A.I. - Time To Throw In The Towel???

[@contemplativecomposer5983]

Christian, I always appreciate how you go into the nuances on your videos, the philosophy of time, thought, ideas developing etc. It brings in both the creative side and theoretical side of the conversation, producing a broad perspective and is just a pleasure to listen to. I am very much a lover of philosophical thought, so its a refreshing journey to go on :)

[@ChrisIzattOfficial]

Brilliant video Christian! Thanks for sharing such a well-rounded perspective! I completely agree with you about getting back to virtuosic music! You've said in previous videos: "Get the F Off the computer" & I think that goes deeper than surface-level screen staring!

[@acdms]

Christian, it's always nice to listen to you.

[@williammonk1020]

Something every composer should have on their radar if they're thinking about AI: the way language models are currently trained involves a lot of finetuning with curated datasets in order for the models to behave in a particular way. Those datasets are then 'owned' by someone or some organisation. That's important because better quality data = better quality performance, so that training data is extremely valuable (and expensive). Also, if you have very specific data, you can tune the model to your preferences to respond in any particular manner, with any specific information etc. It seems realistic that the same sort of thing will happen with music, whereby the big players in every area (labels, DAWs, virtual library developers, producers etc.) will invest heavily into producing the highest quality models or models specifically tuned to their style. If you're a film composer, at some point inevitably, you will want (and eventually need) to have your own curated data which can be fed into a model to produce AI generated material that is adapted to your style. How this is used would vary but you could even imagine film producers 'renting out' your model to test concepts in advance, just to give one example (there are many fun/scary possibilities). I mean, if your works are out there in public (even if they aren't in the public domain), those can and are being used for pretraining, but that is not high quality training data. So yeah, 'turn this tune into a hans zimmer piece' is not going to give you the results you want just because the model's base training has heard his scores.

[@jakobymaster]

The scariest moment for me was, when messing around with some of the generators that are still in beta right now, i had to admit to myself that it was good and i liked the music. And it was no pop or anything "4 by 4", it was contemporary classical, emulations of morricone, japanese classical etc.. Scary scary stuff

[@juno6]

Try Choirs! you´ll be 🤯

[@ABC-bm7kl]

It ‘creates’ modern orchestral, chamber, and small ensemble “art” music extremely well. I have also been making songs for my friends and have woken up on several occasions with these songs running through my mind. Udio in particular is impressive. If you have not yet tried it, write some lyrics and generate a Broadway show-tune. It is preposterously well versed in that idiom.

[@tonus6559]

I’ve been there with the Udio show tunes. I am really struggling to believe that it can come up with these chord sequences and melodies based of text alone. I can imagine it having multiple pre-imagined examples of show tune chord progressions (templates) and then having stems that can be pitch and time shifted… But being a composer and having virtual instruments I can’t believe that all of that stuff is modelled and mixed and produced so quickly without having access to some kind of stems? Of source material? Is anyone with me on this? I’ve heard the best virtual guitar instruments in Kontakt and they are nowhere near what I’ve heard in Udio.

[@ABC-bm7kl]

@@tonus6559 agreed! The sounds in Udio, except for the unintended? “artifacts” and intentional vinyl or tape noise that it adds, are remarkably good. I’m a composer also and have, no exaggeration, almost ALL the sample libraries. The ‘live’ feeling of the performances in Udio is uncanny. And the idea that the sounds are being generated from modeling is hard to imagine considering that other “modeled’ libraries don’t sound anywhere near as real. Wild times!

[@Edbrad]

@@tonus6559it’s hard to believe but it’s hard to believe it can with LLM’s like chatGPT, OpenAI’s Sora is hard to believe. When I see realistic images generated by Midjourney for example it’s really hard for my brain to NOT thing they’re not actual living people

[@mu6ix.musics]

Such a refreshing video! Amidst all the recent doom and gloom, Christian consistently manages to bring a positive perspective. I especially appreciated how the announcement of the new library was saved for the end. This decision reflects Crow Hill's confidence in their products, unlike other companies that often push their libraries down our throats from the very start. Great job, Christian!

[@zillasaurus]

I love this take. I could listen to you pontificate all day. And I agree with everything so far. Good stuff - much respect from across the pond.

[@LoveMeBack]

Because of Computer music tons of studios closed down. Now there are more studios then ever and hardware is loved again. Even tape. We will be fine.

[@agentmith]

AI differs from previous advancements in that it removes the foundational creative necessity from the artist. Granted, someone could dial something in aesthetically, but that’s like doing touch ups on a painting versus a blank canvas and a paintbrush+palette.

[@mranchovydrumbass]

The foundational creative necessity of the artist comes from the artists life experiences and lies deep inside an artists soul which could never be replaced by AI

[@Edbrad]

It’s also because the way the creative AI develop is completely in reverse to other technology. Look at how CGI developed. There was total control, but it wasn’t very good. AI gets better much faster at making a good result BEFORE it gets good at given you control. Image AI’s for example. It can give you a great result, but it still needs improvement in giving you: 1. Exactly what you asked for specifically, & 2. How MUCH specific detail you can prompt it for. If you can give it up to 5 details and it always gets the right, there’s always room for more details. And there’s always a need to change it microscopically afterwards easily. 3. Being able to consistently take objects, people, characters and anything “form” you can think of and carry it over into another image.

[@LeeBlaske]

You make an important point. Just because music professionals survived previous technological innovations doesn't mean they'll survive this one. AI is very different.

[@mranchovydrumbass]

Lol, why has my reply been taken down ? was that AI or Christian ?

[@danvorosmarty9854]

It doesn't differ as much as you'd think though at first glance. The kind of results you can get by giving generic prompts are... generic results. To get anything of actual use or value, You have to use the tool in much more complex and indeed creative ways. I have a question/challenge that touches on this idea: What is the actual specific functional difference between prompting an AI tool to say generate an image, and Beethoven writing a symphony? I would suggest that there's much less of a difference than we think... Because think about it this way. Beethoven learned a special language (musical notation), which was then given to an orchestra (a tool outside of himself) which essentially "automatically" translated his instructions into music that people then listened to. In other words, in this metaphor musical notation = prompt syntax/engineering, and the orchestra = the AI tool. Who actually "makes" the music in this case? Beethoven? Or the people who play the instruments in the orchestra? Are they not behaving as an "automated" kind of machine to translate the composer's instructions into tangible art? If we apply the same thinking most people apply to generative AI currently, then we'd have to conclude that the orchestra is doing the creative part of producing the symphony, NOT Beethoven. My point is though that of course we all agree Beethoven is the creative force/element in the equation, NOT the tools he used to produce the tangible result and NOT the specialized language he utilized to translate his vision into reality. Just saying, Beethoven literally couldn't DIRECTLY produce (most of) his art because it required an orchestra by definition. "All" he actually was doing was writing down symbols on paper. So again, can someone tell me how that's technically/functionally/fundamentally different than the act of using AI tools to create art? (I know "common sense" tells us it's different but I'm questioning that notion because it seems extremely difficult to actually come up with real reasons why it's different. As far as I can tell it's not actually different other than that it's people playing the instruments in the orchestra... Yet still, they're meant to act as automatons and play what the sheet music tells them to.)

[@nicolwilson9320]

I will always be here to play and make music with fellow musicians, There is nothing better !!

[@PDJMDS]

I agree. Been saying that AI will bring back the live music culture. People will want to see music played, by people, in a room or field with other people.

[@ZakRoams]

totally agree. I think we are already at the beginning of this due to social media.

[@geektronic1852]

Beautifully articulated Christian, may human made imperfect music long continue to bloom and unfold into the future

[@mus3equal]

Thank you for the motivation, always gives me a little push to work harder on things. Couldn't agree more with much of what you covered, especially regarding education around creativity. We need to foster eachothers individual creative strengths so that we can learn and adapt from that as well, style may shift with time but what makes us creatively us is only definable by the individual until it is shared then it morphs with others time and experience, sometime it echos back positively and that is the signal we should all reflect and amplify back into all creative realms.

[@uniqornzmusic1020]

I've just found your channel. I love it! These are definitely the types of conversations we need to be having right now in my opinion and delivered in such a hilarious fashion. I've subscribed Cheers Crow Hill!

[@2fioreto]

Always a pleasure to hear your thoughts on the important questions in music composition. I record all of my music live from top to bottom - every instrument and cue is followed and improvised on for the duration of a piece, which takes loads of time (I love every minute). There are imperfections, as I choose not to follow a click track. The flow is unique and can seem strange or ‘uncanny’ and I wouldn’t have it any other way. This is my own special brand of contemporary music after all - organic, human, and filled with the guitars that I want to sing into the world. AI is a solution to a problem that has been solved for ages. Blood, sweat, and tears. I hope my neighbors can find a joy in the work required to make meaningful Art. Cheers mate, and wishing you and the rest of Crow Hill Team the best 🤘🏼

[@crazydigitalmusic]

I’ve been watching almost every scientific video about AI for the last 2 years and reading tons of articles, but this video is one of the best explanations ever! Congratulations, dear music friend.

[@Thegregcoulson]

I like your work and outlook. You make the world a more comfortable place. Thanks

[@stevesm2010]

I love your idea for the CPUless studio! I volunteer in a local school running an after school club (for electronics construction) and I'm constantly amazed at the breadth of the kids musical tastes. Much more varied than when I was that age 45 years ago LOL

[@mk1st]

A friend of mine worked in various LA studios back in the 80s. He kept some of his gear and just sold a n 8 track multitrack machine for big bucks. The buyer drove 10 hours to pick it up.

[@bassManDavis1953]

Nice talk Christian, Ive been playing and learning Bass guitar all my life and yes I use a computer to help me write my stuff! But I always go back to the natural feel of music and I teach this way as well, I'm happy to say that the next generation that I have had the pleasure to teach want it that way.....guitar solo's, bass offbeats and everything that goes with it.

[@BoogieBear]

More food for thought, thanks Christian 🙏

[@qbass00]

This is spot on, and it mirrors exactly the same conversation those of us who work in the creative industry have been having about graphic design / illustration.

[@AllanGildea]

Thought provoking and insightful commentary, Christian. Thank you.

[@steven190]

Thank you Christian, for your ever important insight. A true treasure

[@YoPaulieMusic]

"A return to virtuosity." I FREAKING LOVE THIS AND WELCOME IT! Thank you for such a rational viewpoint.

[@thepanicroommastering2062]

Thanks for sharing your thoughts Christian ✨✌️

[@Edbrad]

Thanks Christian! Finally someone of importance actually embracing this technology. Because like it or not, it’s here and there’s no way to put it back in the box. I’m actually excited that you guys might be one of the first to actually make some cool stuff with AI. At some point a GOOOD open source music AI will come out. You’ll be able to train it on anything (whether one finds this abhorrent or not) and do all kinds of things when you use it. GoogleIO’s “Music AI Sandbox” with their Lyria model showed you can take beatboxing and singing and turn it into drum performances and orchestra/saxophone etc. The possibilities will be endless. And I hope more devs like you embarrass the development of this technology because I want composers and sound designers to be the ones dreaming up new, finely tuned and well directed tools in imaginative and creative and musical ways.

[@bontempo1271]

Henson gets it, very well said, nice encompassing talk on the subject

[@AndyGrayedout]

My take on tech (and I started my journey long before computers were in studios ) is learn it understand it , and make it work for you -:) great vid Mate

[@DensDigitalDen]

Great video and an interesting topic! I think you've raised some really good points about the potential impact of AI on our future. AI's impact is a double-edged sword, but I'm optimistic. Automation can displace, but also create new opportunities. Ethical development is key to a positive future. Let's embrace the change and ensure we steer it responsibly. Thanks for sharing!

[@Yarbro45]

Interesting thoughts. Thank you for sharing. Appreciate your insights from the music industry.

[@simplisticscientists]

I did recently start to get a glimpse of an idea that I could use the AI in such a way that it enables me to create music. Like a sort of band manager. I always used to feel like I needed a manager you know. and I heard a few artists say the same thing... so imagine having a robot that is programmed to sound like your favourite manager sat next to you the whole time y'know!

[@deveyousness]

I love Trance music, its all about the harmonies and melodies which feel otherworldly. The beat is just secondary to that really. AI seems to have the ability to see patterns in things we as people might not be able to see, so I don't believe that AI will only ever be imitation, once it really picks up. 🤖

[@shawnpcampbell]

Thanks Christian for yet another amazing video! Like alot of composer and artists in general, Ive been worried about this topic and Ive heard a great many opinions on the topic. Youve covered alot of what my fellow classmates and my mentors have been discussing over time. General consensus so far is that human made music is definitely going to be around for a long time still...but the previous barrier for entry skillwise is going to have to rise. People (especially in videogames like myself) are going to have to learn to adapt and raise the bar in our virtuosity in order to compete. Not the worst thing in my mind but for some people they might be panicking right now with the pace at which this tech is advancing. I think above all though...its SUPER important that right now us musicians and artists band together to speak with one voice at the negotiating table. I feel we fumbled the ball when the record companies disparaged and underestimated entities like Apple or Spotify and let THEM dictate how much artists should be paid. We have a chance again to make our voices heard this time and push back against letting non artists dictate how our art is valued or paid for. Lets make this different...not just for us but for the artists that follow after us yeah?

[@LeeBlaske]

Sadly, I don't see any scenario where music creators will be speaking with one voice at any negotiating table. There is very little organization in the community. And, the only way pressure can be applied in such a situation is to actually DENY the use of music. The music producing community is too large, diverse and deployed worldwide. If a certain group of producers tried to play hardball and deny the use of their music, others who weren't part of that group would simply see it as their big chance to get some traction. There was a reason the music community didn't join in during the big SAG/AFTRA and Writers' Union strike not that long ago. I think people realized it would have been futile. If push came to shove, there was already tons of music in licensing libraries that would keep content creators going for years.

[@shawnpcampbell]

@@LeeBlaske Im not disagreeing with you...but SAG/AFTRA wouldnt be an organization today unless SOMEONE started it. The Screen Actors Guild started in the 30's amidst the depression with two people. Just two. Thats all it took and now it counts its members in the hundreds of thousands. Someone has to start it. You? Me? Someone else? Why not here? Why not now? There are already some loose organizations for video game composers and songwriters trying to do exactly that. Now is not the time to give up. Now is the time to communicate, organize and make sure we DO have a voice. Start small with your local community and musicians and see just how quickly it can grow!

[@LeeBlaske]

@@shawnpcampbell SAG/AFTRA and the AF of M came about in the grand old days of big studios where all of the work being done was highly observable and easy to monitor. When the work being done can essentially be done by anyone in the world, anywhere in the world on a laptop or even an iPad, there's no way to observe the work being done, and no way to keep a lid on things (EVERY session could be a "dark date"). In order to have any impact, an organization would need to have the clout to bring everything to a standstill if members decided to strike. I don't see any possibility for that.

[@aaronnewberry]

I always love your perspective, thank you for sharing ❤

[@el_arte]

You ramble beautifully. Like me. And you almost look like me. I have been in tech for 30 years almost, in software mostly. I have always been compelled (others made me do it in exchange for money; often they were tasks not fit for humans) to remove humans from processes. It’s only fair AI would remove me from processes now. Producing real art is a different matter; that’s work worthy of a human. That’s the one part we should keep.

[@Zaccyonline]

❤❤❤ Many of my favourite personal production moments are the result of happy accidents, or deliberately going against conventional wisdom. I am comforted to know I have a superpower against Ai, possessing something it can never have; being HUMAN. ❤❤❤

[@BartvanderHorst]

"tv screens take your souls they said' and we laugh. But if you take people from that era to this time... I bet they will panic and believe they were very right.

[@Lunsterful]

Great video. I think you (and many others) are underestimating the inventiveness of AI. There's no magic inside our brain, at least, not of a kind that cannot emerge from the strange mind of a different medium. We're overestimating the technical difficulty of creativity, while simultaneously reducing something very complex and interesting to "just statistics". The truth is a strange thing in-between.

[@zacharywoodard2097]

ROFL That Class A Drug comment was PRICELESS!!!

[@GrantTregellas]

For the last 20 years we've seen the rise in music tech. Started simply with making life easier (no cutting tape), moved onto the basic audio manipulation and then onto every type of plugin/sample library/amp sim/softsynth etc. And now the next thing, AI basically removing the need to even have a human involved in much of the process. We all accepted it as just "moving foward". But its now close to breaking point. A split is going to happen. Your "CPU-less" studio is a perfect example of this. One path will continue in the ultra-perfect digitally created direction and another will go back to an totally analog approach. It has to happen. If we continue on the current path we will just eventually remove the need for humans. There will have to be "Human Only" streaming services. "Human Only" record labels etc. The masses who dont care/know the difference will just keep eating up the music fed to them. And a smaller group of people who understand and appreciate music written and performed by real people will support that.

[@SineEyed]

This notion of yours assumes some kind of discernable, substantive difference between human made music and ai made music. You also assume that music makers i.e. those who know enough to make that discernment, will always default their preference to human made music. Why do you find ai music so unlikable? And what makes you think it will always be unlikable by those with a refined taste or whatever?..

[@paulm3805]

current trajectory taking us to a place where any content can be auto-generated - "give me a techno song with a great catch, steady repetitive beats, with a dark industrial feel", or "give me a spy movie with Michael Caine as the lead actor based on Iain M Bank's culture novels". And probably in our lifetimes. Live performance by robots will also be a thing, but live performances by humans will hopefully be valued more. And if humans decide to stop being artistic, then AI will only have a limited dataset to train on. We owe it to our future selves to keep producing.

[@johnrichardson3297]

4 on the floor, is legit fire chanting and dancing.

[@kensurrency2564]

His point is that the computer can produce a song or jam in perfect time with no deviation or mistake - thus losing the human feel of the natural ebb and flow of a live performance. You also lose the feedback of playing with other individuals, which shapes and colors the outcome of what everyone is doing around you. My band tried to ‘quantize’ one of our songs ‘to the grid’ and it never worked. It sounds clinical and sterile, losing the human feel. I think the debate here is: _human vs. machine._ What do we want going forward? It’s obvious what the humans would choose.

[@jaixiviii]

I asked AI the very same questions and it replied: "Anything you can do, I can do better!".

[@sandwich-breath]

Some people naturally fear change, while other early adopters embrace whatever new technology is invented. This conversation has been ongoing since the invention of the printing press, automobile, television, and cell phone. I was initially resistant to DAWs coming from the analog world of tape and desks. I realized the purist folly of my thinking and have never looked back. The same can be said for AI and everything that comes after. Creativity is our blessing and the gift that distinguishes us from other species and computers... and is our greatest strength.

[@HEAVENLYTABLETS]

It's one thing operating & embracing new technology throughout centuries. It's another converging/ intermingling with new technology to become a new whole. (Transhumanism)

[@Mattened]

@@HEAVENLYTABLETS Exactly.

[@zariisofficial]

@@HEAVENLYTABLETS 100% and the amount of job losses coming is gonna be monumental, i actually have several friends who have lost their jobs to AI (not in the creative industries), it's coming.

[@WillyJunior]

​@@zariisofficialwhich industries, just out of interest?

[@zariisofficial]

@@WillyJunior housing association.

[@janneckchristiansen2778]

We humans will always be interested in human accomplishments above anything else.

[@synaesmedia]

What I was expecting and hoping for from AI was that we'd get instrument plugins trained by skilled musicians, that could then interpret the midi score in a more expressive way. In other words I could write a sax part in my DAW and have it sound like a human playing it, because the neural net had learned enough about how human musicians would approach a line like that. That still seems to be the way that human musicians and composers could get the most creative value from AI. More than from AI trying to chuck out entire pieces in very generic styles.

[@cathbadmusic8489]

Exactly my view on the phenomenon, too. Let AI do the repetitive donkey work and leave the messy creativity to the humans: a much more promising team.

[@BartvanderHorst]

That will happen very soon. And I forsee this will become realtime, in the near future you will be able to sing something and been orchestrated immediately.

[@mokosound]

Thanks for quoting, honoured. For now it seems AI becomes especially good in nostalgia. Maybe that will kick some of us out of our recycling habit and go explore new grounds. Fin de siècle may finally be over? As for the disruption of AI, we already know our technology can extinguish us. Let's hope we are smart and brave enough to manage it.

[@BILLY-px3hw]

Watch out the artists are getting emotional and fearful, arm yourself with and acoustic instrument and fight, write, and innovate. AI should be afraid not us, I personally have been scared about the future of music but the emotional releases I am having musically have been very satisfying. There is literally an army of real musicians and artists who are contemplating an uncertain future, they are frightened and angry. I believe that a Renaissance is upon us, this is going to be historic. We really need it, the stagnation was killing us all anyway

[@genuinefreewilly5706]

No issues with tech or modern manufacturing as long as its on a scale where crafty people can make a living. The demise of the luddites was tragic. The large textile industry isn't doing anything good for a living wage, creates more environmental mayhem than is imaginable but it spits out cheap tacky new clothes a lot of people can not afford or fit into properly. In my mind AI is not going to inspire. The tactile nature of looking through a telescope on a clear night, looking through a microscope, making a meal from scratch, even butchering a piece of meat, inspires. Of course practicing and memorizing a piece of music and playing it inspires. The techno optimists believed we would have a four day work week and have time to pursue our passions, quite the opposite occurred. I am too old and jaded to see much good out of an AI future. No matter how good it gets, drudgery is part of the human condition and bound to get exponentially worse.

[@BinaryDood]

Yes, we should be working 2 days a week with all the surplus production of the 20th century. We work now more than ever, with less benefits, with more insecurity and more needed qualifications, for less money and higher bills.

[@genuinefreewilly5706]

@@BinaryDoodYes, Good point about qualifications. I do not understand why AI companies are focused on saturating an already saturated market of undervalued musical cues. Gains made from tech and AI do not seem to be passed down to ordinary people. There is more grift, fake news, bad products and so on, one has to navigate now than ever before. The music industry was always subject but now, what was serious hobby, preoccupation, a means to an extra income seems far more difficult. This also goes for fine art, video and craft

[@GhostNoteAudio]

The way I see it, I compare generative AI to the popularization of the camera, and how that must've felt to a portrait painter at the time. You made your living by capturing the faces of royals and other wealthy individuals, a process that would takes dozens of hours. Suddenly, there was a black box, with a single button, that could perfectly capture any subject you placed in front of it, without any work necessary. What did the portrait painter do then? They had two options; --- 1; pivot to a new subject area, paint for the sake of making art, not just to replicate faces to hang on a wall. Embrace the artistry and craftmanship that goes into making a painting, and market that as the source of value you offer your customer. --- 2; become a photographer :)

[@samueljbooth9124]

And it will also take over our art work on the cover , sad to say the digital version of our albums. Remember when we got excited to get a record or cd . Look at the art work before we hit play or dropped the stylus

[@rastAsia]

AI.. We need to remember our craft is a service industry. More today than it ever was yesterdecades.. Sonically it used to be 'free entertainment' for bread on table.. But in today's consumer / customer lifestyle we are to be paid wages to make them happy for what it's worth. If there's no exchange of emotion or trade, only then is the art dead. Would consumerism tomorrow be click happy to an AI button? Would our kids be holding on to AI melodies in a playlist that please - or the voices we sang to them since birth to rekindle memories. There's still hope. Keep doing what you do Chris. Cherish our beautiful (human) mistakes. What else would we hear after (if) the nukes drop.

[@shmk1]

What freaked me out about some of the ai generations I have done is the tempo would shift and change slightly like a live band. It was freaky.

[@NeilVanceNeilVance]

Excellent man, you just about covered everything!

[@michaelfaeth]

I say to keep using your organic intelligence to create music and enjoy the process. "Like Prince, Cobain, Lennon, Hendrix, worth more than some high tech trick."

[@paoloramacciotti3532]

Hi Christian. Very interesting speech, with many engaging thoughts. From my standpoint, one of the reasons at the base of the current concern about AI in music is the standing confusion between Generative AI and General AI. For the time being, it is possible to get from transformative algorithms only countless versions of intellectually unaware imitation and crossover, obtained from databases of recorded examples. Right now, generative AI is hardly capable of grasping underlying compositional structures, because it works isolating chunks of sonic data on the musical surface, which is not where the real compositional magic happens, being the acoustic result just the effect and not the cause, in the music-writing process. Thus, there is still a territory temporarily unattained by generative music AI, strictly reserved for composers, that is people who know at deep level how musical languages work. True music creation stems from having absorbed grammatical and syntactic rules, finding a personal and meaningful way to change and infringe them where needed. What so-called expert systems are yet unable to grasp, from a set of acoustic samples, is that when a human composer chooses to change inherited musical norms he is performing a culturally meaningful act, that conveys a vision or interpretation of the world (a symbolic awareness of how an existential experience relates to a specific point in time, and place in space). At present, it doesn't yet exist a musical AI capable of thinking in a self-reflective manner, like a human composer or generally a human being. Hard to know when and how this is going to change, in the next or coming future, but that will be the turning point when the exclusive prerogatives of human music makers will possibly be put into question. In the current context, only musicians used to compound repetitive ideas, replicating simple existing patterns (let us say, low-tier self-appointed social network-based wannabe composers) will be actually endangered by generative AI. For the rest of those capable of managing the intellectual complexity of good composition, there will be abundant time and opportunity for transforming music AI in a useful everyday assistant, who/which will gradually learn bits of technical skill from the practice of his/its human user. When this condition will come, we will better think to music AI assistants like pupils in a Renaissance painting workshop, filling fabrics and backgrounds with colours all the day, in order to relieve their Master from the most menial and less artistically significant part of work.

[@samueljbooth9124]

To create as in so much words from my understanding what you say . The beauty of non-perfection of being human.

[@brunobonaventure2984]

I was kinda of expecting you'd go this way, Christian. Very positive (wishful?) thinking. But... I don't see AI tools for the kind of game music work you talked about out there yet. The higher end stuff is, for the regular person, a musician-replacement tool, not a serious musical tool for the music craftsman. So from that perspective it seems it will be "too late" again. I totally agree with the homogenization. "Homogenized music" is basically what media/movie composers do (and are expected to do, BTW). "Make my film work and don't attract too much attention to yourself" is the modus operandi and it has good reason to be like that. It is a great way to make a living. I love it! But... AI excels at this type of musical output. You read of beautiful analogies with the music technology evolutions of the past. But... There was a time when whale oil was used for illumination. And then whale hunters were no longer needed. Poor guy who was born to hunt whales. No matter how positive their thinking might have been at the time, they had to move on. Is this what will happen to us? Is there any way to avoid it? I was hoping you'd explore this a bit. Maybe you will in a future video in a couple of years, when the smoke has cleared. We'll see if AI was underestimated or not. All in all, my take is that the musician who is a true innovator will make it (for 15 minutes, just before being copied by AI).

[@kuhliloach8842]

Very interesting thoughts about humans becoming more inspired in the absence of distraction. You just reminded me of Koala Sampler which seems to be the new beast of all music tools ever made.

[@GingerDrums]

I really think the outlook is actually more counterintuitive. It is both more hopeful and more pessimistic. It won't be long before AI can create better fitting, more innovative and surprising music than people can. And I mean any one. It's like arythmatic - nobody can compute numbers as fast or as accurately as a computer, and the same thing is happening but with tokens - units of relevance. This will apply across every industry, and this will create a totally different economy. Late capitalism will shade into something far more radical and we will see something new and exciting emerge. Make sure you are registered to vote and let's get more involved in politics to steer this things as best we can. (Edit: spelling)

[@davidbaker231]

I see a man staring annihilation in the face, doing his damnedest to react in some way other than succumbing utterly to his fate. Noble. Admirable. In the end, futile.

[@oggeeboggee]

When will we say: „enough is enough”? When will we rise and take back what is ours? History shows us that it’s necessary “from time to time”.

[@2fioreto]

There are things larger than we, like waves in the Great Sea. Given the current narrative, and the ease with which it combs through the populace, we would be wise to question what sort of deltas they lead us towards. Cheers, and wishing the best 🤘🏼

[@Edbrad]

Ok …but what do you even think that means? What are you imagining when you say that?

[@anatomicallymodernhuman5175]

Man. I have to get back to Edinbugh soon. So beautiful.

[@danielsanichiban]

I think the decades of industrializing art, progressively pushing that to it's limits, has put everyone on edge. As you say, it's just another new thing. Use it, don't use it, whatever. Enjoy yourself. If there's one thing AI can do for us as it removes what little value is left in the recording industry, it's what actually matters and where the real value is.

[@WhitbyStuff]

I have a friend who does FX for television and a couple of years ago I remember saying that I see AI being used for most foley sounds. To that, I think AI will generate most sample libraries withing a couple of years.

[@chrisstrobel3439]

I was there too at the Anaheim NAMM when Midi was introduced, it was midi people were afraid was going to replace musicians .. that fear happened before midi when sampler workstations like the CMI and NED machines came out.

[@g.p616]

MIDI; Plus sampled emulations did and continues to replace musicians.

[@chrisstrobel3439]

@@g.p616 Right, but what I was saying was samplers were around and replacing some musicians before midi was even invented, and us keyboard players were being criticized using em, I used an AlphaSyntauri System on an Apple IIe which was like a poor mans Fairlight myself before midi. The point being it wasn’t midi appeared at NAMM that January and all of a sudden other musicians were suddenly paranoid cause of it .. they were already paranoid before anyone ever heard of Musical Instrument Digital Interface, heck even the Mellotron was putting musicians out of work long before even Fairlight/Synclavier/EMU 😳 lol

[@LeeBlaske]

Interesting video, Christian. I get a little uneasy when I hear anyone predicting that there will be a return to this or that (virtuosic live music, recording without computers, etc.). It immediately makes me remember the rallying cry of the generation that raised me: "The big bands will be back!!" Of course, big bands have never completely gone away, but it doesn't look like they'll be front and center as the main form of pop music again anytime soon. No matter what field people are in, there always seems to be a longing for the old days. But in most cases, we seldom return to the old days. We move on to totally different things that we can't even contemplate now. I personally think, however, that careers in producing recorded music are going to get increasingly difficult, no matter which way things go. Right now, there are supposedly 100,000+ tracks being added to streaming services every day, and that number is increasing. AI is going to boost that number tremendously. Plus, all of the great stuff from yesteryear NEVER goes away. There's simply way, way, WAY too much music out there. The laws of supply and demand are in force. When there's too much of something, it's hard to sell it for a premium price. I see nothing on the horizon that could potentially reduce the supply, outside of a thermonuclear war. Plus, the fact remains that there are tons and tons of content creators out there who just want something pretty good to lay under there production, without paying much at all. AI will satisfy that need, and AI's "pretty good" will continue to get better and better. Not everything needs ground-breaking, amazing music. Not all listeners are even going to appreciate ground-breaking music. Decades ago, avant-garde orchestral composers pushed the envelope too far and lost connection with a lot of their audience. For much of the industry, content producers have often done initial edits with temp music (often, music that they couldn't afford to get). Clients regularly fall in love with temp music. The imitative capabilities of AI are perfect for satisfying such a client. It's mind boggling to think of all of the varieties of music that have been created to date. I do wonder if we will get to the point where it will be difficult to create something "new" that truly knocks people's socks off. Prior to the age of recorded music, it was probably much easier to really impress an audience. Most people (especially outside of major cities) would seldom see a big orchestra concert. People probably did spend more time entertaining each other (lots of pianos and pump organs in homes, and various other instruments). But back then, people most likely didn't spend as much time in a day listening to music. It wasn't underscore to everything. The rarity of music probably made it more valuable to those folks. Is it possible that we've reached and passed an era of "peak music"? An era where people really listened attentively and thoughtfully? I think it's quite possible that in the future, music may be less important to people and occupy less of their time. Scrolling through stuff on your phone has certainly stolen a lot of time young people used to spend attentively listening to record albums. Being able to access any video at any time steals more potential dedicated music listening time. And, how much time might people soon be devoting to escaping into virtual reality? Thinking about what's going to happen might be pointless. Doris Day probably hit the nail on the head. "Que Sera, Sera."

[@FerdiSchwarz]

This is a great take.

[@owendouglas2879]

I agree with all your points except ‘music may be less important to people and occupy less of their time’. No matter how saturated the industry becomes, I think humans will always have an active desire for musical stimulation. It’s something that’s deeply woven into our existence, and the appreciation of music is biologically ingrained into us. It will just make it a lot more difficult for musicians unfortunately, however I hold out hope that this could spark a wave of musical innovation that we probably cannot even fully comprehend yet.

[@topologyrob]

Lots of people still play chess and Go. Lots of people run races, even though we've had bikes for quite a while. As for avant-garde orchestral composers needing to please an audience, that's just mob rule.

[@LeeBlaske]

@@owendouglas2879 There will certainly still be music, but I do believe we're already in an era where there's less focused attention to music for most people than there was decades ago. And again, before recorded music, most people outside of musicians had far less music in their day on a pure time basis. Only so many hours in every day. If you're doing other things besides attentive listening to music (or even doing other things while listening to music), that's less time invested in music.

[@LeeBlaske]

@@topologyrob Maybe, but people's interest in bowling is WAY down. ;) I've also read that Generation Z has lost a lot of interest in sports.

[@ukbleed]

Wow that’s so crazy that you predicted music streaming all the way back in 2022/2023 you must be psychic

[@pablomoralesm]

I'm not a musician but I'm a software engineer and I feel your pain because (ironically) AI is probably going to make people that write software redundant too. For additional context, very good computer code can be beautiful and actual poetry to the trained eye (some of us consider ourselves artists - even if our art is not for mass consumption). That said, you all should try to understanding how modern AI technology works so that you don't under estimate the impact it will have. AI's do not record everything in a giant database or contain a bunch of hand made code to remix and rip-off the training data like some people believe. The only thing that is hand coded in a modern day AI systems is a learning algorithm (basically a set of steps). That algorithm is inspired by the brain and learns like an animal brain does (probably more efficiently because it is using pure math in a way biological brains can't really do). The learning algorithm learns from the training data by adjusting the connections between simulated "neurons". The output of the training is really just a bunch numbers that represent the strength of the connections of the "neurons" in the AI brain. We (humans) don't really understand the neuron connection strength numbers generated by the training in depth (just like we don't really understand in detail how human brains work) but we believe that the AI is learning patterns and relationships between the elements in the input data while abstracting information from it to create "mental models of the world" just like our brains do. Once trained, the AI "brain" can then be given some "inspiration" and use its "mental models" to generate a novel output just like musicians, software developers, or graphic artists can. The AI is doing more or less the same thing we do when we learn something and while it is indeed taking inspiration from what it learns, it is not just regurgitating a remix of the original inputs that it was exposed to during training. The AI brain has real understanding of what it learned, which is why it can do surprising things. There are different ways to train and generate the music using these AI "brains" but one way to do it, is to give it a description potentially including lyrics ("prompt") and some training music to which some random white noise has been added, the output from the AI "brain" is then is compared to the original music and the connections between its "neurons" are adjusted so that it is slightly more likely to remove the noise the next time the same input is provided. Over time more noise is added to the input music until the input is pure noise and the description of the expected output. After doing this billions of times the AI "brain" learns to basically generate music using white noise as its "inspiration" but to that it has had to learn how creating music works. This is the reason why you do not get the same music even with the same "prompt" (the only way to do this would be to provide it the same white noise as input but they probably generate new white noise every time they generate something new). The music it generates is for sure influenced and potentially mimics some of the training data (for example, if one artist is over represented, it will have more of an influence on the "mental models" it learned) but the AI is definitely creating something new every time it is ran and it has the potential to generate totally unique and innovative.

[@arpaddanos9416]

AI samples the "creative voice" of people so that tech venture capitalists can use that creativity to make money for themselves instead of it going to those who had their creativity mapped into the neural networks without their knowledge, consent, or compensation. It is straight out of dystopian science fiction

[@karllindberg]

ML's of today have not been taught to think for themselves, nor to think outside of their boxes ;)

[@GustinJohnson]

Right now it looks like people think code generated by AI is better than it is. I have personally had very mixed results. Some languages it is comically bad at, while being better at others. It still feels a long way away from me being able to rely on it though. Reviewing and explaining existing code it is pretty good at, but generating secure, efficient, and maintainable code it is not.

[@michaelsethcudd]

This is interesting! We truly learn and progress through trial and error, and I think the big advantage AI has is that it can "attempt" (for lack of a better term) much faster than us. To me this means that it can learn many times faster. There's nothing we can do about this, but I do believe we'll need to use the technology in our respective fields where we're "experts" and demonstrate that we can use AI better than a lay person. This is what I hope happens, which still leaves us being able to do things we love...even if it is a bit different.

[@contemplativecomposer5983]

As a member of the IT world by career, I can see exactly how software engineering would be art. Physical Engineering is the same to my eyes. Listening to the design theory and reading the schematics are like poetry, or a wonderful painting as you watch someone construct a Bridge or something else incredible

[@PieterLaroy]

I enjoyed your interesting opinions once again! I agree that we need to embrace this technology for its potential. However, there are some points I don’t fully agree with. Firstly, I disagree with the notion that AI systems cannot be ‘creative’. In fact, they operate in a way that’s quite similar to us. Just as we combine influences from our environment to create something new, AI systems use the data they are trained on to generate new combinations. Therefore, AI can support us in exploring unknown territory by combining statistical possibilities with our human inventiveness. Secondly, regarding larger systems, I believe there is more at play than just simple statistics. These systems often exhibit behaviors that we cannot fully explain. They learn in ways that are not entirely understood, which goes beyond mere statistical analysis. While I’m not suggesting that this equates to consciousness or human emotions, it is certainly more complex than just statistics. For the rest, I couldn’t agree more 🙂.

[@deveyousness]

We still play chess long after the computer did it better. I have no doubt humans will still be making music long after AI starts smashing out hits.

[@sub40hz]

Work with AI. The new kind of “Collab Bro” 😊

[@claursen1]

well said

[@KimonoEtrange]

AI as a tool for creativity is interesting it's the idea of replacing humans that scares most people. It's different than advancements from the past in my opinion.

[@Edbrad]

It may for some things. But at least for a little while human controlled Ai will be able to produce better content. Look at Udio, it doesn’t generate a good result 3 min track with one click. It gives you 30 seconds and you need to choose the extentions and you can direct it somewhat with the custom lyrics (you can also prompt the composition/structure/production somewhat in the custom lyrics). It also won’t be AI music vs human created music, Ai will have more tools for producers. Creative Ai just works in reverse to previous technology, where something like CGI took a long time to get good BUT it gave you total control. Creative AI’s get good first at giving you a good output long before it gives you a lot of control in giving you what you asked for

[@ollied2025]

thats a nice garden and lake. where are you?

[@MattnUska]

10:13 spot on. If you don’t write music that sounds like a human made it then everyone will assume it was created by AI. We have to work on our musical skills again. I let my skills slack for a long time and I’m working on getting it back. Musical skill is not like riding a bike. If you don’t continue to work on your skills you will lose it.

[@NorfolkTraveller]

Something's going to happen... something wonderful 🌳

[@DMTCYMATICSDreamMusicTemple]

Ok HAL

[@timhartnell2472]

You talked about AI being taught the rules of music… I’m not afraid of that. But the intuitive bending of those rules is what makes a performance. Lets make sure no body tells AI how to do that.

[@markkens9]

So much of human composition hangs on the physical interaction with the chosen instrument that an AI can't entwine with note generation or training (yet), that I only see AI music for the middle term as a desert of sound. We haven't begun to digest natural plug-ins let alone sounds beyond imagination.

[@NickThunnda]

I reckon the rise of AI will cause us all to value highly those real musicians playing onstage, mistakes being treasured as proof of authenticity 😊

[@templatemusic]

1:18 wait what are you saying here? That you predicted music subscriptions in 2022? 10 years after that was already a thing?

[@beatsmith]

Adrian Younge has a studio in L.A. that is completely dawless... you should look him up. Fantastic composer.

[@mattfwilkey]

Perhaps we can work to realise a royalty model where composers get paid if their composition is part of a dataset used to train AI and that AI then creates a piece. Something like PRS or MCPS for AI datasets.

[@robertjones9598]

Interesting comments about trance, considering BT did Phobos with Spitfire 😂

[@russcontact]

To comment on one point. You mentioned how people thought midi would replace musicians. It did! Just not the way you’re seeing. With midi one person can control/program machines to play on command, reducing how many humans are needed to perform that work. Today it’s common to see artists perform entire concerts alone, before you needed someone to actually play all those instruments. In that way it eliminated a lot of musicians and that’s still growing today. AI will certainly impact music creation, but over the long run I think you see everything become more personalized. Think like elevator music but customized to how you’re interacting/reacting to it. It wouldn’t just play something based on the Beatles catalog, it would also tweak the music to what it thinks you want to hear. Not sure how I feel about it yet but that’s the direction it’s going. Buckle up!

[@maxbrodie1267]

Expertly put Christian x

[@antsteep]

Love that description of Rupert

[@sosimple3585]

I don't. Going so easy on him his wrong.

[@lit3work]

Mark Hollis’ eponymous album did exactly that Christian- removed the CPU from the process. Don’t suffer for it either.

[@MrMick560]

I think we will rebel and go back to basics, everything artificial will be out !

[@A.I.MADETHIS-ie8bi]

I just made that same Bee-Gee's / loop argument last week only instead of loops mine was "Imagine if you had to invent the words you use every time you wanted to communicate?" Older tech, but the same concept!

[@TheHorse_yes]

There's this brilliant mash-up titled "Stayin' Alive In The Wall (Pink Floyd + Bee Gees Mashup)" by Wax Audio, it's here on KZbin. That was made 13 years ago, so I'd assume it too has been put together with meticulous, manual editing work. Bill McClintock (YT channel) is another brilliant mash-up maestro. I'd like to see and hear more of these absolutely surreal mash-ups in the future, AI could bring us into whole new levels in terms of where you could push the creative envelope. Imagination is the limit, and AI is a great "method enhancer" for that.

[@ivanklass1]

Super Christian, any respectful writer should see A.I as like "A poker player that bluff with confidence by having constant ace cards on his sleeve for the purpose to gain max. profits sum of money." It has nothing to do with human creation of original copyright and observing human rights. Best to Oscar and wishing you well. IK

[@headshotpros888]

Good stuff!

[@soulbill]

‘“… what we’re really scared of is that we will need to work a bit harder.” - yes, THIS! We’ve automated so much already and want to get credit and receive payment for it, even if we are just holding down some notes and letting the pre-programmed samples and patterns do their thing. AI will surely replace that and the lowest common denominator music. We must rise above our tendencies, keep learning, and compose music as humans for humans. I’m a bit afraid as well, but I’m an optimistic realist and think we can rise above what is here and on the way. Maybe AI will provide us with competition we can rise above?

[@CybreSmee]

I think the misconception though is it will be musicians/producers using AI to make music, it won't. It will be end users. You will have an app that creates your favourite music for you, and the film agents will do the same. Some ppl will hold on to the old-school way of playing music manually, but it will become prohibitively expensive and fringe.

[@LeeBlaske]

Great point. It's like desktop publishing, some years back. Desktop publishing put the ability to make great looking print into the hands of many more people who had never done type setting. It wasn't just something that ended up being used by the people who had been dealing with setting lead type, inking platens, and inhaling a lot of carbon tetrachloride fumes while cleaning everything up.

[@Edbrad]

But they won’t be able to get the most out of it. It won’t produce great stuff if you don’t have someone with a good creative ear on the other end. Maybe you aren’t aware of this but most people aren’t very creative. There’s such a thing as good AI and bad AI imagery. The best stuff is made by artists and film makers.

[@CybreSmee]

@@Edbrad Why do you need 'the most'? It just needs to be good. I love Psytrance, its by no measure musically 'good', but I like it. This is precisely why musicians will become extinct, they believe they are the gatekeepers to "real" music. It happened to recording studios, it'll be musos next. Already kinda happening, gawd look at Eurovision for some ear pain.

[@Evilgrapez]

​@@CybreSmee Maybe I'm misunderstanding, but why do you consider psytrance not musically "good?" What's the metric for that?

[@gavinmccabe2908]

I think you're probably right. Non musical people are already doing this and it cuts costs majorly. There seems to be so many hip hop 'producers' out there now that basically buy whatever new software comes out and put things together like a jigsaw, without any real thought or reflection. I wonder, however, if this becomes more and more prevalent, will we reach a peak moment where people just get tired listening to this kind of machine generated music and there is a burn out, or as Christian says will it just take people who are actually musically inclined to take the reigns again and use the tech to create something new.

[@beatowner]

I would love AI software that can help with conforms! Yes, some conform work requires (and benefits) from creativity. But sometimes just a few trims here and there are needed, and with 40 or more tracks, it’s often just a pain.

[@MixedByDotRob]

I'm waiting for the first DAW plugin that makes your music sound more like Suno or Udio ;-)

[@Edbrad]

Look up Google’s recent IO presentation of their Lyria model showing all the new music production tools they’re making. Like singing into a mic and turning it into a saxophone or violin, or tapping on your table and turning it into a drum beat. Or even singing turned into an orchestra. Adding its own harmonies etc etc.

[@SineEyed]

@@Edbrad that sounds like fun..

[@topologyrob]

Reminds me of when schoolkids in the noughties perfected the art of singing with autotune-notch effects.

[@PlatonsArm]

The loop was arguably invented in the 1970s in New York by hiphop-DJs.

[@JacksonTaylorandTheSinners]

I’m embracing AI to help men produce high end demos to my songs quickly. That I can give to my band to learn and then add their own flavor. Then to make even higher end demos with the band. Then when we’re a machine. Goin into a studio with vintage real gear and tape and knocking out the best performances we can. There’s a place for AI. I will use it to make recording traditionally to tape, affordable.

[@hrishiimuni.]

where can i find the intro music?

[@CypiXmusic]

10:09 Thing is, AI is not making music that sounds like a computer