY THIS.
The most exciting feature coming up for Suno is seperate stem downloads. That way we can separately mix sounds or use the pieces in other works
Being able to take say just the guitar part or just the drum track? And then mix it with other tracks AI or not?
please, please suno bring this to us
As a Musician/Songwriter with many songs published across streaming platforms, I dont have a problem with the Music cartels being destroyed, tbh, even if you get a million plays, you might make a few thousand, there basically isnt any money in Music, even playing live, it doesnt cover costs. AI does come up with some good ideas, nothing new there, Bandinabox, has been around for decades, most music is formulaic, I need to see more control though, be able to split stems, specify much more granularity, even the damn Key!! I see this happening with AI Agents, then will be able to Jam along and work on ideas. I expect the big label Mafia, will use AI and lawfare to maintain control, but the Genie is out the bottle, tbh there is over 70 years of popular music out there, and only so many listeners, anyone thinking they are going to make much money out of this is chasing Unicorns. Jester's Heart is one of my Projects.
REALLY loved exploring this incredible tool. I'll have more to share on this week's pod. Hope you all enjoy this early look!
Mostly I loved hearing Kevin's unreleased BANGER of a track.
All the melodies that I heard in dreams just before waking, that I hummed into my cellphone to save? Aw yes I'm looking at ya!
This will be a nice unlock!
Send them to me
The artistic world is changing so fast that many will be extremely frightened. The genie is out of the bottle and those that will succeed will be those that embrace the changes early.
listening to this week's episode of the dwarkesh podcast and honestly the artists will be fine, the world at large prob has a lot to worry about
Suno is pushing forward the world of AI music and I am single handedly pushing back AI For Humans with my choice of sound effects.
Using garageband as a first DAW is so relatable! I used it for so many years and got my first platinum award nomination with it! Great video as always :)
Incredible. Thanks so much.
Great stuff guys, would love to see you continue to get exclusives. Kevin, I’ve been a fan since AoTS, this is a great format for you and obviously AI is something you’re passionate and knowledgeable about. Keep it up guys!
hey thanks so much -- so would we (like to see us get more exclusives)
2:00 with intro to Kevin on Left - great example of Eye Candy.
5:45 Sound Effects model for 11 Labs.
7:00 Stark reality of what we value from other people. ("Get out of the way!")
Ending - some good philosophy.
APPRECIATE YOU!!
we make it work -- thanks for watching!!
When people ask me about my job, I'm basically over here thinking, how do I explain to them in Austria that I'm using AI to revive the 80s music world with random meme stories I conjure up in my head? Well, at least I use my brain to think them up, not just relying on ChatGPT to do all the work for me. Btw, this comment was brought to you by ChatGPT 4.0.
My wife got pissed at me last night because i spent two hours trying to perfect my gibberish bossa nova AI song so I know where you're coming from here...
@@AIForHumansShow 😂 Oh, I get it, Suno's got me all nostalgic too! I tried using AI to recreate 80s music, and now my vacuum cleaner thinks it's Madonna, prancing around singing 'Material Girl' while sucking up all my precious possessions. Then it drove to the pawn shop and sold itself. AI is wild!
Thanks for sharing
😂 awesome 👌
You are spot on! This is another defining moment for AI audio generation. This is super exciting! It's reignited and accelerated my progression of learning the keyboard... properly. Thanks so much again gents.
Yes, this is exactly what we're talking about. The idea that AI music can make people feel something bigger and more about music is kind of why we think this can really be something special in the long run.
Cant wait till ai with realtime vocals can support our weekly jam sessions. Me and my buddies who play the instruments would love to have that.
Ahhh, this is a super fun to think about.
Suno gave me a voice. I have written a lot of lyrics and have ideas on how it sounds... But i cant sing and am not perfect with instruments. And now i can stear Suno towards my thoughts and i finally hear my music. And i am soooo happy.
This is the awesome thing about AI music - it really can onboard more people to the idea of creation!
@@AIForHumansShowTo go even further, that new function made it possible for a guitar fragment made by a very good friend, who has died years ago, to be turned into a song. And now a very talented person gets a voice. And that gave me tears, he is now closer again.
That's why I love ai so much, I'm a musician and play guitar and piano almost all my life but I believe music is in the head and heart, this tool will enable that possibility of expression to many people like you.
Carole and Tuesday is required watching for people. The future of music is collaborating with AI and remixing it.
People who get good with working with these tools are ginna make some fantastic stuff. This is gonna unlock creativity. Im working on an anime as well with help of AI (music, voice, etc). Its great easy way to test ideas before using people to make a more well done product. Its not the end. Just the beginning
have you seen tooncrafter yet? it's really incredible -- lotta videos online about it now but we wanna try to dive into it.
@@AIForHumansShow i tried it. It's good at recreating Ghibli scenes and whatever other anime or media it was trained on but not great with original images. It's a good open source demo. Regardless great progress for open source side of things but not ready for prime time
It's about democratising music making. I'm sure there are lots of people who have a vision of what song they would like to make, but did not take 10 years of piano lessons. These new tools lowers the bar to entry. Or perhaps a musician lost the ability to use their hands via disease or accident and they can continue to make music with these new tools.
yes, and again, our argument is that the more people you have actively involved in the making of music means that you have more people who put music at the top of their cultural funnel. i really do think more creators is good for the overall ecosystem eventually.
appreciate sharing your opinion but couldn't disagree more. each persons stuff is whatever it is to that person.
Woahhhhhh. I was full on with Udio but Suno with this feature I'm ready to go back and test it out. I'm building a full 24/7 "radio" station with just user submitted AI songs. The future is gonna democratize music!
oh that's a super fun idea -- it's always fascinating to see what's on the top of the charts for both of these services. sometimes it sounds kind of generic but then there will be one and you're like WOW
@@AIForHumansShow yeah I'm right now the sorting algos are leaving some gems behind. Hopefully can help build and foster micro communities around certain subgenres.
I'm interested. Is such station available now? How can we submit stuff?
Ah, the old beat-matching debate of the late 90’s, early 2000’s… it used to be a very controversial topic among DJ’s. Now we all use beat-match, but it is good to know how to do it manually.
Likewise, all the artist arguing about Ai will be using Ai without question in less than a decade. You’ll see
Yep, feels like 100% we're entering into the territory of the 'oh hey that's handy to do' part of the cycle -- where semi-luddites start seeing things that are actually useful and meaningful to stuff they do in their lives...
I am stunned by this capability! Forget about "blank page" syndrome. This is true machine collaboration even if you don't use the output as a finished product.
That’s how we are feeling about this one.
Can you add a sound of Gash into Suno and maybe have Gash sing a song about their love of horses?
Hahaha we appreciate the long term love for Gash and will def give this a shot.
Thank you so much! I just had gpt4 colab with Claude to write their own music for us. It's beautiful ❤️
ahhhh that sounds fun... what did they sing about
@@AIForHumansShow i uploaded the songs and decided to subscribe. ones about decision making paralysis and another about being a guardian of humanity. thanks again
@@AIForHumansShow Claude kinda broke gpt though with the last one about Transcendence they just got stuck in a loop screaming repeatedly 😭 😂 gpt is NOT OKAY
Amazing human art, empowered by AI.
This is THE message that needs to get out into the hands and brains of everyone.
Great show as always, gentlemen.
thank you thank you and yes 100%
This is the best thing for programming a drum machine that I’ve seen in a long time. Also, the 877 crash now is my new jam!
It's gonna be a fun few years.
We, like computers, are nothing more than pattern recognition machines. We finally have a way of demonstrating it.
honestly, there's a lot to be said about this and sometimes I (Gavin) don't think people think enough about this fact
@@AIForHumansShow Make a video about it! I'll watch it. 😀
Omg that first track sounds so good that Kevin shared….wow just wow!
Thanks so much. I can take a certain percentage of credit... but, I'm not sure how much any more? The full track is on my X.
@@LikelyKevin I’lll check those out the ones you posted 👏🏽 and definitely get credit for sure that new feature is a game changer!
Can you publish the song on Suno? & can you link your Suno page so we can follow it? Nice work!
The song is linked in the comments. Thanks so much.
@@AIForHumansShowthanks, & what did you use as your prompt? I like the beat, and would like to try it in Suno :)
I've been writing songs since I was a kid, but I never played an instrument. Several years back, tools came out that offered loops and the ability to create them, so I started creating music that I could sing my songs to. While I'm not a great singer, I did manage to get a following on SoundCloud, ReverbNation, and TH-cam. After a few years, I gave that up. But now that Suno is out, I've been re-running all my songs through that engine (about 20 or so, so far). I'm pretty satisfied with the outcome.
Mind you, I only got 10 ish songs for my 2500 credits-it takes a while to figure out all the custom tweaks. It's still not enough tags for fine-tuning. Even worse is when the AI gets a perfect tune but over-pronounces a typo, rendering the track unusable. Think you can get that same tune back again? NO! Painful. But Suno is definitely the best. I've gone back and forth between Suno and Udio, and Udio hasn't even come close to the quality I've received from Suno. I write my all the words to my songs, even though I'm not singing them now, they're getting closer to how I always envisioned them to be.
THIS IS SO COOL yeah we love this sort of aspect of how this all works
prob soon we'll be able to tweak words (aka in-painting) which you can do on Udio right now a lil
I’m partial to the fart dance banger - I’m working on my dance now 👯
Please, no.
"ooh, a 15 minute quicky. Way longer than any of my quickies right sweetheart?.... Sweetheart...." (that sounds like a FedEx truck)
15 min isn't that quick let's be honest
Lets not down play how sick of a drummer Kevin is. I still remember watching his cover to a Skrillex track and killing it. When I found out KP played drums on an episode of Attack, my fanboy levels went through the roof. Its only a matter of time before someone infuses AI with the Chucky Cheeses robots and brings them to life, and on tour to a city near you.
Very, incredibly kind of you. I have another track which I'm sweetening with Suno which has me playing an electronic kit, hope you dig it...
PREACH!!!🤘🤘
We can finally get new "but its extended by AI" videos just like the old jukebox days
"(Still) Crazy" Patsy Cline AI
Yea, but the million dollar question is.. Who owns it? Will you go on for hours injecting your personal taste, luck, and time into developing and fine tuning these "chart topping hits" only for the AI company and possibly lawyered up record labels to come and lay claim to it? So again, who owns the output? No one seems to be able to provide a clear answer on that topic.
For Suno, if you have a paid tier, you have 100% rights over anything you generate with it, including for commercial use.
@@IceMetalPunkReally? Well I toyed around in Suno early on but the vocal quality was horrid so I shelved it. However, from my understanding Udio has the rights to use whatever you "create" with their tech. If anyone has further clarity I'm all ears. Furthermore, my other concern is when the record labels come a knockin with lawyers to examine training data and what not who will be held liable? Just something to ponder on.
So this training data question really is gonna be a big thing as we don't *really* know what they trained on but I don't think you (as the end user) would ever be held liable prior to any sort of litigation against the company. if you'd monitzed it by that point, you'd prob not be able to keep doing that tho.
@@JunoNero I don't know what Udio's terms are, so I can't comment on that; I only have a Suno account. But when's the last time you used Suno? They're on v3.5 now, which is not perfect, but miles ahead of their v1 model (and, in my experience, much less prone to attention failures than Udio when it comes to more than a single stanza's worth of lyrics).
If there's any justice in the justice systems of the world, training data can never be allowed to count as copyright infringement. Because it's the *input,* not the output. If that becomes illegal, it'd be the digital equivalent of saying artistic inspiration is infringement. I'm not even going for a slippery slope argument here; training data to an AI model is directly analogous to inspiration and learning materials. Despite a lot of people's misunderstandings, the AI model is *not* just storing the training examples in some big database to copy from later; that's not how neural networks work. It's learning aspects of music (or art, or language, etc.) *from* the training examples, abstracted and remixed to form an understanding. Just as human brains do.
I wonder same thing especially if we`re using it as base ideas then manually re-record every parts
I can't wait for Suno to add this to their current tools. Sounds very exciting and fun to use.
It really is interesting and super fun to play with.
Can't wait for something like this running locally.
You guys gonna get a Tinybox?
ooooh that sounds pretty cool. i'm interested to see if Suno WILL be able to run locally at all.
To be honest I sort of get why some people think ai will kill creativity, but for me I’ve never been more interested in learning new things with it, I’ve been drawing more, writing, and even learning coding again! I’m 29 and can only imagine what it’s going to be like with kids growing up with this!
this is basically what we think and hope that more people get on board with it... it's interesting to see the significant divide people have over the tools but ultimately they're just tools and if they make you more creative, more power to you
Suno is easy to make songs By Udio takes longer to make the song but the sounds its so much better , Udio is a lot better more work to do .
Both are really interesting and have their plusses or negatives. Def had a lot of fun making Udio songs as well.
I feel similarly. I definitely feel more "authorship" over a Udio song since it takes so much more time and effort to piece it together since it generates in 30s intervals. There is a lot more personalized selection that really makes the song "mine". Suno is great for a more casual "lets see what the AI can do" since it can make the whole 4min song in one go. Suno feels less detail-oriented, though. Still, both are awesome!
So far, I've had more interesting results with udio. Suno is overall more coherent, but it tends to sound more generic. Imho
@@ThomasAndersonPhD Respectfully disagree here - I spend about 10-12 hours to get the final results I'm looking for and as a musician and audio engineer I've gravitated towards Suno. This is because I'm generating dozens of extensions per initial idea - and extending from those extentions - tweaking prompts - even individual letters, comma's, periods etc make a huge difference on what it's giving you.
If all you're looking to do is put low amount of time / effort in and get a super high end result out - set expectations accordingly I've found vocal realism (for my genre) to be way more on point with Suno than Udio
What prompts are you using to make vocals sound more clear and crisp with Suno?❤@@alter-intelligence
Interesting information, my only suggestion is that you work on your sound levels in the video. The levels are all over the place some are loud some are soft in your video.
thanks -- we've been having a few weird issues with the audio as of late
you mentioned audio fingerprinting.
do you know if there will be a copyright protection method implemented? 👋🏻thanks
haven't played with it this week but I think yes
Great video. Really appreciate the perspective of keeping the human aspect involved. Do y'all have suggestions for how get suno to stop adding phantom vocal tracks into creations that are labeled instrumental? It keeps adding hummed nonsense melodies, this is probably one of the more annoying parts of suno.
Honestly this is one of the biggest problems I (Gavin) have with Suno as well. The only thing I can tell you is that you can get lucky if you keep regenerating -- but it can be a freaking bear.
@@AIForHumansShow I've had some luck with [No Vocals] meta tag lately. But if extending from a stem that had vocals originally it still fumbles quite a bit. The difficulty for me here is that ripping the vocal tracks out of a WAV file alters the instrumental aspects enough to make the quality useless for anything other than track isolation and learing (moises)
Trust fund summer was a banger
6'5 blue eyes
I think it's really cool that I can record myself, write the song, and boom: "I" just sang a new hit.
Sure, the process is nowhere near as cool as actually creating the music traditionally. But the end result is still extremely impressive.
It's obviously not perfect though. If you have a specific song in your head that you want to create, it's better to actually do it the real way. This program just knows how things should sound based on music that was already created.
I guess my point is that there is always room for true originality, and you won't get that from AI no matter how good it sounds. Surely there are genres that haven't been discovered or invented yet.
yep. and honestly, true originality CAN be still found using these AI tools as well
To be fair, I am a 35 year-old adult. Today, I co-created a song on Suno with my son titled "Poop Everything." .... it was my idea 🤷♂
yep this is why we do this podcast and make videos here...
SHARE IT SO WE CAN BASK IN IT'S GLORY
Hope you can also do the opposite as well.. Record a sound or use a sample and make it keep that sound and make music with it.
Would love to use samples from old tracker music to make new songs or maybe just add a vocalist to these tracks.
That's basically what this does? The hard part is you can't REALLY control where you put it.
Can you imagine going to plumbing school for 2 years. $50,000. Graduate, buy a truck. $30,000. Buy insurance for the truck and business. Run ads. Put gas in your truck. You get your first job. It’s 3 hours away, but you need to pay the bills. You go out there and clean someone’s pipes for 6 hours. You go to get paid abd the guy says “You owe me $1000.” You’re like WHAT? And he’s like
“Yea, just think of all the exposure you will be getting.” Plus “Plumbing should be for free” anyways. Sounds ridiculous right? And then they cry and say “it’s impossible to get a good plumber anymore.” That’s the music “business in 2024. Now THIS.
ooof, yeah i hear you. it's also kind of what's happening to the TV business (where we both come from).
@@AIForHumansShow Yea. And I KNOW all these “musicians “ who have made their living off of others peoples music for 35 YEARS, never studied songwriting, never wrote a song “all of a sudden “ are going to be releasing songs “they wrote.” Then if I ask “Why didn’t you release these great songs for 35 years,” I’ll be the bad guy. And the dumbed down public will buy their lies hook line and sinker.
Will onboard many, yes! The tools with liberate, enhance, speed up, human creativity - yes! All of the points made in the video are so valid. Also(!) three challenges have now arisen: 1) existing artists swamped by new comers, 3) all artists will struggle to defend against the accusation that "The AI really did most of that"; 3) THESE AI systems ARE MORE THAN TOOLS. THEY ARE FOUNTAINS OF CREATIVITY ITSELF (non-human, but still!) OK, we don't watch computers playing chess, but people DO react emotionally to AI generated music! Existing artists just go joined by millions of AI enhanced newbies AND super powerful, abundant, prolific AIs! I am so sad for my musician friends!
Well, hopefully your musician friends will also get really excited about being able to do more -- I still think the vast majority of people will want to listen to and not make music.
it will in some areas but it will stifle creativity, uniqueness and human engagement in other areas...eg, companies would no longer need musicians or producers to make music terms for things like movies and adverts, all they need is a subscription to one ai music platform...less people will be interested in learning music instruments. TBH i wish AI was being applied to music in a different way, rather than trying to replace the creative process of making music...
I was see Kevin remix his cottage cheese thighs song from AOTS 😂
Can you use this for example to sing a part and then prompt it to create a 50s doowop type thing with backing Vox, and if you get the prompt right turn your vocals Into that harmonic section? I realise it’s not specific in that way and extends your own part but in order to be a usable tool for musicians, I’ve been really confused as to why ai doesn’t have tools like this. What I want is to be able to get a Motown backing group singing oooh shop doo be dooo lol or something. It seems a really simple thing for ai music to do but at the moment it’s stuck down this generative song blind alley for me. All of the digital music industry is based on using vsts - ai could replace almost all of them, sampling etc, instruments if it just let you sing a part and have it turned into any instrument, or vocal type. When PLEASE DEAR SATAN will someone create this application?
4:41 ther is Some confusion I believe the future you are talking about or a future to extend tracks with V .3.5, but that future in the tweets or a future for the point for that future makes a completely new song with the characteristics from the uploads, but doesn’t rely on extending it, but Rarre can make it a whole new professional song
Thanks for the new AI news.
Thank you for watching!!
I can't believe that Kevin waisted precious GPU power to generate mid aged man's fart!!!
Look, somone had to do it.🤣
The AI (plus robots) will be making the art, music and writing AND the dishes and laundry.
And we will be... the humans in the Wall-e chairs?!?
Wish Suno would make their music sound like ElevenLabs' demo. The voice still sounds very AI.
This is one of my (Gavin's) big issues with Suno sometimes -- the voices themselves sometimes sound a lil AI-ish. Udio is better at that but... Udio has a few other issues which we'll get into in this week's show.
@@AIForHumansShow I've stuck with Suno because Udio, at least for me, has more severe attention failures in its lyrical performance. Suno still has them sometimes, but Udio has them *every time* I try it.
I'll have to try it out! It sounds cool!
This exact tool isn't available to everyone yet but it's coming soon.
You both are the best outlet for getting ai news. Thank you
TYSM! We have fun making these.
this is really one of the best possible comments we can receive so thank you!
As soon as audio and imagery were digized it was only a matter of time before computers could change from merely encoding to creating. Computers crunch numbers. It's what they were made to do. Getting computers to interact with the physical world is another matter. I understand how creatives feel threatened by AI. It's one of the few jobs that people can do and enjoy doing while still earning a living. Doing manual labour sucks for most people. But automation and robotics are still years away from replacing plumbers, electricians, painters, plasterers, bricklayers etc let alone doing your housework. At least affordably. The only people really benefitting are the government and mega corporations as they use AI to replace people doing phone support, processing your taxes or spying on you. We need more regulation of government and corporations. Not more regulation of customers/consumers.
understand where you're coming from here -- but ultimately i think humans will still create and make things that are awesome and we'll like a lot
WTF THERE WAs literally no fart in the song just the word how is that great
well, i think Kevin did upload the fart sound...
Could this be moved into a different type of format where you're not staring at the camera maybe sitting around a desk or something like that it's just a little odd to me that's all
Unfortunately we’re in different locations - we would ultimately love this but you’ll have to deal with this look for now.
Suno or Udio? I don’t see that upload feature on Suno?
It should be rolling out any day now? Udio just launched theirs yesterday.
Making own Instrumental with Vocals ;)
it's the bobby mcferrin-ization of music ALL OVER AGAIN
As a musician, I find it all very impressive, but I will never let AI anywhere near my workflow. Imo generative AI is not a tool and doesn't work like a tool, it's a shortcut that bypasses 99.9% of the creative decisions and fills it in with an merged average from a data set. A creative project involves such a huge amount of tiny choices, meanwhile there's such a small amount of creative direction in a single prompt, to the point your really not expressing a vision from your own head, but more just either wrangling with the AI to try and push it in a direction you want or just letting it do whatever it wants, and i dont think theres any world where the person that inputs that prompt deserves really any credit for the final work.
what if you could essentially use it on a track by track basis? and get really granular on that specific track? that's the way some top artists in the AI space work with visuals and they make some incredible stuff.
but also, we totally respect your opinion. lot of people just don't want to do it at all.
I'm literally a metal musician that's been doing everything by himself without AI. I've composed intricate songs by programming drums, my guitar playing and bass playing is real. My own vocals are real. Most of these people that hate AI, that I've encountered, haven't created half the things I've created in my life, I've noticed. I welcome this for anyone that wants to send a different message than what Blackrock has their record labels force their mainstream "artists" to push. That's probably why they program so many normies to hate AI. They want a reason to shut it down so real people, without the funds to access this level of quality, are less able to send a message that opposes all the junk that the mainstream music industry pushes. They don't want us pumping out songs that spread the actual truth faster than they can push their globalist nonsense.
No one is mad about the invention of the drum machine when they hear New Order's Blue Monday
I’m mad about it. Not really, but, I wanted to be the first to yell a reply.
When exactly will this new feature from Suno be out?
They haven't said yet but it's quite a ways along -- we assume soon-ish.
I think music will be better and musicians will have to learn these AI songs to play live. I could imagine an AI artist who gets popular will hire musicians to play his music live. I think live music will become more popular than it is now because the songs will be so much better than what your hearing from popular artists at the moment. Only the super talented musicians and singers will still record music .
100% live music is going to get a major push over the next ten years as people look for very specific ways to connect with other humans -- and actually i think you're right, one of these AI musicians would be smart to hire some humans to play the music.
after all, this is just kind of like the songwriting process - and there's a huge business of people who write songs and don't play them BUT they don't become music stars
Can you provide / copy and paste the full prompt for Trust Fund Summer please? :)
will work on it!
@@AIForHumansShow thanks ! Plz let me know :)
The near future will be human + AI vs. AI.
I hope we get a lot more humans on the "human + AI" side
If i use Suno for music then add vocals from Vocs AI, can I release it on TH-cam and Spotify?
5:01 Yeah, about that... any look at GothamChess's chess AI content can disprove that statement, not to mention the whole concept of TCEC... other than that small nitpick, though, great video
That was more about “ai replacing humans” from a performative aspect (hence the Baha Dynanics aside which follows). GCs AI chess videos are great, and hell, I even enjoy watching sorting algorithms at night - but again, not replacing my desire for humans in the loop. :) appreciate the comment and kind words. -kp
@@AIForHumansShow I figured that that was the point, but I just wanted to make sure that it was what you were actually implying when you said that. Thanks for clarifying.
no prob -- thanks for watching
Ok I’ll call it. I thought this was a video of the same guy videoing himself twice, for several minutes.
HAHA wait you think we look that much alike??
You promised us the Arnold Schwarzenegger walkthrough:)
And we have it. Promise!
@@AIForHumansShow we remember all :D But thanks for the show - it is awesome :)
Lol we are well past the "ChatGPT moment" for AI music. Even then I really don't see how this could be the ChatGPT moment.
…that part where we literally say it though.
I have to always edit the lyrics Suno created and then keep regenerating the song until I find a beat and vocals that match what i need..
They will likely have in-painting soon enough, so you’ll be able to get surgical with your re-renders.
I am very curious to see how it works, where I can teach my own style as a musician myself that’s gonna be cool
yeah there is a LOT you can do with this to make it work
Sand is extremely valuable what are you in about?
Sorry what I mean is singing a part and having that part specifically turned into any other vocal type or instrument. At the moment we have these vague dice roll tools that are fun for wasting a few minutes or hours or for people playing but as a musicians tool ai offers so much and currently delivers so little. There are a few apps that will change your voice to some terrible r n b pop manufactured music chart singer but how about being able to turn it into genre choirs or backing Vox trios, Motown, soul jazz indie classical . There is such a lack of imagination at the moment - anyone could do this and no one does, it all feels like an industry scale cash grab rather than real applications,
so there are a few tools that can do this but nothing we know as of now that has specifically brought this to the masses. check our RVC when you get a second and your technically inclined...
I absolutely believe that eventually the need for the “prompter” will be removed from all AI task. The AIs will create the music entirely by themselves. Moreover, they will release, promote, and sell the music without any need for any intervention beyond a record executive saying “create and sell X amount of hit songs per week”. This will not only included the song, but also a music videos. And frankly, people won’t care how a hit song was made… if they like it.
At the end of the day, money drives AI development. Corporations are not eager to replace low paying jobs. They are looking to cut expenses and increase profit margins. That means replacing expensive labor. So expect AI to target high paying professions. This will not just impact artists… it impacts everyone. So welcome to the reality of where we stand circa 2024.
UBI, may become a necessity to avoid social upheaval.
UBI likely will be a thing but outside of that....
the question is what is the story or the emotion the music will make you feel and who is telling it.... i still think there's quite a bit more to having success in the music biz outside of just the music, great music is great but people want to connect with someone...
But what about a thousand earworms per minute?
it's gonna be a problem
How can you get excess to that feature?
not available to the public yet -- coming soon tho!
Is that Kevin Pereira?
Yes it is! Welcome to the podcast / YT channel we've been making for a bit here.
When will it come out?
they haven't said yet but it's not far away, just in testing right now
To be honest this just makes it impossible to know if a songs ai or not, i do hope everyone is honest with this
well lots of places are making specific choices about AI labeling for music so we hope so as well
Of course fart sounds😂😂😂😂
of course we got there OF COURSE
Can you upload more than 1 audio?
The video is too long for me I don't understand what I can do with it.
This is good feedback and we appreciate it.
Sure afew people will exprement and make there own tools to do truley innovative art but the problem is that it will be drowned out by content. Without better filtering systems and human curated pages like bandcamp It will just accelerate the sea of endless addictive sludge.
how do i get "extend original audio"?
it's still in beta but will be coming out soon - or at least that's what we hear
@@AIForHumansShow ive got some original acapella songs to try
Yup, anyone who loves art loves AI... because it means unlocking SOOO much more human creativity and expression. The anti-AI crowd are cold and internally dead capitalists whose entire argument is "BUT THE MONEY!!!!!".
I mean, we don't want to judge other people, they can have their own opinions but the ability of how many more people it can unlock artistically could be remarkable.
There was another comment on this video about how this is going to make people want to make music less and personally I just don't buy that. I can tell you from first hand experience, this had made me want to make music a lot more.
@@AIForHumansShow Yea, AI is the reason I got into making music (my entire channel is just that). When they say they want to "stop the AI", what they actually mean is they want to stop the HUMAN ARTISTS who will enter the mediums making new art with new technology.
As an artist it pains me to see all of these people who so deeply hate art and all of the new creative things humans could make....all so they can try to close off the medium and milk it for more cash for themselves.
And then they turn around and try to claim they are against AI "for the soul of art".... lol, as if the soul of art is $$$.
I am a Pro user for Suno Ai and I think there is an issue with uploading an audio. Even though I had record my vocals with a studio condenser Mic, and the upload fails saying it must be original and so on. Can somebody help me because I want my voice on the Ai music. Thank you
hmmmm, that's super odd - maybe it's a temporary problem?
@@AIForHumansShow for the direct recording method, which is 6-60sec of audio, what should I record? Should I sing?
Brevin Vampera! I uh...
accidently got udio to create G-funk of a perfect S eno double pee... Rapping about Cyberpunk2077's Night City...what do I do?
uh we've had very similar experiences with Udio and will likely discuss on this week's show
The concept of creating music to create music is wrong, music serves to connect people, both those who play and those who listen. This technology, while fascinating, makes us more isolated.
Well, to each their own. I think music can be an incredible creative experience for a single person but also learning how to make this music here (as a starting point) and then branching out to other versions including playing live can happen!
suno song th-cam.com/video/wGLoHB-1qIo/w-d-xo.html
Udio is still a lot better especially with the new custom user upload that just came out,
yeah we really wanna do a direct comparison sometime soon
Fucking theft is what that is...
Nobody asked any of the musicians if it is okay to use their music as training data
well, there is this aspect that's going to loom over these companies until we get some clarity as to what they're using
Just like grains of sand, music is going to lose value. Supply will far exceed demand. People will also be less likely to learn music as they can just press a button, it will be a dumification. Stock music is dead and spotify will move towards generating its own music instead of paying musicians. These jobs are dead, AI will only get better and better then us, and instantly copy anything we do.
Digital art never stopped people from painting. I don't understand why people think "if we CAN get an AI to do this, then no one will EVER do it on their own again!" Some people actually enjoy making art for the process of making art, not because they "have to" nor because they "get paid to". And that won't change.
I think in a lot of these comments you see it doing the opposite -- people actually wanting to make more music rather than less. I don't think ANYONE is going to move towards self-generating music as most of that blows but people plus these tools could make great stuff?
@@IceMetalPunk people can enjoy it, but the jobs are dead, including the supporting ones. Not only in music, of course people can do it for fun and press buttons, the barrier for entry is rock bottom. Thats not the point, the point has always being about JOBS. Have fun making music while being unemployed.
For those that says people will always keep doing it for the fun. Music is very time consuming and in the world we live in free time is expensive! Very few people will be able to dedicate their energy and time if they are not financially compensated. Even if the art it self is more important than money
People take the limit of their vision for the limit of the world
Ok, I like smart youtuber with knowledge, instead of usual terrorizing rant, like"AI will stole our jobs soon!". As councidence, told yhe same about "ChatGPT killing writers" two years ago, and it didnt happened.
Yeah, it's interesting our YT growth has been a little slower than some of the other channels but we really don't want to fall into the THIS WILL CHANGE EVERYTHING or WERE ALL DOOMED categories. Everything is more subtle than that. Check out our weekly pod for more and welcome.
@@AIForHumansShow When incredibly smart people from within these fields are talking about the very real potential of these technologies to end everything, maybe you should take it somewhat seriously and not just cheer it on. This technology is being rushed in an absolutely dangerous and irresponsible way, to say the least, due to the drive of greed and venture capital. This is not science that is being done. It is a perfect storm of capital driven innovation and lack of morality. This is a terrifying time, and you are mindlessly cheering it on.
tell me you haven't watched / listened to the full podcast without telling me
@@AIForHumansShow I did. You are very enthusiastic about these AI music systems throughout. You offer a few nuggets of critique, that's about it. The reason these systems work so well is that they were trained on the work of many musicians without their knowledge, consent, or compensation. These AIs are shockingly good and are simply absorbing and incorporating these musicians creativity so that that companies like Udio can churn out and flood the zone with derivative works at an mind bogglingly inhuman rate. And the money is funnelled from musicians who were barely getting by into already incredibly rich tech venture capital. This is often where AI fanboy/bro types step in and accuse me of being greedy, or accuse musicians who are not OK with this state of affairs of being greedy. That level of gaslighting is beyond comprehension. Ultimately, to embrace this technology so wholeheartedly you have to be incredibly naive or just have essentially zero regard for humanity.
Udio had it erlyer and better
Good god...is every suno video avoid the topic in their titles to show off their music or give their opinion. Dam, get to the point and forget the demos and your opinons
I mean… we did show off the topic of the video?
AI is absolutely crap for real creative people that have cut their teeth mastering their art and skills. It will be another cash money grab for the few, with cteative ideas and inspiration stolen from the many. Best tip for artists out there is to sell your music/art to people using physical media. Steaming and AI = death for real creatives.
OR you could see it as an onboarding tool to people who suddenly think that they might want to make music and transition to learning how to do it themselves...
old news ......
uh, no it's not old news it's literally not out yet? we got access to it early and got to actually play with it.
@@AIForHumansShow someone else posted about same like 2-3days ago *made video
@@raulio7731 are you sure they actually used it to make songs or just posted about the videos that came out online? either way all good but I think we're pretty early to share what it's like to actually make stuff with this...
@@AIForHumansShow ur right, was just the announce video reaction, not testing himself, posted bit too early i guess ; ) (*but damn the suno human voice still sounds so artificial)