There are crappy combination padlocks that can be picked by turning the combination wheels and pulling on the shackle. When you've turned a wheel into its unlocking position the shackle will shift open slightly, signalling that you've cracked that digit. So a padlock with 3 digits could be opened in at most 30 moves, instead of 1000.
When you have a selection mechanism, the randomness of mutations becomes kind of moot. Each step of the way that produces a new adaptation only has a handful of optimal (or optimal-enough) solutions. Now, suppose that we're an underwater species that's in the process of developing light-sensitive cells, and from our present genome there are 2^64 possible genes that will produce a good protein for the transduction step (of converting light into some other form of energy). Do we need to mutate 2^64 times to find the "correct" gene? No, all of those 2^64 genes are "correct". Our descendants should not fall into the trap of thinking that because they got gene #5541741487894936799 that it was a special outcome. They could have just as easily gotten gene #5541741556614413535.
>How would evolution slowly build a heart with no blood or blood vessels? How would evolution slowly build blood but no blood vessels or heart? How would evolution slowly build a network of blood vessels but no blood to carry or heart to pump?
You start with a fluid-filled body cavity where oxygen is transported by dissolution from the outside to the inside. This fluid already contains oxygen-transporting cells. Then you section off part of the cavity to enable oxygen-transporting cells to move more efficiently. At this point the OTCs are completely separated, so the fluid inside the conduits can be called blood. The animal can pump blood by squeezing its body during its normal locomotion. Eventually muscle cells are added to the circulatory system to enable oxygen circulation independently of the animal's movement, as well as to circulate it even more effectively. Later on the muscle cells become centralized as it's more efficient, not to mention that a big single muscle can pump more strongly than a distributed system of tiny muscles. Each step of the way you have a functional organism and each form performs the function of transporting oxygen more effectively.
I know you're going to say that this is a "must have" explanation, but you merely asked for an explanation. The above is a plausible series of events that could have led to the circulatory system as we see it today, so if you want to argue that the circulatory system couldn't have evolved, you'll need to argue why this explanation isn't plausible.
EDIT: Also, I can't help but notice you ignored the question I asked in my last paragraph.
I mean, everything would seem to indicate that mitochondria were assimilated while our ancestors were still microbes themselves. But that aside, our species is the result of everything that happened prior to its appearance, including for example the Cretaceous–Paleogene extinction event. Arguing that we should not attempt to prevent such events (if we're able to) because it would evolutionarily benefit the survivors is suicidal.
There's nothing special about our species. If things had happened different would the species that would have been in our place (whatever that means) be better or worse by any metric one chooses? There's no point in wondering about that.
>These explanations are always full of "must have" and "must have been".
Any explanation that starts with "must have" requires that there is a single theoretical framework that could possibly apply to the world as we see it today. Since you reject this necessity, you must think there's an alternative framework that accounts for all observed phenomena (or that accounts at least as well as the theory of evolution). What is that framework?
You say: The fact that production in style S (of artist A) can exceed human consumption capability makes the fact that someone's style can be reproduced without bounds irrelevant. You mention as an example all the AI-generated garbage text that no human will ever read.
I say: Whether it's irrelevant is subjective, but that production in style S is arbitrarily higher with an AI that's able to imitate it than with only humans that are able to imitate it objective, and an artist can (subjectively) not like this and seek to frustrate training efforts.
You say: It's all subjective.
As far as I can tell, we're at an impasse. If we can't agree on what the facts are (in this case, that AI can copy an artist's style in an incomparably higher volume than humans ever could) we can't discuss the topic.
I'm not privy to what goes on inside your head, I can only reply to what you say.
>Then, instead of debating what I meant to say (you know, the actual point of the conversation), you took my words verbatim.
The actual point of the conversation is about intelligent entities (either natural or artificial) copying each other's artistic styles. My answers have been within that framework.
>techniques can be copyrighted ... even patented, or protected via trade secrets.
First, what do you mean by "technique"? We're talking about art, right? Like, the way a person grabs a brush or pencil, or how they mix their colors...? That sort of thing?
Second:
>A patent is a type of intellectual property that gives its owner the legal right to exclude others from making, using, or selling an invention for a limited period of time in exchange for publishing an enabling disclosure of the invention.
Now, I may be mistaken, but I don't think an artistic technique counts as an invention. An artist might invent some kind of implement that their technique involves, in which case they can patent that device. I don't think the technique itself is patentable. If you think I'm wrong then please cite a patent on an artistic technique.
Third, how do you imagine an artist using a trade secret to protect their technique? Unless they do something really out there, most skilled artists should be able to understand what they're doing just by looking at the final product.
>I'll repeat this as well: "Fair use"
Okay, repeat it. I don't know why, since I never said that copying someone else's style or technique is fair use. What I said was that it cannot possibly be copyright infringement, because neither styles nor techniques are copyrighted.
>It's a legal defense when you are accused of stealing someone else's work
I'm not going to reply to any of this until you clean up the language you're using. "Steal" is inapplicable here, as it involves the removal of physical items from someone else's possession. What are you saying? Are you talking about illegal distribution, are you talking about unauthorized adaptations, are you talking about plagiarism, or what?
>"research for commercial purposes is not fair use,"
Sorry, what? What does that even mean? What constitutes "research" as applied to a human creation? If you say there's a crapload of case law that backs this up then I'm forced to ask you to cite it, because I honestly have no idea what you're saying.
Yes, someone can disagree on whether a fact is true. That's obviously true, but it has no effect on the truth of that fact.
I'm saying something very simple: If a machine can copy your style, that's a fundamentally different situation than if a human can copy your style, and it has utterly different consequences. You can disagree with my statement, or say that whether it's fundamentally different is subjective, or you can even say "nuh-uh". But it seems kind of pointless to me. Why are you here commenting if you're not going to engage intellectually with other people, and are simply going to resort to a childish game of contradiction?
I'm not sure what you've changed, but I'll reiterate: my copying your style is not fair use. Fair use applies to copyrighted things. A style cannot be copyrighted, so if you tried to sue me for infringing upon the copyright of your artistic style, your case would be dismissed. It would be as invalid as you trying to sue me for distributing illegal copies of someone else's painting. Legally you have as much ownership of your artistic style as of that other person's painting.
>You might believe they don't have the right to say "no," but they can say whatever they want.
You have a right to say anything you want. Others aren't obligated do as you say just because you say it.
>If you studied my (we will assume "unique") techniques and used them without my permission, then let us say I sue you. At that point, you would claim "fair use,"
On what grounds would you sue me? You think my defense would be "fair use", so you must think my copying your style constitutes copyright infringement, and so you'd sue me for that. Well, no, I would not say "fair use", I'd say "artistic style is not copyrightable; copyright pertains to works, not to styles". There's even jurisprudence backing me up in the US. Apple tried to use Microsoft for copying the look-and-feel of their OS, and it was ruled to be non-copyrightable. Even if was so good that I was able to trick anyone into thinking that my painting of a dog carrying a tennis ball in his mouth was your work, if you've never painted anything like that you would have no grounds to sue me for copyright infringement.
Now, usually in the artistic world it's considered poor manners to outright copy another artist's style, but if we're talking about rights and law, I'm sorry to say you're just wrong. And if we're talking about merely studying someone's work without copying it, that's not even frowned upon. Like I said, it's unavoidable. I don't know where you got this idea that anyone has the right to or is even capable of preventing this (beyond simply never showing it to anyone).
>the artists should still have the right to refuse their art being studies.
No, that right doesn't exist. If you put your work of art out there for people to see, people will see it and learn from it, and be inspired by it. It's unavoidable. How could it possibly work otherwise?
Artist A: You studied my work to produce yours, even when I asked people not to do that!
Artist B: Prove it.
What kind of evidence or argument could Artist A possibly provide to show that Artist B did what they're accusing them of, without being privy to the internal state of their mind. You're not talking about plagiarism; that's comparatively easy to prove. You're asking about merely studying the work.
Whether it's irrelevant is a matter of opinion. The fact remains that a machine being able to copy the artistic style of a human makes it so that anyone can produce output in the style of that human by just feeding the machine electricity. That inherently devalues the style the artist has painstakingly developed. If someone wants a piece of art in that artist's style they don't have to go to that artist, they just need to request the machine for what they want. Is the machine's output of low quality? Maybe. Will there be people for whom that low quality still makes them want to seek out the human? No doubt. It doesn't change the fact that the style is still devalued, nor that there exist artists who would want to prevent that.
Tool improvements only apply a constant factor to the effectiveness of learning. Creating a generative model applies an unbounded factor to the effectiveness of learning because, as I said, the only limit is how much computing resources are available to humanity. If a single person was able to copy themselves at practically no cost and the copy retained all the knowledge of the original then the two situations would be equivalent, but that's impossible. Having n people with the same skill multiplies the cost of learning by n. Having n instances of an AI with the same skill multiplies the cost of learning by 1.
I'm using "implication" as a synonym of "effect". If a human learns to imitate your style, that human can make at most a handful of drawings in a single day. The only way for the rate of output to increase is for more humans to learn to imitate it. If an AI learns to imitate your style, the AI can be trivially copied to any number of computers and the maximum output rate is unbounded. Whether this is good or bad is subjective, but this difference in consequences is objective, and someone could be entirely justified in seeking to impede it.
I'm not sure what you mean when you say different implications existing is subjective, since they clearly aren't, but regardless of who has more say in general terms, the author of a work can decide how to publish it, and no one has more say than them on that subject.
And? Even if neural networks learn the same way humans do, this is not an argument against taking measures against one's art being used as training data, since there are different implications if a human learns to paint the same way as another human vs. if an AI learns to paint the same way as a human. If the two were exactly indistinguishable in their effects no one would care about AIs, not even researchers.
There's a difference between playing to win and being frustrated because your opponent can consistently steamroll you effortlessly without even giving you a chance to set your game up.
Couldn't this just be statistical noise? 2% isn't a huge difference, and if you partition stock funds into two arbitrary groups it's almost certain that one will on average perform better than the other, but not by a lot. The next question to ask should be how much better are stock managers who have an odd number of hairs on their head, compared to those who have an even number.
I mean, it's hard to say whether two different actors really perform the same job; speaking for myself the only name I recognize in that article is Christian Bale's. Plus presumably they all have representatives that negotiate their cuts, so if they agreed to work for that amount, I don't think there's much room to complain. Finally, I can't really tell if there's a trend here. The article talks about the cases of three men and three women, where all three women are paid smaller cuts than the men, but is this the case for everyone else as well?
When you have a selection mechanism, the randomness of mutations becomes kind of moot. Each step of the way that produces a new adaptation only has a handful of optimal (or optimal-enough) solutions. Now, suppose that we're an underwater species that's in the process of developing light-sensitive cells, and from our present genome there are 2^64 possible genes that will produce a good protein for the transduction step (of converting light into some other form of energy). Do we need to mutate 2^64 times to find the "correct" gene? No, all of those 2^64 genes are "correct". Our descendants should not fall into the trap of thinking that because they got gene #5541741487894936799 that it was a special outcome. They could have just as easily gotten gene #5541741556614413535.