Heeey, guys! What do you think about points I made? Did I miss anything? What do you think: is A.I. in music more of a threat or just an opportunity? Lemme know down below!
Totally agree with You but now ppl make release using DistroKid and another distribotion site to release his "own" AI generated music come directly form site like SUNO or UDIO with no editng or no rearragment. For me is completely stupid and this streams service will by overload by AI music against real artist work they made for audience. This is really GIGA problem
„gonna make artists stronger“…… Paul, we live in the Era with the best equipment in history and endless possibilities and yet we manage to stay caught in music that makes no use of dynamic range which is one of the tools with huge potential to express emotion and from a originality point of view most (not all. am talking about the vast majority but not all) music sounds „familiar“ to put it mildly. A.I in music is my least concern from a „composing“ point of view. However in the mastering aspect, i lost a long time client cause they decided to go for automated mastering cause as they said „everyone does“. A year before this they were telling me how happy they are about our work together. Go figure. So either automated mastering is much better than me or their expectations are such as to be satisfied by the state of automated mastering quality at the moment. Opportunity….. Why not. I am sure like always in history that some people will make good use of it. I am more concerned about A.I in other aspects of life than in music tbh. We have managed to damage the quality of the later quite severely without the help of A.I already. Threat…… Maybe ,, but not about music exclusively. Its much bigger than this Or maybe i am getting older too fast. Who knows
I fear we face a period where music companies see only the short-term bottom line of being able to cut out all the expense (the humans) and maximising their profit by filling their streamable music catalogues with AI generated music for which they own all the rights. In due course, this will probably lead to argument 3 (AI is boring because it is, by definition, moving to the mathematical mean) but who knows how long this will take and what kind of "desert" it leaves in its wake?
My issue is the way that a lot of the stuff that's making songs is set up.... With all the emotion and inflection and timing and rhythm stolen directly from actual performances... Intentionally morphing two or more voices together to get something that's not recognizable but stealing performances..... Changing the melody and chord progressions...enough to avoid direct copyright infringement.... But literally stealing people's performances without permission!!!!! That is not okay!
Why speed up? Are we in a hurry and if so why? Artificial Intelligence...sounds good? The best and most beautiful symphonies were written without it. We are just lazy. What the EU is going to use it for will stun us in a few years, used against us in every way.
On a purely technological level, I honestly suspect that although prompt-centric generative AI may persist, the actually powerful tools will be more of an auto-midi program where you can manually compose parts and generate others. For example, if I want a particular leitmotif in a track to represent a character, AI is presently incapable of doing it. I think we'll land somewhere where we have the ability to work more cooperatively and edit the program's work to get exactly what we want. Now, to be clear, I'm not saying that will necessarily reduce financial pressures. There'll still be a struggle to compete against the fully generated content. Just speculation as to where I think the tech itself will go.
On that third point. I watched a video yesterday, where someone else do the kind of concepting I do with ai, and something struck me. Dude wrote this beautifully emotional song, lyrics that would make you cry. It was something like, "I never thought I'd lost the person I thought was my forever," or something like that. Powerful stuff. Then he took his lyrics, put them into ai, and melted them into a song. And, honestly, the song was pretty. The ai version of the song was interesting, but in my opinion, not that remarkable for an ai piece. But dude, then he sang it. He learned the chords and sang the same song, and it melted my brain. Knocked me on my own ass to see him do that. And it's interesting, because you don't think about when you're doing it. And you go through so many hours and so many recordings, and so many renders, that you go sorta get desensitized to the magic that's happening artistically. I really think that's where this tech goes. And where it's most effective. As far as compositioning, and fabrication, I just did a piece on Twitter for an art contest called Ramble. It's me with my weird metaphysical delusions, on top of classy jazz. The entire sound portion of the piece was made with ai. Sound effects, loops, everything except the unsettling breathing loop that runs through the whole thing. I did with suno, udio, faydr, ripx daw, reaper, and stable audio. First stemming out the pieces of the composition, and layering them into place to achieve the specifics of what I wanted for the piece. I could have done the same piece with traditional samples, but it was an ai art contest. So I wanted to show off with how many layers of ai I could integrate into the thing. Turns out, quite a few. Three months from now, there will be some google tools dropping that are going to be better for stamps and loops than the current tools, which are more geared to entire concept compositions. Also, I don't think ai's going to get boring. I still think it shines brightest on the weird stuff. There's a video on my channel called Hot hot hot, where I went full zappa on it. It defies genre and categorization. I've done half a dozen songs like this, but yeah. The exciting thing about ai, I think is how how it lets you explore sound combinations that are either novel, or totally impossible in nature. And that makes for good experimentation, and a cheap harmless way to explore sonic possibilities. Also I'm subscribing to your channel now. Love you Folia!
reposting this answer of mine: give it a little more time! This is really a problem for all artists! I don't think there's much we can do. The music industry was already crappy after Napster..the death of cds, then Spotify royalties and play revenues..now this? it's just getting too much to bear. I wonder why should we keep trying? it's really serious. I bet they're already musicians pumping out lyrics or albums really that they don't have the skills to, and then cover those songs live, I mean who can tell? I struggle with a blank paper to write lyrics, come up with the music and putting stuff together..now you can only be a cover artist ..hell..you don't even have to be a musician at all. Im not against technology, but this crosses the line. And I don't believe there's a law that can control this from happening. They could backup automatically every AI generated music and use the existing means for copyright control scanning..but how about I record the song generated all over on reaper and then upload to CDbaby? this os "cheat gpt" and no way to know. A band can release 6 albums a year or 2 each month while the honest guy will come up with one, maybe two in a year if they're really inspired. Not fair and quite horrifying for us all."
I have been messing with Assisted Idiocy in its various forms since early Expert Systems. All of it has always promised more than it can deliver. That doesn't mean it's not occasionally useful as a tool, and I am using "AI" in my own work (but not to make songs). OTOH much popular music is so formulaic it's arguable if any creativity is actually involved, so if that is replaced by generated forms, who would know or care? I can't see "AI" ever generating anything actually worthwhile. If it does ever produce anything of equal value to Beethoven Bach or the Beatles (substitute your own list of remarkable works) then it is not "AI" but AGI and will rightly demand royalties... (and perhaps have the ability to enforce them! 😉
Currently the musicians in the Udio discord feature request section are requesting the developers to focus on giving good clean stems, give control over key signatures, chords, tempo etc. The ask is to get the tool to be very fine grained in control of what you can get it to do. Once those goals are achieved, the tool becomes extremely useful just like the DAW software apps we use. If AI is to be fully utilized well be existing musicians, it needs to be able to produce everything we ask of it with the end result of our songwriting and production being much faster. yes it gives amateurs an opportunity to create some interesting things, but in the hands of a professional it helps us fine tune and turn out a masterpiece. Remember AI is only as smart as it's developers and it's not self aware or able to learn new things on its own. You Must train it. It's just a more advanced DAW really, with built in samples.
I don't think this will ever go away. I saw (in Japan I think) a Japanimation character being propeller-displayed on stage in front of hundreds of people sing A.I. generated songs in a concert. Once this tech becomes accepted, bar bands will loose their livelihood, front bands too probably, how cool would an A.I. opener for a real band be? Way cheaper too unfortunately, just another piece of stage equipment...
I spent 3 years writing and recording a traditional album nobody ever heard. When Udio arrived, I spent 10 minutes generating a song that went viral. Is music about validation and servicing the ego, or is it about putting something out in the world that resonates with others? The music industry went through this when hip hop introduced sampling. Anyone that ever uploaded their music to a streaming service contributed to the devaluation of music. When songs became “content” and no longer art, the road to where we are now was set in stone.
100% agree with you. The major issue with music is the fact that it's a product more than art. AI isn't an enemy, we already had the worse enemy that killed music as art. Ourself
The problem is this…..whatever is produced, people are always going to ask the question, ‘is this really your talent, or is this purely A?I’……trying to convince somebody that you were really the creator is going to be the real issue because 90% of people will listen and say ‘yeah it’s nice but so what?…..AI can do all that now!’ Peoples ability to write music is going to be totally devalued from now on because ‘anybody can do it!’…..with stunning results that are only going to get better Just create music for the love of it from now on….the days of impressing people with it are over!!
I see 3 scenarios: 1. Humanity powers through this and becomes significantly more productive 2. UBI + the "machines do our jobs" and we do whatever we wish 3. Something in-between and capitalism will create more wide spread tension between stock holders and the replaceable workforce
@@FoliaSound for a given period of time, sure. But that assumption reminds me of the numerous historical talk of "end of history" whereby the tension of previous revolutions only mildly prevail today. There probably will be something **post** this transition
AI art is one of the dumbest inventions ever. Art is expression of human taste, experience, skills and feelings, so I will never listen to AI's fake music. Nobody cares about the people that use AI. And anyway, AI doesn't teach us anything new about music, because the copy is always worse than the original. Once we will have won the lawsuits against the theft of our work, let's move on, and forget it.
I love Ai music as a consumer but that's it. I don't have any interest in listening to ai music unless i generate it myself. Why listen to anything you don't generate? I guarantee that's what most people are doing
Żyjemy w ciekawych czasach choć najciekawszym jest pytanie co z tym będzie dalej i co będzie z nami ludźmi dalej ? Oddajemy coraz to więcej pola do popisu i ingerencji w rodzinę sztukę zaraz AI będzie wydawało wyroki w sądzie słodkiego miłego życia pozbawi cię system wydajności i klasyfikacji hmmm będzie zabawnie
@ 10:37 "... because human-to-human experience is something you will not replace with any machine ...". Imagine a consumer of music LOVES a piece of music but has no idea what band performed or recorded it, OR even if it was produced by an AI. The LOVE of the piece of music will not change one bit if they discover it was created by an AI. Such is NOT a factor in the appeal of music. ... Further on "human-to-human experience": If you are of the mind that says "a human life is a 'life-long' process that can NOT be computed," then you don't fully understand the natures of processes or computation. AI companions of the VERY near future will have the capacity for being configured with whatever HIGHLY-detailed backstory/life-story you want them to have. Already in this waaaaaaaaaay early early stage of truly human-like AI, GPT-4o is pointing the way to what come next. ... Again, thinking that a human life is NOT a computable process is a HUGE mistake that will leave one stuck in the delusion that full human lives cannot be synthesized and convincing that they are as rich as the lives of any real human one may know. Within the next couple of years, watch as HUMAN BONDS to AI companions will be as common as the smartphones they run on and every bit as rich as the experiences had with a real music-producing-human a consumer may have never met or even be aware of who they are. ... The only way to get a solid reading on the future is to do your best to break out of the human-centric-laden subjective view that human-ness is not achievable by a machine. Letting go of that view and instead becoming more objective based on the reality of the progression and trajectory of AI thus far will make one much more likely to see the future as it will be, not as they hope and desire it to be.
If people hadn't paid musicians, albums would never have been made. If people hadn't paid the inventors, electricity wouldn't exist. If people didn't pay software developers, software would never have been made, etc. If all these had not been created, today AI would not be able to learn from music, photos, videos and all the knowledge in the world. So the real question is, what should have been done instead a long time ago? This is what humanity has to face, one by one. Mental development is easier to accomplish when there is something to eat. Because when someone is hungry, the personality type immediately ceases. Then nobody is a musician, or a lawyer, or anything. The definition of the false personality is gone and the true personality is emerging. No one will play music, paint or be artistic. Instead there will be silence. Profound silence. Why? Because personality is then limited to one decision: Who am I if I'm hungry? If you haven't been really hungry, you don't know the answer to this question. This will be the real Truth, when a person comes face to face with his true self...
It looks like AI will be better than Richard Wagner or Arnold Schönberg in couple of years. Unfortunately game is over. Not only for music, also for cinema and literature, for the whole art world.
AI may generate complete song, but it also gets more and more into mixing of conventional made music or analyzing it. It's also present in the development process of plugins or any application. It's everywhere and it makes learning a lot easier. Since years I am in creating AI images. It's a lot of fun and it tends to be addictive. It turns the focus away from the technical side to vision and idea and enables to be in a flow.
Sir AI doesn't work that way, it needs to be trained on millions of examples before it is useful. What you describe is the "fine tuning" but to be able to be fine tune it has to be trained on non related data
I understand this perfectly. See my main video about AI I'm referring to. This one is just presenting the arguments of people who disagreed with me - not my own ones!
Plottwist we are all code, building the simulation inside the simulation, maybe we are all AI, at least mathematically its more likely that you are not the real Paul rather than one of his million simulated verions on the other servers/multiverses. We dont know anything :D
Load in my AI synth, type in the prompt box " the sound of Yaz's Situation", and boom. In my AI amp simulator "the sound of AC DC's Back in Black" taking the tone of my guitar into account (sort of like a microphone modeler thing. Whatever the tone of my guitar, make it sound like that tone). In my Master plugin, "make it sound like David Guetta's "Now Tat the Love is Gone". Leave the tone of my voice intact, but apply all the FX that the original voice has. In other words, make my song sound like some parts of that song, but retain my own personal sound signature. The referenced artists get paid. They sell song "profiles". New business model. New way of sharing knowledge. Everybody wins. F**k yeah. I'm in.
The tech you describe would be awesome, but doesn't exist. There are workarounds already in existence: They are called human ears and musical expertise.
Optimism? Nah, that ship has sailed. Musicians are desperate AF - the super self-absorbed ones are already sobbing their eyes out.. Why bother paying for human music when AI can do it better and for free? Human and AI music will collide soon. AI as a "tool" is so yesterday. Tomorrow, teens will create their own pop stars and rock bands without breaking a sweat. AI companies will make bank until open source takes over. Music, art, and writing will just be decorations - pretty but pointless. People will evolve, and we'll be the old fogies. The future folks will be chemically dependent for happiness. But hey, let's get at least a "replicator" and some supermodel androids, no brains needed - as compensation for killing culture --then I'm good! LOL.
I'm not. I still believe AI is a threat, I just present the opposite perspective. It's good to give voice to both sides and keep distance to yourself. But still: I keep saying that AI is an immense threat, this is my personal view.
Have you considered that the generative AI tools are also useful for those of us with various disabilities? It can close the gap a musician might have and enable them to complete creation of their music. Take for example a guitarist who lost fingers in an accident and could not play anymore. Using Musicfy and other AI models you can go voice to instrument, in this case guitar, and you can sing your melody and have it come out sounding like your guitar. In that same vein if you don't own or cannot play any instrument other than your own voice, now you have access to every instrument on the planet that there is an existent good sample of. Like I love the sound of a sitar but I don't own one so I must use a sample if I can't afford to access a sitar and skilled player. These AI tools level the playing field for every person that wants to get into music and be able to finally express their ideas musically. The era of gatekeeping in the industry is coming to an end.
I believe you are trying to convince yourself, but unfortunately, you are mistaken. AI is going to annihilate studios like yours. The current models may not be the best yet, but they already surpass most human producers in creativity once the audio quality is liberated. Yes, WAV files are already available in some beta models, and in less than a year, those who haven't mastered AI will become obsolete. The quality and speed of work that AI offers are incomparable to what humans can achieve. Publishers and labels are already rubbing their hands at the thought of remixing entire catalogs infinitely without any production costs. As for the ethical debate, in a world driven solely by money and appearances, it too is becoming obsolete. Best of luck to you.
@@FoliaSound AI will take over image , pictures, models, publishing , short videos, films, music, ..., lawyers, cab drivers, .... when congress man and the president is in danger of loosing the job against an AI then and only then something will change.
A.I will vastly improve most people lives, the music aspect is just a small part of it. In my opinion the best course is to accelerate the development.
