Stop listening to Spotify or Apple Music or any other streaming source. Respect music and musicians and songwriters enough to listen to their source recordings. God I can't wait for these platforms to go belly up!
After resisting (although I use Apple computers, I deep down hate the damn company), I signed up with Apple Music.
I am two months in and underwhelmed. Lousy search function. When you find an artist's niche, it's attenuated for no discernable reason. The algorithm is deeply convinced I want to listen to nothing but HipHop and Rap (won't touch the crap). Pushes weird new musicians and buries anything remotely interesting--and they have a complete record of everything I've bought from their store so they know my tastes...except they don't.
I'm not sure that the majority of people really care whether their "music" is AI generated or not. It seems to me that most people regard music as something to have on in the background. They don't care what it is as long as they do not have to endure "dreaded silence". For me, AI generated pap, or even musician generated pap, elicits a visceral reaction that sets my teeth on edge. I want real, quality music on in the background even less, because then my attention is drawn to the music and I can't do the thinking that is required for the task at hand.
I recently retired from being a veterinarian because I am tired of having the fight that I cannot work or think with music going in the background. A decade ago no one had radios going in the surgery or working areas of a clinic. Now, everyone seems to have "music" going. All I hear is a thumping beat that makes an already uncomfortably loud working environment (think barking dogs, clanking metallic surgical instruments and busy vets and nurses communicating amplified by hard, shiny, cleanable surfaces), even louder as people need to shout over the top of each other to be heard.
Millennials and Gen Z have been set up for the scenario the streaming services are hustling. They were pushed away from music and arts education and toward various computer skills classes in an attempt to create a pipeline of "college- and career-ready" students, which is yet another way MBAs disguised as educational administrators have conspired across levels to kill liberal arts education. But, I digress....
If you don't have a basic appreciation of melody and harmony and never actively listen to music (and it's all just background for multitasking), then bass and beats with a half-ass topline might be all that you want. The streaming services are counting on this for younger listeners and for their older subscribers to continue grazing on the music of their salad days. As Ted writes, this ain't gonna work for long. But...
Once the technology gets cheap enough and sophisticated enough, I envision that the tech titans will simply sell you a subscription-based, Pandora-like AI app that generates music on the fly based on your location, prior listening habits, age, Fitbit-detected heartrate, recently visited online marketplaces, etc. or sends you other users' previously generated AI tracks, which are uploaded to a cloud service when the user performs a specific action (clicks a "Like" button, nods head or taps the desk to the beat while the phone or laptop camera and mic watches and listens--with "permission" of course). This generated music will sound vaguely familiar because the models will train on thousands if not millions of song snippets and motifs, yet it will be different enough to avoid copyright infringement claims. </music dystopia>
Cantos' reference to Harrison Bergeron is on the money. The move to 'purist' music and the 'reinvention' of vinyl is music taking its rightful place again. Take a minute to select the vinyl pressing, take it slowly out of its protective sheath, place it tenderly on the turntable, lower the needle, and take the time to actually listen. It's like making coffee from scratch. It's the enjoyment of the moments that lead up to the thing itself. Love your writing. Thank you.
Your list of candidates for advancing this cause should include ASCAP and BMI, whose job it is to ensure artists get the royalties they're already owed. This probably has some ramifications for training data as well.
Quality, yes. But why would this not go the way of the drum machine? How's that worked for drummers? Rhythm boxes appeared in 1938 and now you can program a Vinnie Caliuta part and unfortunately it will sound pretty not bad. Years ago I did a track with some late 90's drum samples and live band over the top of a 5/4 shuffle. With what was unlimited budget, I brought Vinnie in to "improve" it. Had a great time. Studio filled with his kit and wit and feel...fun. Had to admit, after the bills were paid, didn't improve it enough to notice, but we felt better. These are early days. AI is in training. We're training it. Every voice message saying this call will be recorded is training the operator replacement. Question is: is this the hill to die on? I vote yes. It was 50 years ago. It's human intellectual and creative property, pay for it. It's time to get creative about how to do that. I tried the union route...failed. Important thread, Ted, as usual.
Thanks for this piece. It is an accurate assessment of AI’s talents and ability in music. The same goes for what AI will offer us in the area of written fiction. No doubt there will be all sorts of “tinkly generic improvisations” (your fine phrase for the “music” [note the quotation marks] that AI can produce.) It will be the same for novels and short stories. The finest in these genres have always come traditionally from the frayed, engaged, beating human heart and its difficulties…the emotions, in other words, expressed in great tragedy and fine comedy, always with gorgeous language. AI is not capable of such forms of expression. Especially when the language used is an AI amalgam intended to sound like a particular human writer’s style, it will be tinkly and poorly improvised, a bad imitation, laughable in the extreme, and worthy of lawsuits from the authors being imitated. What makes great fiction so individualistic is its very unique language, which can be made up and expressed only from a particular writer’s soul. A human writer, in other words. Machine learning’s efforts at Faulkner will result only in the reader’s laughter, at Melville in derision, at Wharton in wounded chuckles, at Sebastian Barry in the reader’s effort to throw the laptop on which he/she is reading at the wall. And original fiction? Get outta here! There will no doubt be plenty of such “writing.” But it will always be dunderheaded and lame. Those who read it will deserve what they get.
I agree with @SeanDennis: this is the inevitable result of removing arts education from the school system. The human heartbeat was once inspiration for the Bossa Nova....now it's just data in a global Fitbit bin.
I'm occasionally asked if I've heard any good new music, and I always say yes, but...it's not on the radio or through streaming (unless I get a referral; someone telling me to check out this artist). It's all local. I live in the Seattle metropolitan area and there's more good music than you can shake a stick at. That's where it's at. Whether there'll be a truly supportable monetary system for musicians I don't know, but I know where good music can be found.
The premise, 'AI is bad for music,' might be more aptly phrased to 'AI is challenging to some music,' especially the loop-based products any 9th grader can cut and paste together before homeroom at 8:00 a.m.
What music is 'inferior' to some may be 'acceptable' to others. Judgment often reflects the judge more than that being judged.
As an artist who uses some AI tools, I find the prospect of surrendering to an 'organizing' power far more compromising to creative production than co-existing with bad taste in robots. Having a 'leader and organization' is creating another problem easily influenced by lobbies.
I'm not a participant or consumer of social media or Spotify but I do understand their value to some but would expect their AI standards satisfy popular culture's market. That's a different market than those who would attend a live performance of a Philip Glass work.
Ok, I'll state the obvious. The problem is not AI. The problem is the US has several generations who have not been exposed to the arts in their public education experience. It is far too late to re-educate them, but not their kids. We do not have a critically aware art/music consumer base. The evidence for this is so overwhelming it is difficult to see until we step back from the fear of AI and look what people are making and consuming.
If we're actually afraid of AI making crap better than people, then I'd say that fear is well founded. If we're afraid AI will dominate that market, great.
But, AI still needs a Taylor Swift to emulate her voice just as Epiphone needs A Gibson to make a Gibson-like guitar.
The problem is we in the US have a population which can't or won't differentiate between the original and the copy.
The challenge for artists and musicians is to make work authentically yours and copyright it. Use the technology we have rather than scare us into some union organization abyss which only guarantees dues, not income and critical reviews.
Today, the marketplace is fair insomuch as a demand-based market of entertainment can be saturated while still making a profit. Not everyone gets the big deals, but Steve Reich has never had a Top 40 hit. 'Value' has many faces.
Some would like to apply a Marxist commodity ideology to the AI music market in that the sound does not reflect or identify its creator so its value must be less. A lot of experimental musicians have toyed with that concept for some time.
For myself, fear and cynicism aren't helpful to musicians and artists who, for a variety of reasons, aren't making a living soley off their work. Open minds open more doors to opportunities.
Stop listening to Spotify or Apple Music or any other streaming source. Respect music and musicians and songwriters enough to listen to their source recordings. God I can't wait for these platforms to go belly up!
After resisting (although I use Apple computers, I deep down hate the damn company), I signed up with Apple Music.
I am two months in and underwhelmed. Lousy search function. When you find an artist's niche, it's attenuated for no discernable reason. The algorithm is deeply convinced I want to listen to nothing but HipHop and Rap (won't touch the crap). Pushes weird new musicians and buries anything remotely interesting--and they have a complete record of everything I've bought from their store so they know my tastes...except they don't.
It's weird, mind-melding junk andtypically Apple-arrogant.
Yes, I've discovered some great music but not with any help from Music's front end.
And if I drop, the music goes poof, disappears.
Orwellian.
I'm not sure that the majority of people really care whether their "music" is AI generated or not. It seems to me that most people regard music as something to have on in the background. They don't care what it is as long as they do not have to endure "dreaded silence". For me, AI generated pap, or even musician generated pap, elicits a visceral reaction that sets my teeth on edge. I want real, quality music on in the background even less, because then my attention is drawn to the music and I can't do the thinking that is required for the task at hand.
I recently retired from being a veterinarian because I am tired of having the fight that I cannot work or think with music going in the background. A decade ago no one had radios going in the surgery or working areas of a clinic. Now, everyone seems to have "music" going. All I hear is a thumping beat that makes an already uncomfortably loud working environment (think barking dogs, clanking metallic surgical instruments and busy vets and nurses communicating amplified by hard, shiny, cleanable surfaces), even louder as people need to shout over the top of each other to be heard.
Millennials and Gen Z have been set up for the scenario the streaming services are hustling. They were pushed away from music and arts education and toward various computer skills classes in an attempt to create a pipeline of "college- and career-ready" students, which is yet another way MBAs disguised as educational administrators have conspired across levels to kill liberal arts education. But, I digress....
If you don't have a basic appreciation of melody and harmony and never actively listen to music (and it's all just background for multitasking), then bass and beats with a half-ass topline might be all that you want. The streaming services are counting on this for younger listeners and for their older subscribers to continue grazing on the music of their salad days. As Ted writes, this ain't gonna work for long. But...
Once the technology gets cheap enough and sophisticated enough, I envision that the tech titans will simply sell you a subscription-based, Pandora-like AI app that generates music on the fly based on your location, prior listening habits, age, Fitbit-detected heartrate, recently visited online marketplaces, etc. or sends you other users' previously generated AI tracks, which are uploaded to a cloud service when the user performs a specific action (clicks a "Like" button, nods head or taps the desk to the beat while the phone or laptop camera and mic watches and listens--with "permission" of course). This generated music will sound vaguely familiar because the models will train on thousands if not millions of song snippets and motifs, yet it will be different enough to avoid copyright infringement claims. </music dystopia>
The use of AI in any medium needs to be fully disclosed ... so people can choose the artisanal, local, human musicians with clarity.
Cantos' reference to Harrison Bergeron is on the money. The move to 'purist' music and the 'reinvention' of vinyl is music taking its rightful place again. Take a minute to select the vinyl pressing, take it slowly out of its protective sheath, place it tenderly on the turntable, lower the needle, and take the time to actually listen. It's like making coffee from scratch. It's the enjoyment of the moments that lead up to the thing itself. Love your writing. Thank you.
Your list of candidates for advancing this cause should include ASCAP and BMI, whose job it is to ensure artists get the royalties they're already owed. This probably has some ramifications for training data as well.
Quality, yes. But why would this not go the way of the drum machine? How's that worked for drummers? Rhythm boxes appeared in 1938 and now you can program a Vinnie Caliuta part and unfortunately it will sound pretty not bad. Years ago I did a track with some late 90's drum samples and live band over the top of a 5/4 shuffle. With what was unlimited budget, I brought Vinnie in to "improve" it. Had a great time. Studio filled with his kit and wit and feel...fun. Had to admit, after the bills were paid, didn't improve it enough to notice, but we felt better. These are early days. AI is in training. We're training it. Every voice message saying this call will be recorded is training the operator replacement. Question is: is this the hill to die on? I vote yes. It was 50 years ago. It's human intellectual and creative property, pay for it. It's time to get creative about how to do that. I tried the union route...failed. Important thread, Ted, as usual.
Hello Ted:
Thanks for this piece. It is an accurate assessment of AI’s talents and ability in music. The same goes for what AI will offer us in the area of written fiction. No doubt there will be all sorts of “tinkly generic improvisations” (your fine phrase for the “music” [note the quotation marks] that AI can produce.) It will be the same for novels and short stories. The finest in these genres have always come traditionally from the frayed, engaged, beating human heart and its difficulties…the emotions, in other words, expressed in great tragedy and fine comedy, always with gorgeous language. AI is not capable of such forms of expression. Especially when the language used is an AI amalgam intended to sound like a particular human writer’s style, it will be tinkly and poorly improvised, a bad imitation, laughable in the extreme, and worthy of lawsuits from the authors being imitated. What makes great fiction so individualistic is its very unique language, which can be made up and expressed only from a particular writer’s soul. A human writer, in other words. Machine learning’s efforts at Faulkner will result only in the reader’s laughter, at Melville in derision, at Wharton in wounded chuckles, at Sebastian Barry in the reader’s effort to throw the laptop on which he/she is reading at the wall. And original fiction? Get outta here! There will no doubt be plenty of such “writing.” But it will always be dunderheaded and lame. Those who read it will deserve what they get.
Terence Clarke
terenceclarke.substack.com
I agree with @SeanDennis: this is the inevitable result of removing arts education from the school system. The human heartbeat was once inspiration for the Bossa Nova....now it's just data in a global Fitbit bin.
I'm occasionally asked if I've heard any good new music, and I always say yes, but...it's not on the radio or through streaming (unless I get a referral; someone telling me to check out this artist). It's all local. I live in the Seattle metropolitan area and there's more good music than you can shake a stick at. That's where it's at. Whether there'll be a truly supportable monetary system for musicians I don't know, but I know where good music can be found.
Another urgent + perceptive call to action. Musicians – you all listening out there?
Too true. Gawd, Taylor Swift is going to be busy spearheading all your ideas.
The premise, 'AI is bad for music,' might be more aptly phrased to 'AI is challenging to some music,' especially the loop-based products any 9th grader can cut and paste together before homeroom at 8:00 a.m.
What music is 'inferior' to some may be 'acceptable' to others. Judgment often reflects the judge more than that being judged.
As an artist who uses some AI tools, I find the prospect of surrendering to an 'organizing' power far more compromising to creative production than co-existing with bad taste in robots. Having a 'leader and organization' is creating another problem easily influenced by lobbies.
I'm not a participant or consumer of social media or Spotify but I do understand their value to some but would expect their AI standards satisfy popular culture's market. That's a different market than those who would attend a live performance of a Philip Glass work.
Ok, I'll state the obvious. The problem is not AI. The problem is the US has several generations who have not been exposed to the arts in their public education experience. It is far too late to re-educate them, but not their kids. We do not have a critically aware art/music consumer base. The evidence for this is so overwhelming it is difficult to see until we step back from the fear of AI and look what people are making and consuming.
If we're actually afraid of AI making crap better than people, then I'd say that fear is well founded. If we're afraid AI will dominate that market, great.
But, AI still needs a Taylor Swift to emulate her voice just as Epiphone needs A Gibson to make a Gibson-like guitar.
The problem is we in the US have a population which can't or won't differentiate between the original and the copy.
The challenge for artists and musicians is to make work authentically yours and copyright it. Use the technology we have rather than scare us into some union organization abyss which only guarantees dues, not income and critical reviews.
Today, the marketplace is fair insomuch as a demand-based market of entertainment can be saturated while still making a profit. Not everyone gets the big deals, but Steve Reich has never had a Top 40 hit. 'Value' has many faces.
Some would like to apply a Marxist commodity ideology to the AI music market in that the sound does not reflect or identify its creator so its value must be less. A lot of experimental musicians have toyed with that concept for some time.
For myself, fear and cynicism aren't helpful to musicians and artists who, for a variety of reasons, aren't making a living soley off their work. Open minds open more doors to opportunities.
ted if one day you go missing we will know it’s because spotify took you out
Apple Music now (iOS17) gives full credits for each song. Musicians, producers, songwriter etc