We Built An AI MIDI Generator and Musicians Freaked Out… 

It still ain’t gonna play my guitar for people in a live concert . Do it up bro!

I haven't been replaced, I've never been known.

Definetly up for it!!! Been waiting for something like that for ages)))))

I think you absolutely have to move forward with integrating AI into your music production tools. The key, which you already alluded to, is to see AI as an augmentation tool; we should start using the phrase intelligence augmentation (IA) more instead of AI. In the context of music production, this type of IA/AI software can help speed up workflows, perform more mundane parts of the music creation process, and help generate ideas that composers can then expand upon. It could even help answer composing questions like what the range of certian instruments without having to leave your project. The possibilities really are endless, and people will find ingenious ways to integrate them into their practice. Just as one more specific use case, imagine in the near future if you were able to integrate this technology fully into your plugin. Then imagine you want to try something with a piece of music you were working on that would normally require several minutes and many clicks, copies, pastes, or other adjustments to set up. Maybe this is something like exploring what a piece you were working would sound like in a different key, tempo, or with a different group of instruments playing a section of the arrangement, etc. Instead of manually having to make all these adjustments, if you could simply tell the IA/AI what you want to do and it can do it for you in seconds, you have now created a tool that will save composers a lot of time, keeping them focused on being creative instead of being a computer engineer.

It's coming, from Musio or a competitor, so if you want to make this good and be at the forefront of this, do it. If people do t like it, don't use it. EDIT: I cannot imagine your competitors are not already investigating / developing AI assistants.

Go for it. People will still compose, just as paint by numbers has not stopped artists from painting original works. As a hobbyist, I welcome anything that helps me turn my musical ideas into finished works. I'm already using Band in a Box to arrange my songs. I am proud of the results. They are still my songs. On the other hand, you will be making a major contribution to a force that may obliterate the ability of many people to make a living at composition.

I have already been using ai to make midi tracks or form small ideas, similar to how you did. I have a guitarist friend of mine who got so angry he blamed ME for not playing his guitar for over 6 months. he said i killed his inspiration... to me that is a personal problem and not my problem and i am not going to take the emotional blame he was trying to put on me. you are correct its been being done by a lot of people with different apps, ranging from singing, to song writing, to you name it. its here and its exploding and its not going away. ai has replaced a lot of art, including drawing and video not just audio. its also improving. there will always be people like my friend who are so angry or jealous they will never except it. it wont stop me from making music. the ai isnt ALL of my music and i dont even always use it. im sure others will use it that way. do i care. nope. cars replaced horses. my father refused to ever buy a computer. said they were evil lol. it is what it is. to all the software guys trying to pawn stuff off for hundreds or more dollars i want to see put down like rodent. music should be freedom of expression.. regardless of the freedom in how i use or create it.

Oh and if it gets us out of the never ending loops standing in for actual music with changes and bridges, I’m all in.

I think you should move forward. You don't want to be left behind. I can already do this with LLMs. Best to have your flavour on this. Excited to see what you do with it and see how it can improve my workflow Thanks.

Is this available for testing?

Thank you for this wonderful Video. I'm heading over to Musio to enter your new Challenge because, I know nothing of your company, but your plugin looks wonderful. Wish Me Luck!!!!

Is this anything like Musenet? I've heard that the site/program went down along time ago.

I work as a web developer and as AI lead. You said it yourself. The perfect way forward is to have an AI copilot with the user driving. A randomization button of a whole track is not creative. A suggestion of one or a few notes may be. So yes, keep moving this idea or someone else will.

Understanding it’s a hot-n-cold discussion…as a hobbyist, I see the benefit to this. I can also see the detriment for the pros out there. With that said - I think it can definitely help with deadlines and writers block as well. Fact is if y’all don’t create this - we all should accept a high probability someone else will come along and jam it down our throat. Keep up the great work, guys. -Scott

First I was gonna say: Do it! But then again, in 2 or 3 years we'll probably have music streams generated on the fly to our current taste and context. Who needs MIDI when no-one needs music producers anymore? Why consume canned music when you can get fresh stuff that no-ones ever heard before?

I'm old enough to remember a time when just making music with synthesizers and drum machines was considered cheating. It wasn't "real" music played by "real" people. I've never been a very good musician myself, but I love making music. I can't play almost all the ideas I have about composing or arranging. I can't write notation either. But with the help of a computer and a DAW, I can create complete songs with arrangements quite decently. I believe that AI will be the next step on this path. I think of it as an assistant who helps either in the initial ideation or later in arranging the song. In any case, I make the final choices. The software doesn't decide for me what sounds good. However, the AI programs I have tried so far are still not very good. But the game is just beginning. You can join this game or not. I believe that advances in computer technology and artificial intelligence may even increase the value and appreciation of skilled players. We still want to hear and see real musicians. We want to admire and envy people who can produce something that we cannot. Artificial intelligence and humans have their place in making music, and they don't necessarily only compete with each other, but also complement each other.

Well explained, lol the lute was a major evolution at the time, similar I use always the example of a traverse flute to the modern flute. Every big step in evolution was not always welcomed, but afterwards no one would turn it back. See for instance modernization in industry and machinery, and transportation. nowadays we can't think them away and no-one would. So, go on with your brilliant plans. You have my vote PRO

Worth exploring however, it might benefit you to then keep an eye on it to decide whether to further develop it or scrap it if it becomes the nightmare that many are thinking that it might..... especially since it does and will impact people's lives, livings and professional potentials.

Very well made video. I like that you didn't ignore the elephant in the room, or disregard the folks that don't agree with anything AI. I already have a lot of midi generative tools. Orb Composer was one designed for orchestra midi mockups but i would say it gave mediocre results. If you can manage to make something that can understand a melody and generate chords for the melody, or vice versa (usually i like to write my melodies, with no help tho) that would be super useful. I haven't found much that can work with what I give it so I would say focus on the input to output situation.

Don't see the issue with midi generation. The artist still has to make tasteful arrangement and sound selection. As a producer making a good "midi" is the most frustrating part a majority of the times. Getting rid of writers block for good is an amazing tool that could allow for endless production productivity, you know the fun part. I'm very keen on sound design and synthesis so taking the midi crafting out of it is one less thing to get blocked by, can't wait for this.

Waiting for it!

Definitely move forward! Loved your lute analogy - music has long been using technology - pianos, harps, saxophones etc...! Technology has made the level of mediocrity high - just compare production tracks of the 70' and 80's to now! Music will only get better and even more special - and by "special" I mean the human element which can go beyond machine learning and produce the unexpected and wonderful. Let's embrace it! Ready to beta test any time, Mike!

Do it! I'm both a software engineer and a lifelong musician, my dad's a professional composer and we talk about this all the time. Firstly, you can't stop technological evolution, if you guys don't build it than someone else will and they might not be as ethical as you guys are... Technology has always advanced (think about what painters said when photography came out, we still have painters) AI will be yet another tool available for creativity. Sure the creative process will be slightly different but creative people will always find a way to be creative. I'm not worried about AI at all.

Thanks for your video and your great platform. I am a musician, non professional. I play music for my own pleasure and I think that AI could help me to find ideas, but it will never replace my own creation that gives me emotion.

I hold mixed opinions on the matter. On one hand, I feel confident in my ability to independently craft progressions, melodies, basslines, and counter lines without relying on AI or external assistance. This creative aspect of music composition is what I find relatively easy. On the other hand, the real challenge is the organizational and technical aspects of music production. Tasks such as managing sample libraries, loading them, maintaining templates, and working with mix plugins pose a considerable cognitive load. The intricacies of specifying preferences, like choosing Abbey Road 60s snare 2, toms from EZdrummer 3, and assorted kick options from Battery or XLN, add another layer of complexity. This is particularly evident in the realm of drums alone. In essence, the concept of generating MIDI via text appears to be more of a novelty to me. While expressing musical ideas through text might be straightforward, the practical implementation involves intricate details and challenges that extend beyond the simplicity of textual input.

For starters, MIDI generators powered by AI already exist in varying degrees, so the cat is already out of the bag. And even if yours ends up being the best of the bunch it's still just midi, which a human on the other end of the equation has to work with. AI for the purposes of quickly coming up with ideas isn't the problem, it's the AI's specifically geared towards replacing any human involvement that is at issue here. As a person who follows AI development closely, your MIDI generator should be the least of peoples worries. A quick little peek at what the behemoths like Google are up with things like MusicLM will give people a glimpse at what music without human involvement will actually look like in 5-10 years. There will be a time when AI is just as capable, if not more so, than humans. To doubt that is to doubt human ingenuity, which is very ironic in this context. But just because something can cook for me one day, doesn't mean my love of cooking will go away, and that I will stop the practice altogether. People still fish and hunt, even though supermarkets exist. Everything will change, and nothing will change. That is my prediction.

Please make it. Please release it. Soon. I love using tools such as Scaler 2 and Riffler to assist me in ideation. These tools allow me to create inspiring parts, which take me in creative directions that would have otherwise been lost to me.

Yes, make it before someone else does. So you are saying it can also take a song and seperate each instrument and convert into a midi for each instrument? Dose it reconize the different pitches in the song? I know there are some converters out there but they seem to pick up single notes better than full chords.

I think we definitely need to move in this direction and as soon as possible. If you give us the ability to use AI to create the same quality music based on content that you provide in Musio, it will be really incredible! If that's the case, I think it's only fair to consider splitting royalties on the music that will be created with your tools. Think about it. Today, there is not a single solution on the market that can generate really high-quality music with the help of AI. I think it's only fair that when an author posts a track, he or she must mention your company as a co-author and give you, for example, 15% of the royalties received. I'm looking forward to Musio with AI!

As a self-taught composer, I taught myself everything I know. I didn't know how to grasp writing and music theory when I was a child. I tell myself that for someone who masters computer music but needs assistance in composing, AI may be suitable. I believe that if you are going to develop an AI-driven tool, you should imagine it as a support tool and not just a creative tool. There are also plenty of composers who compose very well but who don't know how to mix properly. Couldn't your tool have a mixing aid specialty? No one is offering it today. Here in France, training in mixing and mastering is very expensive. If you develop a composition assistance and mixing tool, in my opinion, it is guaranteed success.

We need to look at just how far we've come with Tech! I see no one rushing back to send a handwritten letter off!!! Convenience will typically win! 🤷‍♂

Another thing that is really needed is a ultimate audio to midi software, there is plenty of tries but no one get it properly. And if it could separate and give us the midi by instrument would be great.

Looking forward to it!

If you guys don't do it, someone else will... and more than likely without the sensitivity in regards to how it'll impact the industry...

Here's the thing. Even people that hate ai music, are in some way using ai in their process. Ai isn't a replacement. It's tool for creatives. This one sounds amazing, and I want it.

If you don't do it, someone else will. You can't put the genie back in the bottle. The question is, will you capitalize on it, or be left behind?

Explore it!!! there's no coming back!! I'd love to use AI as a composer partner, discussing ideas, trying new possibilites, helping me when I dry out... Well, that's it, Let it roll!!!

yes keep going i would love to use this!

Fact: Very soon there will be highly personalized live music streams of all genre on the major streaming services. There will also be companies releasing AI tools for musicians, hopefully soon, to be ahead of the creative game with the bonus of consciousness we all have.

As a 55 year old guy, started playing the piano when I was a kid, as time went on I started with a Korg M1, I bought this due to having a sequencer to record my ideas. However, there is one more reason as to why I bought the synth. I have cerebral palsy, and little to no use of my right hand apart from 2 fingers, but this never stopped me from making music, sure I cannot play a tune/song that requires the use of both of my hands at the same time, however, the sequencer in the Korg M1 was a blessing and even using an onboard arpeggiator when I purchased the Korg N364 has vastly helped me to score music. Fast forward to now, using Cubase on an Atari ST to now using a plethora of VST's and DAW's, it has allowed me to be creative in my music making. Despite there being some negativity with regards to AI in music, for me, it is a building block to allows me to create and gives me ideas and to look at the structure of the notes that were/are used to create the music which I learn to use with my left hand. So yes, as a disabled dude, I would love to see this software come to fruition, especially for the wider community.

Whatever is created under my watch will ultimately have my imprint on it. Any tool that helps me to write/arrange/orchestrate more quickly and thoroughly is a welcome addition to my toolbox. It isn't a matter of "if" AI will be used in music composition, but very much a matter of "how" it will be utilized. I'd much rather have you an a company like Musio working on this technology than for someone with zero skin in the game to make those choices and decisions. The AI revolution is upon us whether we like it or not. It's up to US as creatives to define the narrative for how it's used. So I say, GO FOR IT!

I think technology assisted composing has been with us for a while. The actual tangible experience of creating and discovering new instruments will override any immediate gratification from AI, well for us oldies anyway. I want to play a cello on the keyboard and a French Horn section. Great video as always and thanks for Musio.

The threat will still be there whether we jump on the AI wagon or not. I would like to see Musio leveraging it for orchestration suggestions or even mixing automation capabilities. Things that can seriously enhance the producer's workflow.

I would say go for it, because if you don't, somebody else will at some point, but they might have a very different mindset and take it elsewhere (potentially not a very positive direction). It needs to support, not replace or promote lack of creativity - That is the most important. :)

This is inevitable, hell it's already happening, so I'd rather have people who care at the forefront of this type of tech I personally want this type of tech for idea generation so I can try a bunch of different approaches quickly to help me figure out what direction I want to take it which is kind of like the old orb composer a few years back but better Will It be responsible for a flood of very mediocre music? Yes. But we already have aiva and tons of other companies doing it so it's not like it will be the sole reason for it Simply put. Go for it

Why would you not? Go for it! Besides, any tool that can be utilized, especially for those fairly new to composing with DAW's, will be beneficial. It will invariably motivate them to learn more, explore more (without the tool). Plus, you may attract more customers for your sample libraries. It's a win win.

You can't stop change. You're either in front of it, or you're chasing it, but you can't stop it.

Absolutely follow the technological wave or get left behind. Musio is a fantastic service.

hi Mike, if the AI midi generator adds to the musician's creative level, then yep, go for it, it can create ideas when you dont have one or are stuck on a project, its the same thing with for example captain melody, its a tool to help your creativity, otherwise, you could instaid create a state of the art orchestrator module into musio, more sophisticated then the one in sonuscore the orchestra or east west's orchestrator, but without the incorporation of AI, that would also be awsome, cheers Mike

Let's discover this! 💛🙏

I believe in tools to help creativity and lack of knowledge of human creators. :) We could start with some sort of "performance interpretation" loading and adjusting the phrases (midi data) to sound more polished and professional. For example: Change articulations and velocities based on note durations (like an Orchestral Quantize). But it can go further like selecting styles to change to. So you could have a cinematic arrangement and ask AI to change to a different mood or even different instruments (while adapting all midi data to make the new style and new samples sound perfect). If a sample has a faster attack and another slower, AI could help fix based on it's own sample choices. All of this considering Musio's libraries, of course.

This looks fun, I was looking to plan to write a simple midi, but the AI will add to it, such as I have the piano piece but write an orchestral composition

For the composer: Build a tool that can analyze a composition that one has made, and generate variations based on text prompts and a quick interface with ready made suggestions, or some sort of x/y interface. This can then be used for inspiration, or for work, where one might have to produce variations on themes. Probably also an option for an AI suggestion on how to adopt music to cues. But also honestly, many composers today have not studied music enough to be make suitable parts for all instruments. A good composers will almost certainly make changes to practically any suggestion AI will come up with. So it would be a tool. Will there be those taking advantage of the technology instead of hiring a composer. For sure, but those are probably the exact same people that underpay composers today. It might actually be a good thing for well educated/highly experienced composers, as those will know what to do with what the AI suggest, to improve it. And with less of a chance of getting experienced by taking an underpaid job, education will probably get a higher status again. For hobbyists, any tools that help them make music that they can then play around with, and attempt to make their own, is a good thing. As long as it generates midi, in a transparent way, and isn't just spitting out a finished product, I think it can be a great tool. The important part for composers, is that it can analyze work that they have already done, and not just from what is already in a database it has been trained on. It is already said that people don't value music like they used to. And if AI is as good as humans at making music. Well then, it might free people from unrealistic dreams of becoming music makers, as less than 1 percent has ever been able to make it, in that industry anyway.

I quit my programming job to become a musician. It's tools like this that will allow me to have a smooth transition between the 2 domains.

I'm both a musician and tech enthusiast, and have worked with bleeding edge technology for the last 25 years. There are always going to be naysayers to advancements in technology, just like "buggy" and horses were considered the standard over 100 years ago. The invention of the automobile caused havoc in the "transportation" industry at that time. Bring it on....make it real...make it accessible. Using it as an idea starter is no different then hearing a piece of music, grabbing a kernel of an idea out of it, and making it their own. My feedback: If you don't do it and be one of the first to market.....someone else will. There is someone out there watching this video right thinking about how they can leverage & capitalize because of this video alone. (maybe even me)

"we will not replace the creator" it sounds like the typical sentence of shareholders who come to see the employees of a company before being sold and dismantled. everyone knows that the musical creation professions will be reduced to nothing... only the great composers will have musical creations which will sell for even more money because they will have the label "created by a human", this profession is doomed to become poperized

I personally see AI as far less of a threat in the music world as I do to visual arts. Most genAI for visual art focuses on the finished product, while I'd say that most music-based genAI focuses on the building blocks. MIDI generators have been around for ages (like FL Studio's often unusable riff generator), so I'd say...go for it. Pandora's box has already been opened and it would definitely help alleviate the "blank page" problem.

I like the idea of AI making instruments play in a more realistic way. So that we don’t manually have to add automation

Go for it! Many of us are not offended or scared of the big bad AI! I personally love it and it just allows my compositions to travel and vibrate in different directions! Hey I will love to beta test this amazing AI midi/technology that you are speaking about! Sign me up!

I have been engaged in the field of music production for nearly 20 years. The progress of AI is remarkable, and it will undoubtedly continue to advance. It's inevitable that the trend will persist, leading to the disappearance of many less skilled musicians… Utilizing AI not only for those seeking ideas but also for those curious about the reasons behind certain outcomes seems to be a valid approach. Providing comments that succinctly convey the structure of a song generated by AI could offer a new learning opportunity for many composers. It may go beyond mere music theory, incorporating trends from recent music charts and potentially adding a new dimension to one's music.

I would like to take a piano map composition, and dictate the instrumentation to a more fully realized recording... whether it be Orchestral or Future Bass. The most frustrating thing for me as a composer is that I have to now be a producer, performer, and an engineer.

It will definitely come, if you don't do it, somebody else will do it. From this point of view, go for it. With all contras, I think it could be also a good advantage for professional composers as a tool for inspiration and to reduce routine work. When going too much into detail, it's most probably not safe regarding copyrights. Therefore I would use it only in a rough way, like "give me a driving heroic rhythm 5/4 over eight bars including ethno drums" or such. I could imagine it also comfortable for temporary music in film scoring, when directors ask you to compose something in the style of xy. So you could just ask in the prompt for a suitable midi but make then your own story with it. This definitely would save time where you had before to search, listen, and imitate the asked score. Of course the good question is, how to feed the AI. I think the legal situation is not really clear yet and new laws may be expected. It would be also interesting how good the AI could translate more complex scores into midi, like f.e. an orchestral piece. But in general I am confident enough, that AI doesn't steel my job but could give me instead an advantage to work faster. So I definitely need to go for it before others do 😅

I personally take pride in invention. Listening back to something which I came up with, not something created by AI which I simply edited or arranged. It also allows me to explore ideas which are unique, rather than based on algorithms which rely on convention and predictability. All that said, I'm not against anyone who uses this if it makes them happy. And I'm totally against A.I. - for me, the areas in which I feel A.I. would be most helpful will be repetitive and less creative "admin" musical tasks - stuff like audio editing, phase detection, audio repair, midi editing and alignment, delay compensation issues, mixing utilities, and all the hours spent on stuff like volume & dynamics automation just to get something to sound right, or to match longs and shorts, etc. From a writing perspective, I wouldn't mind it assisting with flagging stuff like voice leading issues, truly accidental accidentals (whoops, forgot that bass was supposed to be G#!), and perhaps even making suggestions of potential harmonic movement which one could turn a dial from fully diatonic to chromatic for interesting and unexpected suggestions. Also sample and library organization & management, and basically all the many other things you'd want an assistant for. This way I can really just focus on writing which is my favorite part of the whole thing.

I love the idea of taking a pencil sketch and having AI flesh it out in a variety of ways. I don’t see the harm in having it generate chords either. There are so many AI plugins that do that already anyway. Go forth Musio!

As others said, its coming no matter what, so you might as well work on it. There will be displacement as any technology advance causes. Your vision of creating tools to increase a composer’s efficiency is the right balance. AI is not yet capable of composing incredible sounding music, in both composition and production but augmenting with human creativity can.

More 2 cents: Use AI to control samples in a more intuitive and dynamic way. Maybe integrate Sample Modeling into Musio's Arsenal. Users could record any instrument (including voice, whistle, guitar, etc) and have this translated to the sample library with all the minor nuances like vibratos, tune, etc. It already exists for voice to voice (Moises.Ai) when you sing and AI changes your voice to another one available in their trained models.

The fears will be mostly reality but this is going to happen. Just as with art, programming, engineering, analytics, manufacturing. What I think I would most likely want to use it for is to help in orchestration/arranging my own music. Like all machine generated content, a human will still need to input idea, edit the output, etc to get a quality that is not obviously generated, at least for now.

I assume when composers began to be able to create scores on a personal computer (beginning perhaps with MOTU Professional Composer in 1984 with the release of the Macintosh computer), somebody lost their job. Or at least didn't get as much work. Anybody could and still can use pencil and paper to write a score. But using a computer is pretty nice and very helpful and used by many. Mike, you said it yourself that Cinesamples/Musio is a "music technology company". So, I think you guys should create great new, innovative music technology and we'll all figure out how to use it to help us get jobs done. And if someone wants to use a pencil and paper and hire a copyist, that will still work, too.

Go for it! And add microphone selection as well 😉

I feel vindicated in giving up all of my creative hobbies. This is hell.

Imagine all the people out there that have incredible music in their mind, that will never be realized because they lack the dexterity, or training to play an actual instrument. It's exciting to think I could compose a piece, that I could never play with my own hands, lacking training and years of practice, yet could produce with AI the way I imagine it, for professional musicians to play on an actual instrument, adding their own unique feeling to it. That's extremely exciting for me as I am starting to suffer from arthritis in both hands and won't be able to play anything in the near future. The music is still in there....this could allow me, and I imagine many others, to get it out there. Count me in ! You can't stop progress so you should be on the forefront of it all. 🤔

You’ve mentioned ethics in other posts and I want to dig into that (apologies if this has been asked and answered): what do you see as your obligation to compensate musicians whose work is used by the llm to generate this midi? What kind of licensing seems appropriate?

To add to my comment, I don’t actually need a creative assistant. What I dream AI might become is a truly brilliant studio assistant. Because in a business that has become equal parts creative work and highly technical button pushing and knob twirling, it is always the technical tasks that pull me out of the creative work. I would like an AI assistant that can listen to human commands, learn my workflow and take care of those technical tasks on command. It would need a PhD in midi engineering, audio engineering, DAW programming, notation programming, music theory and literally dozens of other things the perfect assistant would know. It should be able to seamlessly work across software programs. It should be able to respond in my language, in a CG voice that doesn’t make me laugh or cringe; an assistant who will set-up a session to specifications that I give, stay with me as I work-add tracks, delete tracks, transpose tracks or bars, change keys or time signatures, quantize an unquantified track, add expression marks, etc.-all on command, all in a flash. And these examples barely scratch the surface of what it should do. There are so many times that my creative process is interrupted by having to put on a different hat to do one of those necessary but mundane (or sometimes very complicated) tasks. Those interruption, which intrude on the creative process, are more than a time suck; they also kill the flow of ideas. This AI assistant would be available at any hour, for as long as I want to work. My current assistant is a very capable human and he does a lot of pre and post session work. But he cannot sit by my side while I generate new music, cannot anticipate my needs based on knowledge of my creative process. An AI assistant like I have described above (admittedly not very well) would be a godsend, would save me untold hours. I would pays serious money for that.

Rather than making AI that makes music, we need tools to enhance creative process. For that reason Scaler 2 beats AI based generators and helps a lot with learning music theory. We need more of these solutions.

There is a basic role " if you didn't do it, someone else will do it ", So by that statement it's better to do it in your own way of creativity, let me give you an example there is a lot of plugins that generates chord, complex chords, melody .. etc., at the end what you will choice is what you as a composer will feel better, creative and in your own style, yes may be Ai will reach in a way that you can hear a perfect composition but for sure it will lack a lot of feeling, taste and the signatures that the composer do to his work. Another example if we think of sampling industry , imagine an instrument player such as cello or French horn 30 to 40 years back and we ask him if he would agree to record his instrument with a lot of articulation to be used later in a VSTplugin at that time he will refuse or he will feel scared that his job will be dropped in the future but now in music label & orchestration you can see that listening the composition it in real instrument after sketching everything in a DAW is something else . SO I think go for it let an AI tool be in an option in musio and let the people explore more and decide the future .....Thanks Mike, will be waiting to use your tool in my musio by the way 😁

RU-vid's algorithm suggested your video to me (probably cuz I've been checking out AI vids and also Music Production videos) and it's NOT lost on me how that is also a form of technology learning from humans to cater to us. And that brings up the philosophical question that I believe you're trying to figure out - HELL, we're ALL tryna figure this out. When does AI assistance encroaches beyond Human creativity and ingenuity? I don't have the answer to that. But.. I am APPRECIATIVE of companies and people like yourself that are asking the actual demographics that will use it if this something that you want moving forward or not. And that is where, figuring out what demo are you targeting? Newbies that are intimidated by music and unable to wrap their heads around music theory? Hobbyists that want to get ideas/concepts out of their heads produced at a professional level? Pros that are seeking ways to organize streamline their workflow so that they can accomplish more and let the tedious mundane tasks get automated by AI assistance? Figuring out which lane, could help what direction that you would like to go in catering to your demographics. As for me, I'd prefer having AI take things that I've already composed (or composing) and suggest different alternatives to how MY works would sound. That along with finding ways to catalogue my songs and break them up into loops where I could listen to my own collection but mixed and matched. These are tools that don't currently exists today, but would be something I'd be interested in. But even with my suggestion, what I'm doing would be moving towards more independent music production and less collaborative produced music. And that is it's own philosophical question of: should music creation be a collective process or an individual one? I realized that I didn't provide any answers - oops But those are some thoughts that popped up after giving your question some thought. Hopefully, there'll be a law or some regulation indicating what music is entirely AI based and what is ol' fashioned HUMAN based that is passed QUICKLY. Because these questions are impossible to resolve on a case by case basis.

If you don't simply someone else will. People choose how much technology to use in making music, always have, always will. As always, such tools enable people with widely different backgrounds enjoy making some music. Tools like Scaler help people create chord progressions with no understandings of harmony or some, or a lot of knowledge. We have algorithmic rhythm generators also, so AI will hugely push into the space of music creation. There's also some built in limitations: The copyright office will NOT copyright a work created solely by an AI with no significant human contribution (same for patents). AI will replace ALL JOBS eventually, so no worrying about the job aspect. I am happy to use these types of tools where they will make suggestions I haven't thought of, or as you say, as a helper. Lots of people can create simple 4 or 8 bar beats, but what to do next to make it into a song? Perhaps a tool that does some of that noodling, presents some alternatives of where to go next with choices like "extend verse", "create chorus ideas", "create middle section", etc.

What I'd be now interested in, is an AI modi generator that could write additional parts to songs that I compose.... Because I didn't play everything! And I can't afford to constantly hire other musicians

Hey Mike, you got the experience, the knowhow and the skill. You realise it's possible. No one else in the tech industry is going to stop. You care for you and your family. Don't stop. I super freking hate it. I just decided I was going to deep in my music making more committed to make it profitable after 20 years/ I learned about computer music just at the middle of the pandemic. I decided, made the investment, got the equipment, learned about reverb routing in freaking LOGIC and it was painful LOL/ And finally I can say I'm at it and PUM! A freaking I. Super damn it. But, well, are they going to stop? nop. So you shouldn't either.

I think I'd rather a respected and respectful entity like you and Cinesamples do it right and do it better. Others will be creating it regardless of what all of us think. It's a great point for discussion. I think a certain amount of composers tend to have this sort of lofty "Oh, AI will never replace true human creativity" idea that makes them feel safe I suppose. I think it's better to have eyes wide open, face the fear, and understand all likely scenarios. We can't put the toothpaste back in the tube at this point. It kinda blows, sure. Let's see what we can do at this point.

Explore it. Music needs technology in order to be created. Go for it

I like idea of AI tools as an assistant, a facilitator who can help me in some creative and production steps. Certainly the AI theme linked to complex issues such as copyright, jobs, etc.... But in relation to the creative aspects, I believe that we are all facing a time that opens up interesting innovations. If the music that comes out of it decreases in value, it will not depend on AI but on human beings who do not know how to best use these new tools. For example: take a pianist. If he makes bad music it's because he doesn't know how to use the instrument well. AI and piano are just instruments. Full speed ahead! Peace

Please release it!

It's going to happen anyway, we can't stop it, so we might as well use it. Our competitors will. So go for it.

As others say, the only reasonable choice is to move forward with it. The genie is out of the bottle, so to speak, and it isn't going back in. Whether you develop AI music creation tools or not doesn't affect that others will be doing it. So, why not be a face for the positive aspect of AI in music creation? Especially if you have the opportunity to "do it right".

How about a questionnaire-led interface. What mood, bpm, song structure (ABAB etc), key and scale, instruments to assemble, song title and even your own lyrics BEFORE the AI kicks in and produces something. We all have to deal with mandatory fields in forms, why not do it in your generative software? The producer has then already invested creative thought, and would surely want to own and modify the good/bad result of your wonderful AI system. Everybody then wins…Oh, and if it spits out DAW friendly output like for Ableton session view, that would be the gold subscription/ price tier…

I'd rather have a company that is conscious of the concerns of composers making the best version of this than a company that doesn't care at all.

in voice of Palpatine "Do it!"

I do very minimal techno, so often i play with a set of 2-3 notes, i utilize microtonal scale sets to create atonal stuff. i use Euklidian generators, to generate stuff fast, to find new rythms - but these are only tools, not sure about an AI companion. Its not a real band mate - when youre in a Band, often there was one guy, generating the ideas, and hands out the song to the players, and making a vision come true. I am not so sure how democratic the general band process is - this is why i went from metal to techno, where i had more control over the musical vision. Would AI learn your vision? If its a online medium, would we hand out our vision? Is it run locally? Would our vision influence other users, using your AI ?

Where can I sign up to beta test?

I would really like something like this as a plugin, actually. Forget the web idea. Just make it a straight up AI plugin - something like Izotope’s products - and I’ll buy it. I know we have things like EastWest’s and Sonuscore’s “orchestrator” that kind of does the same thing, but that only has premade patterns in it. What you’re proposing is a lot more open-ended and possibly more useful. So what do I think of this? Well, I think AI is definitely a threat. To what extent remains to be seen. But does that mean we should avoid it like the plague? No way. Why? Because a plugin that provides midi samples of written prompts could be very useful for idea generation, brainstorming, and general sketching. For playing with possibilities, ideas, and maybe even arrangements. Useful ideas to build on after they’ve inspired you. To me, this is a lot different than what Sonuscore is doing atm with their “The Score”. That thing not only has premade patterns, but will literally build entire scores of music (and easily alter their dynamics for simple arrangement parts). To me, that’s too easy. It reduces music to a craft of mere buttons, of a few simple procedures. And music shouldn’t be the result of a few simple buttons. It’s a craft that should be worked on with care and development. And so for me The Score is really using technology in the wrong way. Instead, composers and musicians should put time in music theory and their craft and really avoid relying on something like The Score, because that’s a crutch that sets us all back. I will say, though, that in the right hands, The Score can be very useful for sketching and brainstorming as well, especially for playing with arrangements. But the fact that it can build entire scores effortlessly still bothers me. And so, as we see here with The Score, AI can definitely be a threat when used improperly. But this midi generating AI that you’re talking about sounds like a great “inspiring” tool, like you said. And I don’t see anything wrong with that. As long as it doesn’t build entire scores for you on its own, I’m all for extra tools that can assist in the creative and workflow aspects of composing and music production.

They said music videos would kill music on the radio. They said recorded music would kill live performances. At every point there's been naysayers wanting to stop advancement because they're afraid of change, and they're almost always proven wrong.

We’re not going back, so it’s better to just adapt and learn to use it to our advantage. 🤷‍♂️ One can’t ignore that seasons change, new tech arrive. I try to embrace and see how it can ease up my daily tasks and visisons. We can’t comprehend expontential advancement, that’s why we’re all confused and it’s hard to keep up. It’s sad to see so many jobs go away though. We’ve seen a tremendouse shift in games industry alone. Every industry will be effected one way or the other. 2023 (or Nov 2022) was the year of AI public introduction. 2024 is where the real AI products (and robots) starts to roll out. Fasten seatbelt.

One guiding principle could be: let’s free people from having to reinvent the wheel every time. If a single note is a letter, then music has a vocabulary of words which many musicians learn and then put together to create sentences and then build whole stories, ie, songs. Although musicians and poets do occasionally create new words, most of the creativity and meaning exist at the level of sentences and stories which utilize a vocabulary that most musicians share in common. Give us words as wheels.

All I want in life is to generate awesome shred guitar leads and stuff, using AI. I cannot play guitar to the level I used to (due to a serious hand injury, and onset old fart arthritis), but I'd like to use MIDI to create lead solos for my original songs. 😀 Oh, and to crush my enemies, see them driven before me, and hear the lamentations of their women. lol

What exactly are we talking about AI? A training model that uses licensed music as input, royalty-free music or just an algorithm that simply applies “rules” of harmony and arrangement? It is certain that against blank page syndrome, this could sometimes be useful. But will development and maintenance of this tool be worth it in the long term compared to the use to which it will be put? I mean, most tools in the same genre are, to my knowledge, very little used. I understand the "marketing" advantage of such a tool for Musio, after the thing that could counterbalance the "automatic music" side would be to add pedagogy options to the tool, with bull info that would explain how it is composed a melody and then harmonize it (question/answer type, musical sequences, etc.) using blocks like programming tools for children or in any case really add a learning side.

As always technology in and of itself is neutral. How we humans use it, gives it positive or negative value. We can use to to enhance or devalue the human experience. It's still up to us.

I see benefit in a tool that could help flesh out the orchestration of some melodic ideas. But as someone with some background in AI and software development, I think that such a tool would need to be much more advanced than this demo in order to be regularly useful for me. I don't think this demo is "60%" of a working tool. It looks like

This could take Musio to the next level

It is what it is, I'm tired of be scared of the future. If you put time in effort into it, then I dont care that it's AI.