AI Amp Modeling Next Gen: A Frankenstein Scenario

[seamonkey]

I'd say that the extent to which even use of a AI as a tool might affect copyright still needs to be tested in court. It's already been ruled that you can't simply have an AI write all possible melodies over a weekend and claim that your human input is selecting all of them and copyrighting them all. So simple selection may not be enough but what sort of additions, assemblage or whatever is required remains to be seen. These are uncharted waters and the law on the use of randomness or algorithmic composition is mixed

Agreed, it needs to be tested in courts, but I don't think there's anybody who knows how to do that yet. Legal experts better hurry up, it's happening now.

The melody and lyrics are the elements that are protected by copyright. Although you can copyright either separately. You can register the copyright with sheet music. I doubt anybody could tell by looking at the sheet music how it was created. And what's not to say that given a sheet of music that someone plays or sings in their head is not enough to participate in the copyright? The may make an artistic change to it, phrasing, note choice, lyric change.

AI does have the benefit of scanning all existing copyrights to be sure the song is original.

 

[ardgedee]

AI does have the benefit of scanning all existing copyrights to be sure the song is original.

LLMs do not have the means to do research in a way that exceeds humans' ability to do it.

In fact, so far LLMs have been extremely crap at research, with failure rates exceeding 50% (according to Google's internal testing of Bard, their own LLM). We do not have AI yet, although we have a number of companies eager to make us believe we do.

 

[N. Kentucky Bass]

They played on later albums. They wanted to play on them from the start

Those dang producers! Always get in the way of creative aspiration!

 

[seamonkey]

LLMs do not have the means to do research in a way that exceeds humans' ability to do it.

I just meant 'AI like' tools in general. I can hum or whistle a melody into an app on my smartphone, and it will find songs that match.

LLM are getting better all the time. There's a lot of people voluntarily training them!

 

[beans-on-toast]

I just meant 'AI like' tools in general. I can hum or whistle a melody into an app on my smartphone, and it will find songs that match.

LLM are getting better all the time. There's a lot of people voluntarily training them!

I see the potential of all these things but it scares me. Rather than an assistant as emerging now, people will loose skills and the ability to think. We will become too dependent on the technology, or what it wants to show us.

 

[agedhorse]

I see the potential of all these things but it scares me. Rather than an assistant as emerging now, people will loose skills and the ability to think. We will become too dependent on the technology, or what it wants to show us.

Agreed.

 

[seamonkey]

I see the potential of all these things but it scares me. Rather than an assistant as emerging now, people will loose skills and the ability to think. We will become too dependent on the technology, or what it wants to show us.

Very true. I've lost maths skill since calculators are ubiquitous.

I observe that one driving force of Business adopting AI is to replace low wage, low skill jobs. The kind of jobs that people use to first enter a company through. Get a starting job, then they'd stick around, learn the business, which can take years, and climb up the ladder.

Now, How will people get a starting job in a company? Companies don't have starter jobs that soon won't be replaced by AI or Automation.

Face it though, with software tools, one engineer can do the work of many engineers of the past. It's been sneaking up on us for years.

 

[N. Kentucky Bass]

I observe that one driving force of Business adopting AI is to replace low wage, low skill jobs. The kind of jobs that people use to first enter a company through. Get a starting job, then they'd stick around, learn the business, which can take years, and climb up the ladder.

Beg to differ. The most attractive target to be replaced with GenAI are high paid, high skilled workers. WAY more bang for the buck with dumping these people. Historically, this target has been insolated from tech changes because the tech focused primarily on replacing mechanical processes. GenAI focuses on intellectual processes, the kind that have required college and advanced degrees, and experience. Corporations are zeroing in on this question: Why pay unnecessary high paid workers when GenAI can do it better, the same, or even "good enough" for a fraction of the cost?

 

[agedhorse]

Beg to differ. The most attractive target to be replaced with GenAI are high paid, high skilled workers. WAY more bang for the buck with dumping these people. Historically, this target has been insolated from tech changes because the tech focused primarily on replacing mechanical processes. GenAI focuses on intellectual processes, the kind that have required college and advanced degrees, and experience. Corporations are zeroing in on this question: Why pay unnecessary high paid workers when GenAI can do it better, the same, or even "good enough" for a fraction of the cost?

Actually, I think it's a little bit of both.

 

[beans-on-toast]

Very true. I've lost maths skill since calculators are ubiquitous.

I observe that one driving force of Business adopting AI is to replace low wage, low skill jobs. The kind of jobs that people use to first enter a company through. Get a starting job, then they'd stick around, learn the business, which can take years, and climb up the ladder.

Now, How will people get a starting job in a company? Companies don't have starter jobs that soon won't be replaced by AI or Automation.

Face it though, with software tools, one engineer can do the work of many engineers of the past. It's been sneaking up on us for years.

India has many consulting companies that are used for software development. Those domestic North American jobs have long been lost. Some predict that in the near future, India will suffer a crisis of major job losses in high tech due to AI.

 

[AgingEngineer]

Very true. I've lost maths skill since calculators are ubiquitous.

I've lost spelling skills due to autocorrect...

Last year I was writing on a computer that had spelling auto-correct disabled and discovered that that I have apparently been spelling JUST as JSUT. Along with several other three and four letter words. Thanks to 2 decades of Microsoft Word automatically fixing the mistakes and muscle memory telling me "you're doing great, man!" as pages of typo-free copy scrolled forth, I thought I was a typing machine.

Apparently not.

I had to disable it on my computer just to force myself to relearn how it feels to type these words correctly.

 

[N. Kentucky Bass]

One of my clients for whom I do ghostwriting uses Grammarly, an app to help with correcting grammar and spelling. In my first encounter, Grammarly interjected its automatically suggested ideas so often that I lost focus on what the content was about. Had to shut it off after five minutes, thinking, "This is the perfect app for a helicopter kid."

 

[larrysb]

I honestly feel like there will be a backlash against “machine generated” music.

I’ve said that before on this board and got a very nasty private message calling me an idiot for being against electronica dance and trance. Okie dokie, not what I said.

Already today, there’s some growing backlash against modern production practices of pocketing everything to a time grid, so it is easy to edit and cut and paste as well as against the gratuitous practice of pitch correcting absolutely every vocal track so it snaps to the line.

Both practices take the humanity out performances, and some forms of music are supposed to be raw and expressive, not homogenized and sterile.

Machine learning (properly called) can be used to compose music and lyrics based upon example data. Generative ML methods can successively refine outputs to maximize a value function. That function could hits on website, or purchase conversions while browsing a web store, or mimic and cater to the preferences of large groups people.

The music business isn’t about music. It is really the crowd generation business. An act that draws large crowds because they play so well is no more valuable than an act that draws a large crowd for showing boobies or other outrageous acts.

From a business standpoint it is a lot easier to find and replace boobies than it is to find highly skilled musicians and artists who perform together well and consistently.

Same for machine learning generated music. Falling sales only mean a software upgrade is needed to return to sales growth. That’s just a bill to pay, not the hard work of finding, cultivating and retaining human talent.