Maybe you were never all that creative to begin with
I've been cutting back on some of my addictions, trying to upset the rhythm of consumption - one so slowly familiar, one I didn't even really notice myself slip into. It's irritating. Life takes on a particular flavour when you start each day with 50mg of nicotine to the gums, and as I began to recognize my rationalisations for being just that, rationalisations, it began to set in, just as everyone around me had been saying for months, that the time had come. I conform to the rituals despite wailing about them every now and then, so to make good on my word of dissent, I am off the platforms as well, for the most part. The particularly noxious short form content, which is still insufficiently recognized for the mass inundation it facilitates, is on my black list now. It's been a broken record of total brainrot, AI-based search engine optimisation, and a bunch of zeitgeist-adjusted, photogenic village idiots, speaking on topics they are not equipped to understand; a never-ending cacophony of memetic squawking; information currents coursing unimpeded through a cultural superconductor. Not a cause for concern, at all, that most people, myself included, participate in it for hours every day.
One of the particularly irritating topics of discussion has been that of the morality of the AI-boom. If you'll forgive me my aggravation, I've seen too much nonsense, particularly on this subject, to keep my manners mild. You have two camps, let's call them camp Luddite and camp Mephistopheles.
1.) The Luddites
The luddite opposes the technology due to a perceived threat to their existence. The historical luddite was a worker scared of being replaced by a power tool, the contemporary luddite are the quote-unquote 'content-creators' as they choose to call themselves, whose singular means of sustenance, their attention capital, has been snatched from them by a class of technologies far more adaptable to the task.
A content creator's job is to maintain the attention of a platform's audience, by prancing around and flashing lights at them through their monitors. Often this results in genuine art or educational material, but as these often require more effort to both create and appreciate, profundity is outpaced by multicoloured, incoherent nonsense in terms of price-to-performance to the platform owners. Now, platform owners have simply driven out the human volunteers to this racket by stealing the last thing they didn't yet own, the means to produce huge volumes of information cheaply. Content sources have had their racket stolen from them, they're now redundant.
The luddite position is held by those who use their artistic capabilities to earn money online, and they see themselves as the victims of a system-level breach of intellectual property rights. AI requires a lot of human-made information to effectively create attention-grabbing content, so in a way the weight matrices must contain a small amount of information which was lifted out of a human's intellectual property, and the owners are mad about it. It's easy to emphasise; after all, we're convinced that creators are the most 'like us', at least in concept. Most people have, at one time or another, tried themselves at the content-dance by going to a computer or a smartphone and making a video, it's a fun and outrageously popular pastime. The most famous new personalities on the internet mostly have one thing in common; they leveraged these platforms to attain a meteoric surge in popularity, at a velocity unimaginable to even the most consolidated media moguls of the mid 20th century. It's the individual hitting it big, happening on a windfall with no hand holding; the american dream if ever there was one. Everyone my age seems to dream of waking up one morning validated by crowds of anonymous strangers, and therefore they empathise with those for whom this dream is a reality. A theft committed against the creator is a theft felt by the audience, and the result is near constant outcry.
My problem with the luddite position is that the adherents are quite literally all bark and no bite. Working by the espoused logic of the VC-culture which they view as responsible for this mess, wouldn't we expect to see competing alternatives to the media ecosystems appear spontaneously? If everyone is so aggrieved by this theft, why do they continue to provide the thieves with training data, day by day? The luddite continues to work the factory, making the theft all the more lucrative and effective; with every post complaining of the unfairness of AI-competition on these platforms, the agents themselves become more effective at reproducing attention-grabbing signifiers. Their racket is being stolen back by the companies which lent it them in the first place: the attention they helped maintain was never actually their property to begin with, it was owned by the platform maintainers and their advertising partners. Their days in this mode of transaction are numbered; this is obvious even to them, and so the grit with which they cling to what used to be their gravy train is nothing more than cynical self-destruction.
I never really understood intellectual property law for art in the first place, beyond the idea that it served the much mutated industry of popular media, that without it as its basis, the business model is impossible. Many marxists have made the argument that this situation is proof that the law is always going to side with capital, and I think that analysis is true insofar as it remains amoral. It's not a violation of IP law to publish an album, while secretly having musical influences which you don't disclose. You could get caught up in one of those farcical trials where lawyers draw up some drivel to try to prove that it's possible to own clusters of integer ratios, even though they're unbound to the world of things, but it's more or less impossible to prove you were inspired by a particular artist, it can only be hinted at through choices of style.
The crux of the argument appears to be the question of whether or not it should be that one can own a number. It would be silly to start selling deeds on numbers like small integers. What would it even mean for someone to own the number 3? What does this owner get to do with the number 3 which everyone else is excluded from? Where this changes, is when the numbers get large enough (that is, require a lot of bits to express) that they can't trivially be arrived at, so the idea of property reduces to knowledge of this number. Any computer file is basically just a long number, just a microstate of some wacky configuration space, arrived at through some algorithm which a human's actions either are or aren't a direct part of. So the logical basis of the argument for AI as system-level IP theft just doesn't exist, in my opinion, because the AI does not reproduce any training data absolutely verbatim, it's designed not to. Even if you could prove that your art was used to train the AI, unless it spits out exactly what you made, there's just no case to be made for plagiarism, insofar as the theft of the knowledge of the underlying number is concerned. Let's say you loosen the constraint here, and suppose there exists a case for plagiarism where the numbers *approximate* each other, maybe they differ by one bit, or a couple, or a couple thousand, and so long as they meet some threshold for similarity, they are considered equivalent. Then you run into the ship of theseus... you can nudge any file bit by bit until it represents any other file, and every step of the way will have had one owner before IP law. The contradiction is that either IP law is so specific that it's useless to establish ownership within reasonable bounds, or it's so malleable that everyone already owns every piece of information. Either way, ownership of digital art is impossible to rigorously define, in a way that implies that AI companies are guilty of IP violations.
Luddites urgently need to shut up and direct their energy elsewhere: they're fighting a provably losing battle, and have been for all of history.
2.) The Mephistophelians
These are the people who embrace these postmodern malefactions and declare their optimism. They act like these technologies are helping our civilisation perform tasks which used to be impossible, rather than just making millions of sentient beings materially inconsequential. Their position is more obviously idiotic than that of the luddites, because their naive and self-serving accelerationism is arrived at through less deliberation than the luddite's objections. They celebrate the idea that a human life can be made economically worthless, and that the technology to do this exists now. Most of them are operating on the belief that they will be spared because they own the machines; they are wrong. This mathematical paradigm will eventually swallow and appropriate every single thing which makes a human mind, and they're helping it all happen, cheering it on.
My stance here is this. We are at a point in history where we are getting shafted by reality itself; late capitalism will continue to deteriorate, it's happening on its own, without steerage, and it will not stop. What we're doing on our planet is proof of technological determinism, which itself might even be proof of genetic determinism. One could argue that it's written in our very DNA, that we would one day gain this near-unfathomable technology and simultaneously we would be, at large, entirely unable to discuss this in anything but braindead, pedestrian terms. To oppose it is to kid yourself, to hang on to the irrational belief that we are better than we are. If a tensor can reliably store everything you like about your creative process, and steal from you what you view as your own special creative formula, maybe you weren't all that creative to begin with, or creativity itself is meaningless.
Sad, but rigorous. And if you're apalled by what I've said, cherish your disgust, because it's proof that you're more than a consumer, I'd ask you to draw some hope from that, if nothing else. What I should also mention is: it might turn out to be true that biological life is a more optimal facilitator of thought than anything we can create to imitate it, i.e. the mathematics of learning in neural networks contains scaling limits which explain why our brains aren't larger than they are. Maybe this technology will annihilate selection pressures beyond those required for a machine to equal a human. Maybe biology catches up to technology. Who knows. We don't really have a way of knowing if all these hypothesised advances will happen until they are actually brought about. I find not knowing the answer rather exhilirating, which is why I can't stand these talking heads who seem so disgusted by uncertainty. Both Luddites and Mephistophelians are epistemic cowards.
Comments
Post a Comment