'Music is as good as gold or oil': A man spending billions on old hits
theguardian.com
theguardian.com
Article I, Section 8, Clause 8, U.S. Constitution. https://www.law.cornell.edu/constitution-conan/article-1/sec...
(Clearly, US application.)
Whereas Printers, Booksellers, and other Persons, have of late frequently taken the Liberty of Printing, Reprinting, and Publishing, or causing to be Printed, Reprinted, and Published Books, and other Writings, without the Consent of the Authors or Proprietors of such Books and Writings, to their very great Detriment, and too often to the Ruin of them and their Families: For Preventing therefore such Practices for the future, and for the Encouragement of Learned Men to Compose and Write useful Books; May it please Your Majesty, that it may be Enacted ...
Statute of Anne (England, 1701). http://www.copyrighthistory.org/cam/tools/request/showRepres...
The countries of the Union, being equally animated by the desire to protect, in as effective and uniform a manner as possible, the rights of authors in their literary and artistic works,...
Berne Convention https://wipolex.wipo.int/en/text/283698
Courts may, and do, invalidate law (or interpretation) inconsistent with Constitutional mandate or legislative intent.
The most recognizable feature of today's internet is the web, which wasn't part of ARPANET.
You could say the internet is a descendant of ARPANET, but it's not the same thing.
And even if the argument if flawed, the concept underlying it is solid: what something was "originally meant for" is not the same as what something was actually usef for, or found eventually useful for....
It's not an abuse based on its actual, de facto, purpose: to sustain profitable back catalogue sales and aid the entertainment industry.
So whether the "original purpose" is important is the whole point in determining abuse or not...
I'm sure that, if you are serious about the discussion, we can easily agree that both are not equivalent or comparable.
Not exactly. When the grandparent (metalliqaz) wrote "This is not what copyright is for", I felt like they were the one confusing/conflating the "spirit of the law"(about copyright) with what copyright is actually used for, and intented to be used for, by the parties passing copyright laws and benefiting from them.
Copyright might not have been "for this", in some original "spirit of the law" sense, but it very much is for this in the de facto, 70 years plus running, sense.
So, I'm saying that there's an expressed spirit of the law and an operating (de facto) spirit of the law. Or, if you will, what the law hypocritically pays lip service to, and what the law is actually drafted and passed to achieve.
Those laws might have started back in the day with the intention (spirit) to promote the arts and protect creators etc, but for nearly a century they have been extended with the intention of making the entertainment business more profitable and protecting a multi-trillion US industry.
That's the actual spirit those laws are drafted, proposed, passed, and enforced for 70+ years now...
Few (if any) in the lawmaking and in the industry cares for the original "spirit of the law", the original wording is just left there as a historical artifact, to justify a different (and even opposing) application of the copyright law.
When filing for a patent or copyright, say “I believe this is worth $1m” — so you pay a $20k tax on filing/renewing it. If you’re not confident that your idea can actually redeem 2% of that stated value, then maybe it’s too risky to actually be worth that much. So you say it’s worth $100k because you can definitely make at least $2k from it.
But if you’re really only confident in making $2k from your idea, and someone else is allowed to “buy license to use your material” at $100k, you’d be absolutely floored to make so much money from that idea. So if you end up being wildly successful, you’ve still made a lot of money (and can keep using your idea), but now competition/reuse is allowed past your stated value threshold.
As a society, we want land put to valuable use. We don't want a piece of super valuable land (could be producing food, etc) that is dormant simply because someone refuses to sell.
But we also don't want to be the arbiter of what counts as valuable. So you state the value of the land and pay the tax on that value. If someone else can extract more value, they can buy it from you at the stated price.
I think this sort of tax makes less sense for intellectual property, because unlike land many people can use the same idea at once.
Not just directly (e.g. costs to grant patents, hear court cases etc.) but indirectly - people can make things completely on their own and end up being sued for it because of an obscure patent violation. This has a chilling effect on innovation in general.
Because of this, making the person with the special right have to pay for the costs of that right seems fair.
However, concentration of valuable IP in the hands of the few is also a problem. Limited duration copyright provides the greatest benefit to individual authors (by maximizing their bargaining power while limiting the bargaining power of IP-hoarders through erosion) and the public domain.
I can’t see how that would work now that the data is digitised and the filings are completely automated. At best it might work for some older music, but it wouldn’t work for anything produced after the rule was introduced because services would be created to automatically do the filing for artists.
An immeasurable treasure of music and other works of art in the public domain would still be in the hands of these rentiers on that basis.
You used to need to file a copy with the Library of Congress to obtain copyright. Then they'd start filing it, but immediately taking it out on loan, and lose it. Then they'd "file" it and take it out on loan before ever giving it over; the copy for the library never existed. Then they decided to just give up on the farce, and not make you file.
To renew a work, you should have to actually file a (non-DRM) copy with the Library of Congress, and not take it out on loan. If it's important enough to renew, it's important enough for the Library to demand an actual copy.
Now, I don't support having to file with the Library of Congress for the initial copyright, that'd disproportionately hurt independent people who publish on the Internet.
"IP" is an odd collection collection of things that actually work pretty different from each other. Patent limits the use of some kinds information; copyright limits the copying of other kinds of information. Both of these are naturally "non-rivalrous" - I can watch a movie or build an engine at the same time that you watch that same movie or build that same (kind of) engine and we don't interfere with each other. We artificially treat them as if they were rivalrous - where "there are only so many to go around" - so that the market will better reward their production. Letting them eventually fall into the public domain, however, ameliorates the dead weight loss we're causing by artificially driving up the price.
Trademark is a bit different. Someone choosing to do business with me, because they have had good experiences doing so in the past or heard about me from a friend, is a limited resource in a way information alone is not. Letting others confuse that by providing their own goods and services under my mark doesn't increase the total value available.
So your question isn't really a natural one. These things work differently and it makes sense that we treat them differently.
Having said that, your question does make me think about an issue I've seen crop up occasionally, where a long-standing brand known for high quality will self destruct by cutting the quality without cutting prices. If the reduction in quality allows for sufficient cutting of expenses, and the reduction in quality is not sufficiently, immediately obvious, this can radically increase margins without significantly decreasing volume and make people a lot of short term profit until the suckers buying the product figure it out. On the one hand, simply capping the length would presumably make this worse, not better. On the other hand, making them contingent on something beyond a certain point might be an interesting space to look for a solution.
Trademarks protect the consumer
For example, you would see "YC for budding young novelists" if copyright were more long-term and secure.
You can pay the living expenses of a whole lot of would-be-starving-artists-but-actually-baristas, for an enormous number of years, if you can collect investment revenue from the one who turned out to be Stephen King.
This idea is an extension of what copyright is for, which it to allow creators to make money, thus incentivizing them.
I support unlimited-duration copyright.
even with unlimited-length copyright, who would do that? if I'm the investor, I would just find a bunch of aspiring artists and offer to pay their living expenses for a year in exchange for ownership of the copyright. I'd ditch the ones that didn't make money and offer better terms to the successful ones for their next work. with unlimited-length copyright, I might be willing to give more people a one-year shot, but the end result would still be that most of them go back to their barista job.
Wow. That's an unpopular position around here!
Even among those of us who support copyright and cringe at the advocacy for piracy you often see in these threads, I think there is a lot more support for shortening copyright duration.
Zero copyright facilitates exploitation by certain corporations (telecoms, media giants), while unlimited copyright facilitates exploitation by a different set of corporations (IP-hoarding publishers). Limited duration copyright is the best option.
> unlimited copyright facilitates exploitation by a different set of corporations
"Exploitation" is a loaded word. Nobody is going to pay $500 to read Harry Potter. What you would see with unlimited copyright is what you already see: there is a reasonable maximum that can be charged for (say) books in a given market. In other markets (e.g. 3rd world), prices are lower.
Scientific publishing is an entirely different issue, BTW. Science journal publishers rape universities and the general public. That is not really because of copyright. It's a different system of abuse that I don't want to raise here.
Does anyone think that way? They shouldn't, because for almost all books, the income past 20 years is zero.
Now, corporations are a bit different, and if you're Disney making a movie, the revenue past 20 years is actually something you consider. Still... would the Avengers movies, say, not get made without 95 year copyright? I doubt it. They made enough money in the first year that they would still have been made even if copyright was quite short.
So I can't buy this argument that "almost-forever" copyright incentivizes creation. I mean, it does, to some degree. But with the current duration, we are ridiculously far past the law of diminishing returns.
This is false. It helps corporations and hardly anyone else. There was incentive to create music and other works before copyright and there would still be without it.
I really don't get this. Other people have already replied to you but not specifically on "long-term" vs your later "unlimited-duration" stand.
The USA adopted the Berne Convention in 1989. It stipulates that the duration of the term for copyright protection is the life of the author plus at least 50 years after their death. For some categories of works, the minimum duration is shorter. For example, the minimum term for applied art is 25 years. Movies have a minimum term of 50 years. Countries may choose a longer term of protection, and most countries have done so.
If you're a potential "budding young novelist", will you eschew that career because life + 50 years isn't unlimited and isn't long enough for you?
https://en.wikipedia.org/wiki/List_of_countries%27_copyright...
My proposal would undoubtedly allow literary agencies to take more chances, funding more new authors for longer periods of time.
Check out the Tupac Shakur voice on https://vo.codes
It's kind of broken, but it took zero effort. And that's the direction we're headed.
r9y9 on Github has amazing results synthesizing lyrics. Better than vocaloid at this point.
Look what people are making:
https://www.reddit.com/r/VocalSynthesis/comments/i3egph/even...
https://m.soundcloud.com/r9y9/sets/dnn-based-singing-voice
My prediction is that in ten years, music copyright is going to be very different than it is today. We'll still have live artists putting on performances, but we'll also have entirely virtual artists like Gorillaz (with neural-backed lyrics) and recreations of Tupac, MJ, Elvis, and so on.
100% agree. I like the indie genre better than pop, though even that's somewhat manufactured by labels as you say.
I recently found some incredibly underrated Zelda remixes that are jaw dropping and have barely been heard at all: https://youtu.be/4r9S2yEiuME
So I agree with your argument.
A new factor in the equation, however, is how much cheaper and less labor intensive it becomes to create new music. If you make it so that anybody can produce vocals, that changes the game for a lot of genres (eg. EDM).
But these techniques won't just be viable on vocals. Everything is going to be easier to create. Melodies, instrumentals, the full stack.
That's going to have a profound effect and definitely rock the boat the incumbents are riding.
With DAWs, home studios, VSTs, 24/96 interfaces, and so on, and most hits being electronic anyway, plus widespread mastering and mixing knowledge and services, the margins to create new music are already as low as it gets.
Cost is not the problem, we have the exact opposite issue: overproduction without corresponding demand, and with difficult to get listener attention.
The ones making it wont be the ones producing more music cheaper, but the ones promoting their music best.
Those are 1.6 million views. For mere Zelda remives.
Hugely influential albums in the past and hugely important works in all genres have gotten less than 100K buyers. The Velvet Underground first album is said to have initially sold 30,000 units in its first years.
And let's not go into bands like The Residents, The Meat Puppets, The Long Ryders, and so on.
1.6M for some ...Zelda remixes (of the 1000s that have been made) are hardly "barely heard at all".
1,601,808 views.
That's about 1.6 million more views than I would expect from "Zelda remixes". So we both see artistic injustice in the world, just from a different angle.
I will grant you it is better than 1982's "Pac Man Fever" though I suspect that made more money.
It's very rare for me to really like a song the first time I hear it. Once my brain has become accustomed to where the song goes (even for a jazz song), then I tend to like it a lot more. Another layer to this is the connection that a song can have to a certain period in your life, bringing back the emotions of that time with it.
I think there is probably a wide range of human experience in this regard, since I have the complete opposite experience: I know with extremely good certainty (~95%) how much I like a song the first time I hear it.
To your point, while the value of old songs usually drops over time, I don't see new music, AI-generated or not, driving it down more. A familiar song takes you back to a different time and place, and there's no substitute for that.
"thug life" is kind of a horrifying one to hear. On another note it would be interesting to have a sort of test for these sort of systems, if they can say x sentences without screwing up, they can say almost anything. A "turing" test if you will.
What does "zero effort" mean? I'd be surprised if you told me that Tupac's voice was recreated without using any copyrighted material.
This fascinates me and opens up a wholly new can of worms.
Technically, it's mostly Linda Johnson's speech data, which is in the public domain. Most voice models incorporate this as the base for learning how to transform phoneme to spectrogram, spectrogram to waveform. They'll then transfer learn on top of this.
Could you tell the difference between Tupac and an impersonator? What about 10% Tupac, 90% impersonator?
What if it's really 1% Tupac, 9% impersonator, 90% Linda Johnson? Or no Tupac at all? Can anyone tell the difference?
Where did the copyrighted data go? Is there any left?
Probably not, but learning to impersonate Tupac isn't "zero effort" (and I'd imagine that it required access to copyrighted material).
To be explicit: copyright is awful and I'd like to see it abolished. It's important to realize that nifty stuff like Tupac vocoders are not "zero effort" and require public access to data that intellectual propertarians want to have monopolies on.
That's not the same as artist connection to the audience (which is a big thing even for pop music), or popularity.
We've had such schemes for ages, even without AI (e.g. 3D or anime artists with songwriting teams behind them) but outside of a few cases (mostly in Japan) they're irrelevant...
A single case, so not enough to generalize of any major demand for this sort of thing. And they played off the people knowing it's Albarn behind the scenes (who was already popular with Blur).
OF COURSE it will feel formulaic. You will never get anything but as long as you're simply mashing different source training material together, which is all these things are doing
Why do you want to make artistry extinct?
If you look at history, art has been transformed more than once. First, when we managed to make enough food that some people could be full time artists. Second, when we invented mass media, which made most village artists unemployed while elevating a select few.
This! This is the future I want to live in and help create.
People are going to have all sorts of media tweaked to their interests. The system will have to be smart enough to introduce novel changes in interest so that it never gets boring, or so that people's interests can be expanded.
If we're careful, we can characterize and destroy the filter bubble while we're at it.
How can you appreciate the things you like if you never have to experience things you don't like?
Life isn't supposed to be a constant pursuit of cutting out any unwanted or negative things. It's about experiencing things in general.
A world like you describe sounds pretty bland and lifeless to me.
This sounds like the anti-rock and roll argument wearing a different cloak.
> How can you appreciate the things you like if you never have to experience things you don't like?
I doubt this would remove all displeasure and pain from life. We're a long way away from dopamine drips and AI overloads.
> Life isn't supposed to be a constant pursuit of cutting out any unwanted or negative things.
Life isn't supposed to be about anything. We're apes on a mote of dust in an infinite void.
> It's about experiencing things in general.
This increases diversity of experience. It's something new. People aren't going to stop making music because it got easier. That strikes me as a Luddite argument.
> A world like you describe sounds pretty bland and lifeless to me.
The world gets better every year. These arguments are, frankly, a nostalgia for youth.
Times change.
Says who? My goal is 100% to cut out all unwanted things. Even going so far that I only want things that I know bring me joy. Automate all chores and stop doing anything that I have the slightest doubt about. I'm at a stage now where I don't care if I never experience anything new. I could die happy continuing doing what I'm doing for the rest of my life.
1) It would do irreparable damage to the social elements of listening to music. The question "have you listened to Foo, by Bar?" would be obsolete, because of course they haven't, it was generated just for you.
2) Any music with lyrics generated or tweaked by such a system would ring completely hollow, with the knowledge that it's not actually words a person wrote. Worse, if this is common enough, there will always be doubts about whether a purportedly human-written song was actually written by a human or not.
3) Put yourself in the shoes of an artist forced to confront that no one is actually listening to the song they wrote, but merely automatically generated variations on it which they did not compose and do not feel ownership of.
> If we're careful we can characterize and destroy the filter bubble while we're at it.
I do not understand how you reached that conclusion. I think the most likely outcome is the exact opposite occurring: a reinforcement of filter bubbles and social isolation.
As a more general point: I've come to see the drive towards "personalization" (and by implication commoditization) of everything is a social sickness. It denies shared experience, and leaves us feeling more alone.
As an example, one of the beautiful things about music is that because it stirs our emotions so directly, it allows us to participate in extremely meaningful shared experiences with other listeners. Music is a vehicle for communication.
One of the last things in modern life that brings people together is live music. We can argue whether AI can actually produce art or merely a veneer of art, but there are far more talented musicians, writers, and filmmakers in the world than will ever be able to make a living at it.
So why crowd them out by flooding the zone with machine-generated shit? Why customize the culture so much that there are no more common experiences, that every person is a musical island?
The people listening to it now will listen to it in 30 and 50 years. And, judging by social media comments, teenagers in that era will also appreciate it.
There used to be this notion that only "great" music survives (probably judging from classical music or the history of literature), and "transient" pop hits don't, but we've seen every one-hit-wonder and tons of pop hits from the 70s and 80s survice just fine, and even get new listeners.
It's basically a kitch pop hit, but it's not just people listening to Toto's Africa in the 1980s that appreciate it now. Or stuff in Madonna's 80s catalogue, etc.