Here's What Gemini Thinks About the Whole Creative AI Argument

[pdxindy]

Okay, you're talking about the performance. I want to do that myself too. I don't want some robot playing my song for me even if it can do a better job. I like the feel of sitting down at the piano or whatever instrument I'm emulating.

So you don't use sample libraries or synth presets... only real physical instruments?

[wagtunes]

Okay, you're talking about the performance. I want to do that myself too. I don't want some robot playing my song for me even if it can do a better job. I like the feel of sitting down at the piano or whatever instrument I'm emulating.

So you don't use sample libraries or synth presets... only real physical instruments?

Oh for the love of God just stop. You know what I mean and I refuse to get into a semantics battle with you any further. Just go away.

[wagtunes]

Not semantics to me as they're two distinct aspects, however related.

You're too defensive.

And you're a PITA.

[wagtunes]

Look, here's the bottom line. I have no trouble with people using AI if they want to use AI. I use it for vocals. That is the ONLY thing I will use AI for. Everything else I want to do myself. That is my choice, for whatever reason.

Every day, more and more, I'm reminded why I hate this place so much.

[BONES]

@Bones - Chord structure. You know, like a typical 1950's ballad structure like from "In The Still Of The Night" which is C, Am, F, G.

Yeah, it still doesn't make any sense to me. I really don't do chords at all, I don't like the way they sound. I use notes that sound right when played in sequence, which I imagine would conform to some chord or another in aggregate but I wouldn't have the first clue what chords they might be. Or maybe they wouldn't, maybe they'd conform to some progression or other? At the end of the day it's not relevant information to me. It all feels very arcane.

As for that sketch, I saw it yonks ago, probably on Channel 10, and I thought at the time that as long as you were singing in the same key, you could probably sing absolutely any 4/4 song to that backing.

Anyway, enough OT banter. Let's get back on track. At the end of the day AI is what you, the user, make of it. If you're a moron, you'll just use it to pump out shithouse music that, like most music, contributes nothing to anything. But truly creative people, people with a vision and purpose, can leverage the power of AI to make the creative process easier and faster, while the results could well be better than if they hadn't involved AI at all. That's just the way it is and there is no point in denying it. It's already at least as good as we are and it's only going to get better at an exponential rate until it leaves us all in the dust.

[pdxindy]

Okay, you're talking about the performance. I want to do that myself too. I don't want some robot playing my song for me even if it can do a better job. I like the feel of sitting down at the piano or whatever instrument I'm emulating.

So you don't use sample libraries or synth presets... only real physical instruments?

Oh for the love of God just stop. You know what I mean and I refuse to get into a semantics battle with you any further. Just go away.

It's not semantics... and 'what you mean' is a misunderstanding of the tool. Playing midi data to a sample library is no different than playing it to an AI instrument (besides the AI not currently being real time, but that will change). If 5 years from now, AI instruments are the most expressive and natural instruments to play via midi, why wouldn't you use them?

[pdxindy]

Every day, more and more, I'm reminded why I hate this place so much.

My apologies... I wasn't mocking you or trying to upset you. I was simply making a point... slowly.

Anyway, here is an example that is like what you are currently doing. The vocals are all Ace Studio and the rest is regular DAW and live playing.

[BBFG#]

Every day, more and more, I'm reminded why I hate this place so much.

I'm not trying to mock or bait you into any discussion. I understand you're still dealing with pain. In fact I was agreeing with you and only offering my added perception of the same thing. I realize you have a problem with me at the time. That's fine and hopefully it will pass. BTW, the video you posted was fun and took me back to a teacher I had that would play "Blue Heaven" while singing different lyrical melodies to it, finally ending with "Don't It Make My Brown Eyes Blue".
I forwarded it to my spouse and she enjoyed it as well.

I'm fairly certain I will eventually use AI myself just as soon as I can see the benefit of doing so. I'm much more interested in LRM than LLM though. Especially since LLM is what I generally get from many players hiding behind their cognitive distortions.

[wagtunes]

Okay, you're talking about the performance. I want to do that myself too. I don't want some robot playing my song for me even if it can do a better job. I like the feel of sitting down at the piano or whatever instrument I'm emulating.

So you don't use sample libraries or synth presets... only real physical instruments?

Oh for the love of God just stop. You know what I mean and I refuse to get into a semantics battle with you any further. Just go away.

It's not semantics... and 'what you mean' is a misunderstanding of the tool. Playing midi data to a sample library is no different than playing it to an AI instrument (besides the AI not currently being real time, but that will change). If 5 years from now, AI instruments are the most expressive and natural instruments to play via midi, why wouldn't you use them?

Okay, here's my questions. Let's say 5 years from now I can play a MIDI note and in real time, AI will convert that note to whatever sound I want.

1. Do I still have to describe to the AI what I want the note to sound like? I assume yes. So what if the AI comes up with a sound I don't like? See, with a sample library the auditioning of sounds is seconds. With AI, it would be hit and miss and still potentially take a long time to come up with something I like. So this now becomes time wasted. And with many of my sample libraries, I know exactly what I want without even auditioning.

2. What's this going to cost me? Is the AI service going to charge me per sound. If so, that can get very expensive over time. A lot more expensive than just buying a library and never having to pay for it again. Now if it's a yearly fee, like with Audimee for vocals, that's different. Then it might be worth it. But then there's the pesky problem of having to describe the sound I want to the AI. I mean what if I say "give me a violin sound that sounds like the Chris Hein solo violin" and it doesn't know what the hell that even is? Then what? If I had the Chris Hein solo violin I wouldn't have to go through all this crap.

Again, I have no problem with AI. But there are cases where I just don't think using it will make my life easier or better. I prefer doing those things the way I'm doing them now.

Besides, it's not 5 years from now and AI can't do this in real time.

[wagtunes]

Every day, more and more, I'm reminded why I hate this place so much.

My apologies... I wasn't mocking you or trying to upset you. I was simply making a point... slowly.

Anyway, here is an example that is like what you are currently doing. The vocals are all Ace Studio and the rest is regular DAW and live playing.

Very cool. Yeah, I got no problems with vocal AI. Thanks to it, I will never have to sing on a recording again. Just sing well enough to give the AI something to work with.

[pdxindy]

Okay, here's my questions. Let's say 5 years from now I can play a MIDI note and in real time, AI will convert that note to whatever sound I want.

1. Do I still have to describe to the AI what I want the note to sound like? I assume yes. So what if the AI comes up with a sound I don't like? See, with a sample library the auditioning of sounds is seconds. With AI, it would be hit and miss and still potentially take a long time to come up with something I like. So this now becomes time wasted. And with many of my sample libraries, I know exactly what I want without even auditioning.

In Ace Studio, you select a violin. The app has several. It's basically the same as selecting from among your sample libraries. The limitation with sample libraries of course is, if the library, or libraries don't sound like what you want, you have to go buy a different library.

With such AI instruments, you do not have to describe anything. Just select AI Violin and play.

[pdxindy]

2. What's this going to cost me?

Considering that it's easy to spend $20,000 or higher on sample libraries, my answer would be... less

[KBSoundSmith]

To get to the essentials of what Gemini said, I'm in agreement. If an artist insists on personal agency, there is nothing stopping an artist from using AI as a powerful amplifier, rather than a substitute for the artist themself. It helps you work faster, provides useful critique and feedback -- you never have to allow it to generate a single word or chord, there are other ways to use it. AI is multiplicative -- if you input Zero, well Zero Times Anything = Zero. Your fault.

[wagtunes]

Okay, here's my questions. Let's say 5 years from now I can play a MIDI note and in real time, AI will convert that note to whatever sound I want.

1. Do I still have to describe to the AI what I want the note to sound like? I assume yes. So what if the AI comes up with a sound I don't like? See, with a sample library the auditioning of sounds is seconds. With AI, it would be hit and miss and still potentially take a long time to come up with something I like. So this now becomes time wasted. And with many of my sample libraries, I know exactly what I want without even auditioning.

In Ace Studio, you select a violin. The app has several. It's basically the same as selecting from among your sample libraries. The limitation with sample libraries of course is, if the library, or libraries don't sound like what you want, you have to go buy a different library.

With such AI instruments, you do not have to describe anything. Just select AI Violin and play.

Again. what if I don't like the sound the AI comes up with?

As far as the cost of sample libraries, for me it's a non issue. I have almost every sample library out there and yes, I have spent tens of thousands of dollars. But it's done. So I don't need AI.

[Grizzellda]

@Bones - Chord structure. You know, like a typical 1950's ballad structure like from "In The Still Of The Night" which is C, Am, F, G.

Yeah, it still doesn't make any sense to me. I really don't do chords at all, I don't like the way they sound. I use notes that sound right when played in sequence, which I imagine would conform to some chord or another in aggregate but I wouldn't have the first clue what chords they might be. Or maybe they wouldn't, maybe they'd conform to some progression or other? At the end of the day it's not relevant information to me. It all feels very arcane.

As for that sketch, I saw it yonks ago, probably on Channel 10, and I thought at the time that as long as you were singing in the same key, you could probably sing absolutely any 4/4 song to that backing.

Anyway, enough OT banter. Let's get back on track. At the end of the day AI is what you, the user, make of it. If you're a moron, you'll just use it to pump out shithouse music that, like most music, contributes nothing to anything. But truly creative people, people with a vision and purpose, can leverage the power of AI to make the creative process easier and faster, while the results could well be better than if they hadn't involved AI at all. That's just the way it is and there is no point in denying it. It's already at least as good as we are and it's only going to get better at an exponential rate until it leaves us all in the dust.

So BONES, ultra cool to know... "you don't do chords", except the last time I checked, chords had something to do with music, but you do not need that...alright understood...

So, are you gonna rely on Artificial Intelligence to riff for you? I don't think so, reading your posts. You seem to be willing to use it though, at some level, and I can meet you halfway, I feel.