Can we just gut copyright. Life of the author +70 years is such an unreasonably long time.
Fish_Totem on
Okay whatever, but they shouldn’t be allowed to literally pirate stuff for training like Meta did. It would cost them relative pennies to pay for it.
ProfessionalCreme119 on
I have a mental image of a Chinese AI engineer walking into work tomorrow, showing this headline to his coworkers and they all share a hearty laugh before getting to work.
maglifzpinch on
So AI is only LLMs now? Nothing of value will be lost then.
yacatecuhtli6 on
Wah wah wah
MayorofTromaville on
Don’t threaten me with a good time.
ThatRedShirt on
I think this is one of the places where I disagree with the rest of this sub pretty strongly, and I’m pretty sure I’m going to get some heat for it. But, normally, this is one of the better subs on economic issues and less willing to give into the popular reactionary, populist narrative.
To start, monopolies are bad. A firm with the sole legal right to use an idea lacks the proper incentive to improve that idea, to reduce costs, provide a better product, etc. That said, innovation and creativity are not a given, and the incentives to develop new solutions to existing problems are just as important as the incentives to be efficient and reduce prices. So, in my view, intellectual property rights (specifically, copyrights and patents), are a good compromise. They provide a clear incentive to innovate. If you can develop a novel product, the government allows you to benefit from it by providing what is, in my view, a temporary monopoly. You benefit either by having a bit of breathing room to take the time to raise capital and develop your product, or you can just sell your intellectual property rights to a firm that already has the means to bring that idea to market.
So, taking as this as my foundation, the belief that patents and copyright exists first and foremost to provide an incentive to innovate, I don’t really see why copyright law should apply so stringently to AI training. The successfulness (from a business prospective) of a piece of art, like a novel, should be judge by how many *people* consume it.
Put another way, what _dis_incentive does allowing AI to add your novel to the pile of the millions of others being fed into this machine create for up-and-coming artists?
I can really only think of two arguments against this?
The first is based on some notion of “fairness,” the artists deserve to make money, and big tech companies have lots of money, so some of that money should go to the artists. At the end of the day, though, this feels very “rent-seeky” to me. People often say that AI is only able to “imitate” art, combining, mixing and rehashing ideas that humans came up with, and therefore, the AI is creating nothing original, so the humans ultimately deserve the credit (and the reward). But I don’t view this much differently than how human’s “create.” We’re free to absorb knowledge (I can get just about any book I want from the library for free), and most of our “creations” are ultimately mixes and rehashes of old ideas, just brought together in new ways. Entire genres (like fantasy) are often all based on the same world-building elements (magic, dragons, wizards, zombies, dwarves, elves, etc) and similar story-telling elements (the hero’s journey, the coming of age, etc.).
The second argument is also a bit “rent-seeky” and protectionist, which is a skepticism of AI, and a belief that creative work should remain the domain of humans. This argument, I believe, ultimately seeks to slow the progress of AI by making its development prohibitively expensive. This sub usually understands the benefits of economic growth, so I won’t belabor this point too much. The one thing I will say is that, I believe, innovation is almost inevitable. And if we aren’t the ones leading the charge on AI (and by we, I mean Liberal nations in general), less scrupulous people will be. I personally believe a big part of why Europe’s seeing such a long period of stagnation (especially compared to the United States) is due to their skepticism and regulation of the technology industry, which has been responsible for so much of the growth and improved productivity elsewhere. The AI race is currently being led by the west (primarily the United States), and I really don’t want that to change.
7 Comments
Can we just gut copyright. Life of the author +70 years is such an unreasonably long time.
Okay whatever, but they shouldn’t be allowed to literally pirate stuff for training like Meta did. It would cost them relative pennies to pay for it.
I have a mental image of a Chinese AI engineer walking into work tomorrow, showing this headline to his coworkers and they all share a hearty laugh before getting to work.
So AI is only LLMs now? Nothing of value will be lost then.
Wah wah wah
Don’t threaten me with a good time.
I think this is one of the places where I disagree with the rest of this sub pretty strongly, and I’m pretty sure I’m going to get some heat for it. But, normally, this is one of the better subs on economic issues and less willing to give into the popular reactionary, populist narrative.
To start, monopolies are bad. A firm with the sole legal right to use an idea lacks the proper incentive to improve that idea, to reduce costs, provide a better product, etc. That said, innovation and creativity are not a given, and the incentives to develop new solutions to existing problems are just as important as the incentives to be efficient and reduce prices. So, in my view, intellectual property rights (specifically, copyrights and patents), are a good compromise. They provide a clear incentive to innovate. If you can develop a novel product, the government allows you to benefit from it by providing what is, in my view, a temporary monopoly. You benefit either by having a bit of breathing room to take the time to raise capital and develop your product, or you can just sell your intellectual property rights to a firm that already has the means to bring that idea to market.
So, taking as this as my foundation, the belief that patents and copyright exists first and foremost to provide an incentive to innovate, I don’t really see why copyright law should apply so stringently to AI training. The successfulness (from a business prospective) of a piece of art, like a novel, should be judge by how many *people* consume it.
Put another way, what _dis_incentive does allowing AI to add your novel to the pile of the millions of others being fed into this machine create for up-and-coming artists?
I can really only think of two arguments against this?
The first is based on some notion of “fairness,” the artists deserve to make money, and big tech companies have lots of money, so some of that money should go to the artists. At the end of the day, though, this feels very “rent-seeky” to me. People often say that AI is only able to “imitate” art, combining, mixing and rehashing ideas that humans came up with, and therefore, the AI is creating nothing original, so the humans ultimately deserve the credit (and the reward). But I don’t view this much differently than how human’s “create.” We’re free to absorb knowledge (I can get just about any book I want from the library for free), and most of our “creations” are ultimately mixes and rehashes of old ideas, just brought together in new ways. Entire genres (like fantasy) are often all based on the same world-building elements (magic, dragons, wizards, zombies, dwarves, elves, etc) and similar story-telling elements (the hero’s journey, the coming of age, etc.).
The second argument is also a bit “rent-seeky” and protectionist, which is a skepticism of AI, and a belief that creative work should remain the domain of humans. This argument, I believe, ultimately seeks to slow the progress of AI by making its development prohibitively expensive. This sub usually understands the benefits of economic growth, so I won’t belabor this point too much. The one thing I will say is that, I believe, innovation is almost inevitable. And if we aren’t the ones leading the charge on AI (and by we, I mean Liberal nations in general), less scrupulous people will be. I personally believe a big part of why Europe’s seeing such a long period of stagnation (especially compared to the United States) is due to their skepticism and regulation of the technology industry, which has been responsible for so much of the growth and improved productivity elsewhere. The AI race is currently being led by the west (primarily the United States), and I really don’t want that to change.